1. We're built for engineers

What does this mean?

• It means we ship fast and iterate based on user feedback.
• It means engineers do support – all product teams have a weekly support hero.
• It means we have extensive public and private API endpoints, and a powerful SQL query builder.
• It means we built SDKs for all popular (and many unpopular) client-side, backend, and mobile languages and frameworks.
• It means we make it easy to test in production, conduct phased rollouts, run beta programs, and so much more.

2. We're an all-in-one platform

Heap is mainly focused on product analytics. This means you need to adopt additional tools for things like feature management, experiments, and surveys. They're all built into PostHog – we even have a built-in data warehouse that integrates with Stripe and Hubspot. You can replace half a dozen tools with PostHog, save money, and get more from your data.

3. We're totally transparent

• How much will PostHog cost? Use our pricing calculator.
• What are we working on? It's on our public roadmap?
• How does sales work? We have a whole page on it.
• What do we care about? We explain everything in our public company handbook.
• How do we make money? That's in the handbook, too.

Oh, we're open source, too. Go take a peak at our code if you like on our GitHub repo.

Comparing PostHog and Heap

Platform

💡 Good to know: PostHog can replace multiple tools, such as Hotjar, Google Analytics, and LaunchDarkly. This makes it a lot easier to extract usable insights, since you don't have to constantly switch between tools. Running PostHog on both your product and website makes it easier to understand how marketing activity influences signups and usage, too.

Product analytics

💡 Good to know Every PostHog user gets 1 million events for free each month. You can also save money by sending us anonymous events for non-identified users, which are up to 80% cheaper than identified product analytics events. Anonymous events are ideal for tracking behavior on marketing websites, or mobile apps with large consumer audiences. See our docs on anonymous vs identified events for more.

Session replays

💡 Good to know: Session replays are an essential tool for understanding how people use your product, especially for early-stage companies searching for product-market fit. Both Heap and PostHog offer session replay, though Heap lacks many developer-facing features like a DOM explorer, performance monitoring, and network events, which are useful for fixing bugs and performance issues.

Feature flags

💡 Good to know: Feature flags make it easy to roll out features to specific users or groups, and safely test in production. Our feature flags are also tightly integrated with other features so you can target session replays, surveys, and more using existing feature flags. See our guide on the benefits of feature flags for more.

Experiments

💡 Good to know: Our experiments integrate seamlessly with our feature flag product. This means you can easily deploy the winning variant of an experiment with a single click from the experiment UI.

Security and compliance

💡Good to know: Additional compliance features, such as HIPAA Business Associate Agreements, advanced permissions, and audit logs are available on our Teams add-on, which also includes our managed reverse proxy and white labelling for surveys and shared dashboards.

Frequently asked questions

Can I migrate from Heap?

Yes. See our Heap to PostHog migration guide for more.

How long does it take to deploy PostHog?

Deploying PostHog takes mere minutes. Just include our tracking snippet in your product or website and (thanks to autocapture) you're immediately capturing events. Once up and running, you can enable session replays and other features via your project settings. See our guide to getting started on PostHog Cloud for more info.

Can PostHog also replace Google Analytics?

Yes. PostHog can replace Google Analytics for many use cases – our marketing team uses PostHog, for example. You can integrate PostHog into your website using Google Tag Manager. See our comparison of PostHog and Google Analytics 4 and an intro to PostHog for Google Analytics users for more.

How can I estimate my usage?

The easiest way is to sign up to PostHog, integrate our snippet, then check the projection on your billing page after a few days. Alternatively, you can guesstimate by multiplying your current monthly active users by an estimate of events generated per user – 50 to 100 per user is a good starting point. See Estimating usage & costs in our docs for more.

Does PostHog block bots by default?

Yes. See the full blocklist in our docs.

What about ad blockers?

We recommend all users deploy a reverse proxy, which enables you send events to PostHog Cloud using your own domain. Events sent from your own domain and are less likely to be intercepted by tracking blockers, ensuring you capture the best data possible.

We offer a managed reverse proxy on our Team plans, and we have reverse proxy setup guides for AWS Cloudfront, Caddy, Cloudflare, Netlify, Vercel, and more, in our docs if you want to run your own.

Can I use PostHog with a CDP? (Segment, Rudderstack, etc.)

Yes. See Using PostHog with a CDP in our docs.

Can you use PostHog on e-commerce websites?

Absolutely. PostHog is easy to integrate with Shopify and WooCommerce. You can easily install PostHog on other e-commerce platforms using our Javascript snippet – see our guides to setting up Webflow analytics and Wordpress.

How does PostHog compare to Amplitude and Mixpanel?

Amplitude and Mixpanel offer similar features to Heap. Read our PostHog vs Mixpanel and PostHog vs Amplitude guides for more info. You may also find guide to the most popular Heap alternatives useful.

Sure, ChatGPT is pretty cool, but when it comes to AI chatbots that try to replace a human in a product support context, I've generally been underwhelmed. The experience generally isn't great, and I'm often left with more frustration than before I skeptically asked the question in the first place.

But it's not all the fault of LLMs. It also has to do with the available information about a product that an LLM consumes. And for a technical product like PostHog, the documentation has to be extensive enough to answer very nuanced questions. (This is tricky as information is constantly changing and there's always room for improvement!)

As the product owner totally official "webmaster" of PostHog.com, I've contemplated how we can make use of AI to help answer questions... and until now, I've lacked a good solution.

Why AI chatbots have a bad rap

There are a lot of "wrong" ways to do AI chatbots – and there are many examples of this even before we collectively started using true AI.

How many times have you tried to create a support ticket but then:

• been first prompted to review a list of "related questions" that may solve your issue
• been blocked from submitting until you explicitly confirm that none of the suggested answers are helpful

And now in a world of actual AI, we see a new set of issues – largely centered around LLMs not understanding the question, the context, or straight up hallucinating.

In no world would we want to subject our very technical audience to these types of dreaded experiences.

This isn't to say all AI chatbots are bad. A couple of exceptions come top of mind mainly Intercom and Shopify. But the status quo that the industry has to overcome is the negative connotation of historically bad chatbot interactions.

Our approach

The reason we've been so cautious to introduce AI is that we're very intentional about earning (and not losing) the trust of our audience. We do this in a handful of ways:

1. We don't use fancy words that don't add actual value
2. We try to be relatable and show a personality
3. We traded in the value-based sales model for a no-BS approach (self-service, fair, usage-based pricing)
4. We avoid gimmicky marketing tactics (like popups and jamming CTAs into every available whitespace)

So for an AI solution to mesh well with our brand's vibe (yes, we're about the vibes), it needs to be accurate and succinct in its replies, and never make up a false answer.

After all, if you interact with an AI chatbot and don't get the answer you're looking for, you're likely to write it off entirely. And that's exactly what I want to avoid.

Where we decided to implement AI

The most natural place to include an AI chatbot would be in our docs. But I was hesitant to start here, as the conversational nature of AI meant we'd lack control over the quality and accuracy of the responses. So I decided to try something different.

Our community questions section is where users can ask questions directly in the docs. And we currently have... a lot of open (unanswered) questions. But because they're posted in a forum-style format, there's no expectation of a live chat experience. And this presented an opportunity...

Today customers may wait hours (or even longer) for a response from a human. So what if AI could answer only the questions it was highly confident it could answer, then leave the rest for a human?

Using this approach, AI can add value by instantly responding if it has a high degree of confidence, but it won't detract from the experience if it doesn't have the solution, simply leaving those questions open as they normally would.

Introducing Inkeep

This whole idea came about from an Inkeep ad on LinkedIn – which is incredible to say because I'm generally allergic to ads. But Inkeep's ad spoke to me.

I took the bait, and they quickly set me up with a custom sandbox to try them out. They're indexing our docs, tutorials, blog posts, handbook, existing community answers, and even GitHub repos. (This also means it's aware of bug reports, feature requests, and open pull requests - huge!)

They even themed the sandbox to match our brand. <3

To vet the quality of the answers, I started copy/pasting unanswered community questions into the sandbox and seeing what Inkeep could produce. And to say I wasn't let down should tell you something. It was good enough to where I was willing to try it out on our real questions.

How it works

Inkeep's pre-built library of UI components works great for most companies. But out of the kindness of their hearts, they went the extra mile for us.

Here's how our setup works:

1. When a new question is posted, we send it to Inkeep's LLM.

2. Typically within ~20 seconds, we get a response. During this time, we show that we're searching for an answer.

3. In the response, we get a confidence score.

   const KnownAnswerConfidence = z.enum([
     "very_confident",
     "somewhat_confident",
     "not_confident",
     "no_sources",
     "other",
   ]);
   

   If the confidence score is high enough, we show the answer to the user, credited to our hedgehog mascot,

   .

   If Inkeep isn't confident in the answer, we show a message that says we couldn't find an answer, and that a human will appear as soon as possible.

When an AI answer is presented, we show feedback buttons so the original poster can tell us if the answer was helpful. We use this information to improve the quality of the answers.

Demo

Here's what it looks like to submit a question and receive an AI answer from Inkeep.

How we're tracking feedback

As you can imagine, every time a customer rates a response, we're sending the information into PostHog as a custom event. Along with the event, it sends properties like the sentiment ("helpful" or "not helpful") and the category of the question. (This will help us identify if certain areas of our documentation needs to improve.)

Now that this is live, we're monitoring feedback in Slack. PostHog's recently introduced support for realtime destinations in data pipelines.

This allowed me to configure notifications when we get feedback about an AI answer.

So how's it working?

Overall, it's working pretty well! In fact, it worked so well in one case that the original poster replied to share his delight.

In other examples, we were receiving feedback that the answer wasn't helpful. But after reviewing the response, the best way to accurately summarize the bulk of feedback is: "The answer is correct, but I didn't like the answer." (This is usually the case when a feature isn't yet supported and it's not the answer the customer was hoping for.)

This resulted in a lower overall success rate. So we modified the response options to better understand the different reasons why a response might not meet the mark.

With these additional feedback options, we're able to use it to inform areas for improvement. For example, if someone says, "I think this is a bug", we'll take a closer look. Or if they choose the "Answer is wrong" option, we'll take a look into what might be missing from the documentation.

It's also won over some internal support from developers on the PostHog team:

We're still early in our days of exploring AI – and don't worry, AI is coming to the PostHog app when we feel it's ready.

But I couldn't give you a better rationale for a slow rollout than how a colleague explained it to me:

Hopefully the next time you ask a question in our community, you'll get the answer you're looking for. And even better – if you get it even faster than a human could answer.

Founders have no end of excuses for this. It’s more important to build your MVP and validate it, or you need to find product-market fit and make money so you can survive.

Creating a cool website won't help with either, right?

Wrong. Let me explain why.

"What would stop me from using this thing?"

This was the most important conversation Tim and I had before launching PostHog.

It was going to be free on launch day, but we knew users would still be wary of wasting their time deploying a piece of software produced by a two-person startup.

Being open source helped because people could inspect the code, but we concluded the majority would turn away at the website, the docs, and the readme first.

That's why we set out to create a website and brand that felt both unique and intentional, including a public handbook that explained what we were doing, why, and how.

An obviously templated or generic website signals:

1. The team behind the product isn't very strong. The product probably won't improve quickly, or become a "must use" tool. Those startup launches that everyone hears about? They've put energy into their appearance.

2. The product may not exist in a few more weeks. Why should a potential customer invest time in a product if it's obvious the team making it threw its website together in a day, or ripped it from an existing template?

So, as frustrating as it might be to spend extra time and effort on something "superficial" like a website, remember you're providing your product to human beings. Everyone judges books by their covers.

You need to create your brand for someone...

One of the most important things we did was to create an Ideal Customer Profile.

When we launched, we weren't sure whether to focus on engineers, product managers, or both. But we soon realized that we could go all-in on engineers and technical founders, who now make up around 80% of our users.

Our brand and product has become more oriented towards these users as a result.

Here are some ways we got more specific for this audience after we zeroed-in on it:

• We added code examples to the homepage, such as how to track an event, or identify a user, and examples of what you can do with our API.

• We made buying PostHog easy – the "You'll hate PostHog if" section of our homepage calls out the pet hates of engineers when buying SaaS products.

• We chose a transparent, flexible pricing model. It's more complex, but also much better value. We’ve found technical users tolerate a little more complexity if it helps them save money, and only pay for what they use.

• We focused on power user features, such as SQL access and a built-in data warehouse, so technical people can get to the underlying data and debug things.

• We redesigned our entire UX to have more flexibility, greater information density, and a dark mode. We want PostHog to look like a dev tool, not an analytics product.

...but your brand must be a reflection of you

We’re often impressed at the level of polish others achieve. But, quoting Cory, our lead designer:

"We aren't the best in the world at being polished, but we can be the best in the world at being ourselves".

We sometimes make business decisions based on who we are, and what we want to do – not always what may, for example, generate the most revenue growth.

We chose this path because there was less competition for a self-hosted product. We could build a refined point solution that anyone could self-host, and not worry about competing with cloud-based competitors with over $100m in annual revenue.

But we soon realized we just love shipping new features. After our first hackathon, we wound up with session replay – i.e. a new feature big enough to be its own product. It took off and this resulted in our current multi-product approach.

Eventually, we removed our paid, self-hosted offering entirely, and went all in on multi-product in cloud, and "getting in first" with early-stage companies, not enterprises.

We kept the open-source product, though we stress it's for hobbyists because it's cheaper and easier for companies to have us host it for them.

This strategy works because:

1. Being broad lets us compete by being way wider than everyone else.
2. Being broad also means we can charge less by cross-selling multiple products.
3. It reflects who we are – passionate engineers who love shipping new products.

Had we stuck with the original plan, we’d have lost interest because we weren’t doing what we enjoy. We apply the same mindset to our brand.

If you don’t love it, why will anyone else?

Your brand isn't just your website...

It's how your company is experienced by others. Your company is the product and:

• How much you invest in each function in your company relative to each other.
• What you look for when you hire people, and the process you use to hire them.
• How you decide what to build. Do you ship features early, or do you focus on meeting the needs of your largest customer with every new thing?
• The degree to which you’re led by design, engineering product, sales, or marketing. How many people are in each function, but also your processes for deciding what to work on.
• The frequency and style with which you email your customers, for marketing, product updates or your activation flow.
• The kind of people you hire into customer success or sales.
• The background and experience of those in your marketing team.
• The tone of voice on your website.
• Who is on your board.
• How your pricing works.
• How your social media works.
• How helpful you are to random founders that message you.
• The marketing channels you pick.
• What you talk about in your blog. I'm writing this piece for brand reasons –transparency / equipping developers with knowledge to build successful products. I don’t care if it converts people, I just want to leave an impression. And it's who I am – I'm writing this to pass the time on a long flight.

Everything matters.

You can't 80:20 everything

A lot of design or brand work has an increasing marginal return. This is because a great brand's purpose is that you stand out. You do not stand out by being average!

So, if you're serious, go nuts on your brand. Go weirdly hard.

To give you an example, we built a merch store:

• We started with me emailing users for their address and size, ordering merch and having it delivered to my house, and me then posting it internationally to them. Do things that don't scale.

• Next, we set up a store on Shopify. I just threw our logo onto an amusingly wide selection of items – PostHog shower curtain, anyone? It was cool to just have merch for sale and a way of giving it.

• Then we created custom components and built it into our website directly, so it's perfectly integrated.

• Then we booked a photo shoot and took custom photos of our team wearing everything.

• Now we're working on limited editions of certain items to create collectible series of t-shirts

Who knows what happens next!

This single thing is easy enough to copy, but it's just one activity among hundreds that builds a brand.

Doesn't this imply doing everything to a really high standard and a ridiculous workload?

Yes. You can be iterative with it, though.

Not everything was amazing on day one, but we keep improving and tweaking everything over time.

So get out there, and be bold

If your brand isn't going to turn off some people that might have bought your product, then it isn't strong enough to stand out.

Engage your best customers with your brand, and put off your worst ones.

Matomo focuses on tracking sessions, making it ideal for running analytics on large content and e-commerce websites. It's designed to look and feel similar to Google Analytics 3, AKA Universal Analytics.

PostHog can also replace Google Analytics, but it offers a wider suite of tools for product analytics, web analytics, session replays, feature flags, A/B testing, user surveys, and much more.

How is PostHog different from Matomo?

1. It's an all-in-one platform

Matomo charges extra for features like funnel analysis, cohorts, path analysis, A/B testing, and session recording. These are also considered add-ons and are not tightly integrated with the main analytics product. In PostHog, all of these tools work together and each comes with a generous free tier.

2. It's built for engineers

Rather than focusing on marketers like Matomo, PostHog focuses on the tools engineers need to build better products.

On top of product analytics, this includes A/B testing, feature flags, a complete API, direct SQL access, and more. PostHog also includes a data warehouse so you can import and query data from other sources as well as destinations to send your data to external warehouses or other tools.

3. Seamless integrations with the tools you already use

PostHog is built to work with the tools you already use.

That means you can import data and query from sources like Stripe, Hubspot, Zendesk, S3, and more. You can also export data in batches to your data warehouse like Snowflake or BigQuery as well as a range of real time destinations like Google Ads and Slack.

Feature comparison

Platform

Both PostHog and Matomo offer a range of tools for tracking and analyzing your site and product. PostHog offers more tools for understanding and improving your product, while Matomo focuses more on marketing analytics.

💡 Good to know: Matomo's first paid tier comes with 1,500 heatmap pageviews and 150 session recordings per month.

Web analytics

Both PostHog and Matomo offer all the features you expect from Google Analytics alternatives.

Product analytics

Product analytics reveals the evolution of both tools. While PostHog has always focused on product analytics, Matomo has expanded its offering from a focus on web analytics.

Integrations

A simple way to compare integrations:

• PostHog has more integrations with data and product tools.
• Matomo has more integrations with ecommerce and CMS platforms.

But this doesn't mean either lacks those types of integrations.

💡 Good to know: Although PostHog doesn't have dedicated integrations for CMS or ecommerce platforms, our script snippet makes it easy to use PostHog with basically any of these including Shopify, WordPress, and Webflow.

Security and compliance

Matomo positions itself as a Google Analytics alternative that protects your data and customer privacy, but PostHog offers all of its privacy and compliance features (and more).

💡 Good to know:

• Matomo's on-premise offering can be made HIPAA compliant, but not their cloud offering.
• PostHog is self-hostable under a MIT license without guarantee or support.

Frequently asked questions

How much do PostHog and Matomo cost?

Matomo's Cloud offering starts at $29/month for 50,000 hits. This includes 1,500 heatmap pageviews and 150 session recordings per month.

PostHog is entirely usage-based. Its free tier includes:

You'll never pay anything if you stay within these limits and you can set billing limits to avoid surprise bills.

Does PostHog offer a free trial?

Matomo offers a free trial, but PostHog doesn't, what gives?

Instead of a free trial, PostHog offers a free forever tier as long as you stay under the free limits. You don't even need a credit card, but adding one unlocks all the features (we have billing limits you can use too).

Can PostHog also replace Google Analytics?

Yes. PostHog can replace Google Analytics for many use cases – our marketing team uses PostHog, for example. You can integrate PostHog into your website using Google Tag Manager. See our comparison of PostHog and Google Analytics 4 and An intro to PostHog for Google Analytics users for more.

Can I migrate from Matomo to PostHog?

Yes. See our Matomo to PostHog migration guide for more.

Is PostHog hard to install and use?

Although PostHog has many features, it isn't necessarily complicated. Our web analytics dashboard is similar to Matomo's:

Our install snippet is also only a couple of lines:

<script>
    !function(t,e){var o,n,p,r;e.__SV||(window.posthog=e,e._i=[],e.init=function(i,s,a){function g(t,e){var o=e.split(".");2==o.length&&(t=t[o[0]],e=o[1]),t[e]=function(){t.push([e].concat(Array.prototype.slice.call(arguments,0)))}}(p=t.createElement("script")).type="text/javascript",p.crossOrigin="anonymous",p.async=!0,p.src=s.api_host+"/static/array.js",(r=t.getElementsByTagName("script")[0]).parentNode.insertBefore(p,r);var u=e;for(void 0!==a?u=e[a]=[]:a="posthog",u.people=u.people||[],u.toString=function(t){var e="posthog";return"posthog"!==a&&(e+="."+a),t||(e+=" (stub)"),e},u.people.toString=function(){return u.toString(1)+".people (stub)"},o="capture identify alias people.set people.set_once set_config register register_once unregister opt_out_capturing has_opted_out_capturing opt_in_capturing reset isFeatureEnabled onFeatureFlags getFeatureFlag getFeatureFlagPayload reloadFeatureFlags group updateEarlyAccessFeatureEnrollment getEarlyAccessFeatures getActiveMatchingSurveys getSurveys getNextSurveyStep onSessionId".split(" "),n=0;n<o.length;n++)g(u,o[n]);e._i.push([i,s,a])},e.__SV=1)}(document,window.posthog||[]);
    posthog.init('<ph_project_api_key>', {api_host: '<ph_client_api_host>', person_profiles: 'identified_only'})
</script>

Our users range from individual indie hackers to YC-backed startups to massive enterprises.

Why image-based updates?

Image-based posts perform better on X than linking out to another site.

So that works great for the occasional visual essay, but for changelog posts, we had a problem... we didn't have the bandwidth to create images for every new changelog post – nor should design be a blocker for the marketing team. And I didn't want to force anyone to use Figma to create these images.

So we started to think... could we productize this so our team could generate high-quality images automatically? After all, our changelog updates already include most of the information we'd need.

A couple years ago, we had done something similar for auto-generating custom open graph images for most pages on the website, so we thought... why not do something similar here?

Figma mockup

So I got to work, mocking up a template that would pull information from our self-rolled CMS.

How we did it

I wrote up an issue outlining how I thought we'd be able to combine data from changelog updates, small teams, and team members for use in the rendered images.

Changelog data

Our changelog entries already included title, description, screenshot, update_type, and associated product and team.

Product branding

Each PostHog product has an icon and color associated with it. These are defined in a JSON file that builds our navigation submenus.

{
  name: 'Data warehouse',
  url: '/docs/data-warehouse',
  color: 'lilac',
  icon: 'IconDatabase',
}

(The color maps to a HEX value defined in Tailwind.)

Customization

Then it just comes down to visual polish, so we added a few options.

Hedgehogs

And what PostHog graphic wouldn't be complete without a hedgehog?

Recently,

This made it easy for our front end developer

Now we can choose any hedgehog image from our library – as long as it's uploaded to a folder in Cloudinary.

Generating the image

We used html-to-image to render the final product. When a team member clicks the download button, the image is generated in a very clever way, then downloaded to the client.

The final product

And voila, here's an example of our first image produced with this new tool. You should check it out on X - and give us a follow while you're there!

How is PostHog different?

1. We're an all-in-one platform

PostHog goes beyond replays and user surveys by integrating them with product and web analytics, feature flags, A/B testing, and a data warehouse. It's everything you need from a single app with a single contract. A genuine single source of truth for your product and customer data.

2. We support both web and mobile apps

We offer integrations for all the most popular web and and mobile SDKs, including Next.js, React, React Native, Android, iOS, and Flutter. Unlike Hotjar, we also support session replay on mobile apps through our iOS and Android SDKs – we're working on React Native and Flutter support, too. This means you can track all versions of your app using PostHog.

3. We're cheaper and have a generous free tier

Every customer gets 5,000 recordings and 1 million events for free each month, no matter what plan they're on. Our generous free tier means more than 90% of companies use PostHog for free, and you'll pay less when you do use up your free credits.

See the price comparison section for more.

Product comparison

The best way to imagine PostHog is as an alternative to Hotjar, Heap, and LaunchDarkly rolled into one.

Analytics

Product analytics isn't a core focus for Hotjar, but it does offer basic event tracking, funnels, and trend insights on its most expensive Scale plan.

Session replay

Session replays are an essential tool for understanding how people use your product, especially for early-stage companies searching for product-market fit.

Both Hotjar and PostHog are good options if you're looking for a Microsoft Clarity alternative, though Hotjar doesn't support replays on mobile apps and has no public plans to do so.

Note: Our mobile session replay is free while it's in development, and we're actively working on session replay support for React Native and Flutter, too. See our mobile replay docs for more info on development and how replay works on mobile apps.

Heatmaps

Heatmaps visualize where people click and navigate to on your app or website and, when combined with session replay, give a clearer overview of how users behave.

Surveys

Hotjar supports more survey types than PostHog, but fewer targeting options.

Using PostHog's product analytics and surveys together unlocks more precise targeting options for surveys, such as using feature flags and pre-defined cohorts.

Price comparison

Note: Prices correct as of Aug 21, 2024.

PostHog is cheaper than Hotjar for both session replay and user surveys, regardless of whether you choose Hotjar's most popular Business tier or the more expensive Scale tier that includes Hotjar's basic analytics features, and a dedicated account manager.

Session replay pricing

Hotjar charges based on 'sessions per day', whereas PostHog uses monthly recorded sessions. We've multiplied Hotjar's daily sessions by 30 to get a monthly equivalent.

At 75k recorded sessions per month, PostHog is $1,980 cheaper over 12 months compared to Hotjar's Business tier, and $7,236 cheaper than Hotjar's Scale tier, when paying monthly.

PostHog also offers a more generous free tier that includes 5,000 replay sessions, 1 million analytics events, and 250 survey responses for free every month.

Hotjar doesn't offer any free usage outside of its extremely limited Basic tier.

Surveys pricing

It's similar story for user surveys, where PostHog is cheaper than both Hotjar's Business and Scale tiers. At 2,500 survey responses each month, PostHog is $324 per year cheaper than Hotjar's Business tier when paying monthly.

You can see how much PostHog will cost you using the calculator on our pricing page.

Platform and integrations

Data sources

PostHog has a built-in data warehouse, so you can import, combine, and analyze data from multiple sources, and combine it with analytics events. This means you can monitor deal counts using data from Hubspot, analyze gross revenue using Stripe data, and track support ticket SLAs using Zendesk data, all from within PostHog.

Data destinations

Hotjar and PostHog support a similar range of data destinations, though PostHog can send more data because it collects more actionable data via product analytics. Hotjar is limited to sending basic event data, session replay, and survey responses.

Privacy, compliance, and security

Regulatory compliance can be a critical need for many teams, especially if they operate in financial or healthcare industries. Regulations such as HIPAA and GDPR can require teams to store data in certain locations or protect it in certain ways.

Frequently asked questions

Can PostHog replace Google Analytics?

Absolutely. We include an easy-to-use web analytics dashboard for tracking marketing and content performance, and you can use our product analytics to do more in-depth analysis as well. See our PostHog vs Google Analytics comparison for more.

Can PostHog replace Mixpanel, Amplitude and Heap?

It's what PostHog's designed to do. The beauty of PostHog is replaces multiple tools in your analytics and product stack, so you can stop juggling data between multiple tools, and save money by consolidating everything. See our comparisons with Mixpanel, Amplitude, and Heap for more.

Can I choose which sessions are recorded?

Yes. You can choose to only record sessions on specific pages, or sessions with a minimum duration that you set. You can also set a sampling rate – i.e. only record 20% of sessions – or trigger recordings based on a feature flag. See How to only record the sessions you want for more.

Does PostHog offer discounts for paying annually?

Yes. You get a 20% discount by default for paying annually upfront. Hotjar also offers a 20% discount for paying annually.

Can I use PostHog with a customer data platform like Segment?

Yes. See our guide to using PostHog with a CDP for more on that, though many of our customers use PostHog as their CDP and utilize our data pipelines and data warehouse for moving data in and out of PostHog.

Got another question? You can ask the PostHog team anything you want!

Typical users: Product managers, engineers, support teams

What is LogRocket?

LogRocket is a product experience platform that focuses on helping product managers, engineers, and support teams fix issues. In addition to session replay, it provides tools to identify the most impactful issues, such as crash reports, stack traces, performance monitoring, and product analytics.

Supported features

How much does LogRocket cost?

Its Team plan starts at $199/month for up to 10k sessions and includes most session replay features, but excludes other features such as product analytics and some error tracking features.

The Professional plan starts at $420/month for up to 10k sessions and includes all their features, but requires an annual commitment.

Why do companies use LogRocket?

Reviewers on G2 use LogRocket for:

1. Identifying problems: LogRocket's combination of error tracking, performance monitoring, and session replay makes it uniquely powerful at finding bugs and issues.

2. Improves user experience: LogRocket helps reviews fix issues with their apps. Reviewers find it provides all the tools to improve their user experience, mostly by solving what's wrong rather than feedback and planning.

3. High usability: LogRocket provides a lot of functionality out of the box. It captures the details users need, provides useful visualizations, and automatically triages some issues.

Bottom line

LogRocket is great choice for teams whose top priority is identifying and mitigating issues.

2. UXCam

Typical users: Product managers in enterprise companies

What is UXCam?

UXCam is a platform built solely for mobile apps. It includes a full-suite of analytics features as well as session replays, heatmaps, and crash analytics.

It's designed to provide deep insights into user behavior, helping product managers understand how users interact with their apps.

Supported features

How much does UXCam cost?

UXCam's pricing is sales-driven and completely opaque, so it's not clear how much it costs. The company's sales team will give you a quote after you've spoken to them.

Why do companies use UXCam?

Looking at G2 reviews, companies choose UXCam to:

1. Insights on user behavior: Shows what they are doing, where their attention is going, and how they are spending their time. Users love the realizations they get with UXCam's tools.

2. It's easy to use: Reviewers note how easy UXCam is to set up and use. This means they can go from idea to actionable learnings quickly.

3. Heatmaps and scrollmaps: Seeing clicks and scroll depth works well and creates actionable insights for design and product teams. Helps improve UX and page designs.

Bottom line

UXCam is ideal for product managers who need in-depth insights into user behavior. Despite opaque pricing, its ease of use and actionable insights make it a valuable tool.

3. Microsoft Clarity

Typical users: Marketing and content teams

What is Microsoft Clarity?

Microsoft Clarity is a 100% free session replay tool that supports most replay features and heatmaps. It integrates with Google Analytics 4, so you can view your GA data in Clarity.

It includes most of the session replay features you would expect, but doesn't support any additional features that some of the other choices on this list do, like crash reports or debug logs.

Supported features

How much does Microsoft Clarity cost?

Clarity is completely free.

Why do companies use Microsoft Clarity?

According to reviews on G2 and Capterra, it's because:

1. It's easy to set up and use: Setting up Clarity doesn't require much technical knowledge and users mostly like the simple user interface.

2. It helps uncover how users navigate: Clarity users are mostly using it to understand how people navigate their app and identify opportunities for improvement.

3. It's free: The short 30-day retention limit is easy to forgive when it's free.

Bottom line

If you only care about session replay and dont need other features, Clarity is the obvious choice - especially if you're already using Google Analytics. On the other hand, software engineers will find Clarity's lack of debug logs disappointing.

4. PostHog

Typical users Engineering and product teams

What is PostHog?

PostHog (that's us 👋) is an all-in-one platform built to help engineers create better products. It includes session replays and a whole bunch more, such as product analytics, A/B testing, and feature flags.

It's designed for product-minded engineers, growth teams, and product managers who need to move fast and iterate based on reliable, actionable insights.

Supported features

How much does PostHog cost?

PostHog has transparent pricing based on usage. It’s free to get started and completely free for the first 5,000 recordings. After this free monthly allowance, pricing starts at $0.0050/recording (or $50 for 10k recordings), and recordings cost progressively less the more you use. This makes PostHog significantly cheaper than all the other companies on this list (apart from Microsoft Clarity).

Why do companies use PostHog?

According to G2 reviews, companies use PostHog because:

1. It's many tools in one: PostHog can replace tools like Mixpanel (analytics), UXCam (session replay), and Statsig (A/B testing and feature flags). This simplifies workflows and ensures all product data is in one place.

2. They need a complete picture of users: PostHog includes every tool necessary to understand users and build better products. This means creating funnels to track conversion, watching replays to see where users get stuck, and testing solutions with A/B tests.

3. Pricing is transparent and scalable: Reviewers appreciate how PostHog's pricing scales as they grow. There's a generous free tier. Companies eligible for PostHog for Startups also get $50k in additional free credits.

Bottom line

For teams looking for all the tools they need to improve their products, PostHog makes for a great choice. This is especially true for startups and scaleups thanks to it having a generous free tier.

5. Smartlook

Typical users: Engineers, business analysts, product managers

What is Smartlook?

Smartlook combines session replays, product analytics, visualizations, and crash reports to generate an overall understanding of user experience. It focuses on mobile apps with specific tools like native rendering and wireframe mode.

Supported features

How much does Smartlook cost?

Smartlook's free plan records up to 3k sessions and includes most of their replay features. Its pro plan includes all replay features and starts at $69/month for 5k sessions or $99 for 10k sessions. Pricing becomes progressively cheaper with more monthly sessions.

Why do companies use Smartlook?

According to G2 reviewers, Smartlook users benefit from:

1. The integration between replays and events: Smartlook connects event-based analytics and sessions. This enables users to dive deeper into user behavior than a single tool provides.

2. Understanding visitor pain points: The analytics and visualizations make it easy to understand where users are running into trouble. Reviews use this to improve the user experience and conversion in these areas.

3. Real user monitoring: Smartlook shows how real users are using your app and monitors the quality of their experiences. Reviewers use it to figure out what areas are confusing or used improperly.

Bottom line

Smartlook's combination of analytics, session replay, and crash reports make it a good choice for teams who want a full picture of their user experience.

6. FullStory

Typical users: Product managers, customer success, and support

What is FullStory?

FullStory describes itself as behavioral data analytics platform, which is code for session replay and mobile app analytics, with a side of product analytics.

Supported features

How much does FullStory cost?

FullStory doesn't publish pricing publicly. However, according to Failory, Fullstory will cost you between $250 - $1,000 per month.

Why do companies use FullStory?

According to reviews on G2, companies use FullStory for:

1. Easier collaboration: As an accessible tool for non-technical users, FullStory facilitates collaboration between product, UX, and engineering teams by allowing all teams to access useful, reliable data.

2. Viewing user issues: Support teams use FullStory to replay sessions to understand user hard-to-replicate problems, and identify bugs that need fixing.

3. Improving conversion: FullStory users like to combine funnel insights with replays of user sessions to understand pain points and improve conversion.

Bottom line

For a complete behavioral analytics suite, FullStory is a solid choice. Users wanting more features, like error tracking or A/B testing should switch to other choices.

7. Contentsquare

Typical users: Product managers, ecommerce teams, and marketers

What is Contentsquare?

Contentsquare is an analytics platform that combines heatmaps, customer journey analysis, and frustration scoring to help find and fix issues and friction points.

Supported features

How much does Contentsqaure cost?

Since Contentsquare caters mostly to enterprise companies, its pricing is not available publicly and you need to speak to its sales team to get a quote. According to Capterra, pricing starts at $10,000/month, but no further information is given

Why do companies use Contentsquare?

According to G2 reviews, people like Contentsquare because:

1. Data visualization and analysis: Users like that Contentsquare makes it easy to visualize and analyze large data sets. Advanced features make it easier to identify and fix problems big and small.

2. Improve your UI and UX: Contentsquare makes it easy to understand the impact of various design elements on a screen. This helps developers understand the impact of elements on conversion rate, revenue, and user journeys.

3. Detailed session replays: Users appreciate being able to match quantitative data with qualitative data, giving them a better understanding of how best to solve problems.

Bottom line

From data visualizations, to behavior analysis, and detailed session replays, Contentsquare has a lot to offer, especially for larger companies.

Further reading

• The 9 best mobile app A/B testing tools
• The 7 best mobile app analytics tools
• The best UXCam alternatives & competitors, compared

Best for: Engineering and product teams

What is PostHog?

PostHog (that's us 👋) is an all-in-one platform built to help engineers create better products. It includes product analytics and a whole bunch more, such as A/B testing, feature flags, and session replays.

It's designed for product-minded engineers, growth teams, and product managers who need to move fast and iterate based on reliable, actionable insights.

Key features

• 📈 Product analytics: Custom trends, funnels, user paths, retention analysis, and segment user cohorts. Also, direct SQL querying for power users.

• 📺 Session replays: View exactly how users are using your app. Includes event timelines, console logs, network activity, and 90-day data retention.

• 🧪 A/B tests: Experiment in your app with up to nine test variations and track impact on primary and secondary metrics. Auto-calculate test duration, sample size, and statistical significance.

• 🚩 Feature flags: Rollout features safely with local evaluation (for faster performance), JSON payloads, and instant rollbacks.

• 🏠 Data warehouse: Combine data from all your different sources for easy analysis and comprehensive insights.

How much does PostHog cost?

PostHog has transparent pricing based on usage. It’s free to get started and completely free for the first 1 million analytics events. After this free monthly allowance, pricing starts at $0.000248/event (or $2.48 for 10k events), and events cost progressively less the more you use. You can also set billing limits to ensure you don't get surprise bills.

Why do companies use PostHog?

According to G2 reviews, companies use PostHog because:

1. It's many tools in one: PostHog can replace tools like Mixpanel (analytics), UXCam (session replay), and Statsig (A/B testing and feature flags). This simplifies workflows and ensures all product data is in one place.

2. They need a complete picture of users: PostHog includes every tool necessary to understand users and build better products. This means creating funnels to track conversion, watching replays to see where users get stuck, and testing solutions with A/B tests.

3. Pricing is transparent and scalable: Reviewers appreciate how PostHog's pricing scales as they grow. There's a generous free tier. Companies eligible for PostHog for Startups also get $50k in additional free credits.

Bottom line

For teams looking for all the tools they need to improve their products, PostHog makes for a great choice. This is especially true for startups and scaleups thanks to it having a generous free tier.

2. Mixpanel

Best for: Product managers, desginers and marketing teams

What is Mixpanel?

Mixpanel is one of the most popular product analytics tools on the market. Founded in 2009, in recent years it's deprecated additional features, such as A/B testing, to focus on product analytics alone.

Key features

• 📊 Product analytics: Track user behavior, KPIs, and core metrics with trends, retention, and flows.

• 📝 Collaborative boards: Build analysis in collaborative boards that can include reports, text, videos, and GIFs. Embed these boards in other tools.

• 🚨 Alerts: Get automated notifications when there are anomalies in metrics or if they fall outside a positive or negative range.

• 🔎 Filtered data views: Hide and filter data on a per-team basis to reduce noise and separate data for privacy reasons.

How much does Mixpanel cost?

Pricing starts at $28/month for up to 10k events and scales with events. 1 million events cost $168/month.

Why do companies use Mixpanel?

Looking at G2 reviews, companies choose Mixpanel to:

1. Eliminate the need for data analysts: Mixpanel helps reviewers structure large volumes of data, and make data-driven decisions, reducing their reliance on dedicated data analysts to produce insights.

2. Track and target campaigns: Marketers appreciate the ability to create user segments and target specific users, enabling more personalized campaigns and improved user engagement.

3. Understand user behavior: In common with most analytics tools, Mixpanel's users mostly want to understand user behavior, identify bottlenecks, and monitor core metrics like conversion rates, activation, and retention.

Bottom line

Mixpanel is great for teams who want powerful yet easy to use analytics and don't need any of the extra features that other alternatives on this list provide.

3. UXCam

Best for: Product managers in enterprise companies

What is UXCam?

UXCam is a platform built solely for mobile apps. It includes a full-suite of analytics features as well as session replays and heatmaps.

It's designed to provide deep insights into user behavior, helping product managers understand how users interact with their apps.

Key features

• 📈 Product analytics: Custom trends, funnels, user paths, retention analysis, and segment user cohorts.

• 📺 Session replays: View exactly how users are using your app. Includes event timelines, console logs, network activity, and 90-day data retention.

• 🔥 Heatmaps: See where users tap, where they scroll to, and the areas that get the most attention.

How much does UXCam cost?

UXCam's pricing is sales-driven and completely opaque, so it's not clear how much it costs. The company's sales team will give you a quote after you've spoken to them.

Why do companies use UXCam?

Looking at G2 reviews, companies choose UXCam because:

1. They need insights on user behavior: Insights show what users are doing, where their attention is going, and how they are spending their time. Users love the realizations they get with UXCam's tools.

2. It's easy to use: Reviewers note how easy UXCam is to set up and use. This means they can go from idea to actionable learnings quickly.

3. Heatmaps and scrollmaps: Seeing clicks and scroll depth works well and creates actionable insights for design and product teams. Helps improve UX and page designs.

Bottom line

UXCam is ideal for product managers who need in-depth insights into user behavior. Despite opaque pricing, its ease of use and actionable insights make it a valuable tool.

4. Amplitude

Best for: Product managers, data analysts, marketing teams

What is Amplitude?

Amplitude was one of the original product analytics tools. Many large enterprise customers, like Ford, NBCUniversal, and Walmart rely on it. In recent years, it’s also added A/B testing and feature flags.

Key features

• 📈 Product analytics: Funnel and retention analysis, user paths, behavioral cohorts, custom dashboards, and more.

• 🧪 A/B testing: Test new features on specific targets and analyze with primary, secondary, and counter metrics.

• 👦 Customer data platform: Combine analytics data with third-party tools for data governance, identity resolution, and data federation.

• 🤖 AI insight builder: Generate insights based on natural language requests, like "What is my purchase conversion rate?"

How much does Amplitude cost?

Amplitude is the most expensive platform on this list. Pricing starts at $49/month for up to 1k events. 100k events cost $849/month.

Why do companies use Amplitude?

According to G2 reviews, people like Amplitude because:

1. It's simple to use: Amplitude makes it easy for non-technical users to get insights about their product and make improvements. Amplitude is built for users like product managers and marketers, making it a popular choice for them.

2. It offers built-in A/B testing: Amplitude offers integrated experimentation features. This enables companies to run experiments on existing cohorts, and then analyze the data in a single place.

3. It helps them become data-driven: Amplitude users appreciate it helps them become data-driven. It becomes easy to add data, visualize it, and make decisions, and they can use it as a source of truth thanks to its built-in customer data platform.

Bottom line

Like PostHog, Amplitude is a good option if you value powerful analytics and experimentation in one. If you don't need session replay or heatmaps, it can be a good choice.

5. Firebase (Google analytics)

Best for: Apps already using Firebase's other features

What is Firebase?

Firebase is Google's mobile app development platform. It includes analytics, realtime databases, push notifications, in-app messaging, and more. Its analytics feature is powered by Google Analytics.

Key features

• 📊 Comprehensive analytics: Gain in-depth insights into user behavior and app performance.

• 📈 Real-time data: Access real-time data on user interactions, app crashes, and performance metrics.

• 📲 Push notifications: Engage users with targeted push notifications based on their behavior and preferences.

• 💬 In-app messaging: Send personalized messages to users within your app to guide them through key actions.

How much does Firebase cost?

Since Firebase is powered by Google Analytics, it follows the same pricing. It has a generous free tier that's sufficient for most startups. Larger enterprises with higher data collection and export requirements will need to upgrade to GA360, which starts at $50,000 per year.

A good rule of thumb is that if you're exporting more than 1 million events per day, or querying data with more than 10 million events, you'll need to upgrade to GA360. See Google's breakdown for more.

Why do companies choose Firebase

According to G2 reviews, companies choose Firebase because:

1. It's a comprehensive toolkit: Firebase offers a broad range of tools beyond analytics, such as push notifications, in-app messaging, and Crashlytics, making it an all-in-one tool for app development and management.

2. It's easy to use: Reviewers appreciate how easy Firebase is to use, making app management simple.

3. It integrates with other Google services: Firebase integrates smoothly with other Google services, providing a powerful development and analytics ecosystem.

Bottom line

Firebase analytics is a great choice for companies already leveraging it's other features or any of Google's other products.

6. Appsflyer

Best for: Marketing teams

What is Appsflyer?

Appsflyer is a platform that focuses on marketing analytics and, unlike other platforms on this list, it doesn't include product analytics. Instead, it focuses on attribution to help teams measure, analyze, and optimize their marketing efforts.

Key features

• 🔗 Attribution: Appsflyer provides detailed attribution data, allowing marketers to see which campaigns, channels, and touchpoints are driving installs and conversions.

• 🔍 Marketing analytics: Turn insights on campaign performance into actionable strategies to improve your marketing.

• 💰 ROI measurement: Understand your true mobile marketing ROI, enabling you to optimize your spend effectively.

• 🚨 Fraud protection: Advanced mobile app fraud detection and prevention tools help protect your ad spend.

How much does Appsflyer cost?

Since it focuses on marketing analytics, Appsflyer charges $0.07 per conversion events. Examples of conversion are app installs, in-app purchases, and sign-ups.

Why do companies use Appsflyer?

According to G2 reviews, companies use Appsflyer because:

1. They need comprehensive attribution insights: Appsflyer enables marketers to understand which campaigns and channels are driving the best results.

2. It's easy to use: Reviewers note how easy Appsflyer is to use. This means they can go from idea to actionable learnings quickly.

3. It has a wide range of integrations: Appsflyer can connect to many different marketing tools, making it easier to manage and analyze data from multiple sources.

Bottom line

Appsflyer is the top choice for marketing teams looking to optimize their ad spend and gain deep insights into their campaigns.

7. Countly

Best for: Teams in strict regulatory environments

What is Countly?

Countly is privacy-focused product analytics platform. It emphasizes 100% data ownership and privacy compliance, supporting regulations like GDPR, HIPAA, and CCPA. Users have full control over their data, which can be self-hosted or managed in a private cloud.

Key features

• 📊 Product analytics: Track user behavior, application performance, and key metrics with customizable dashboards.

• 🌐 Privacy compliance: Ensure compliance with GDPR, HIPAA, CCPA, and other privacy regulations.

• 💬 Push notifications: Engage users with personalized push notifications based on their behavior and preferences.

• 🚨 Crash analytics: Monitor and resolve app crashes with detailed reports and real-time alerts.

How much does Countly cost?

Countly has different tiers based on number of active users. To give you an idea of how much it charges per active user, below is a list of its first four tiers:

• Free up to 1k MAU
• $80/month for up to 2k MAU.
• $190/month for up to 5k MAU.
• $350/month for up to 10k MAU.

Why do companies use Countly?

Based on G2 reviews, companies choose Countly because:

1. It's focused on privacy and compliance: Countly ensures that companies can adhere to strict regulatory requirements.

2. It's many tools in one: From push notifications to A/B testing to crash analytics, Countly provides many features for building great apps.

3. You have full data ownership: With options for self-hosting and private cloud, Countly enables companies to manage their data securely.

Countly is a great choice if you're building a strict regulatory environment. Many users complain about its bad user interface though.

Is PostHog right for you?

Here's the (short) sales pitch.

We're biased, obviously, but we think you'll love PostHog if:

• You value transparency (we're open source and open core)
• You want tools to ship, track, and analyze new features – like A/B testing, feature flags, and session replays
• You want try before you buy (we're self-serve with a generous free tier)

Check out our product pages and read our docs to learn more.

Further reading

• The 9 best mobile app A/B testing tools
• 7 best session replay tools for mobile apps (iOS & Android)
• The 9 best GDPR-compliant analytics tools

1. We're an all-in-one platform

PostHog combines product analytics with web analytics, session replay, feature flags, A/B testing, surveys, and a data warehouse into one tightly integrated platform. Everything you need from a single app with a single contract. A genuine single source of truth for your product and customer data.

2. It's built for developers

This means you get support from the engineers who actually build the product, extensively documented APIs, and a SQL query builder, so you can analyze data how you want. We're open source, so you can inspect our source code and vote on what we should build via our public roadmap.

3. Transparent and cheap pricing (forever)

We default to charging as little as possible while still making a profit – we also have a generous free tier on all our products. We can do this because we're efficient. We don't splurge on massive sales teams, or D-list comedians to host an annual convention you'll never attend. Want to know how much we'll charge? See our pricing calculator.

Comparing PostHog and Mixpanel

Platform

PostHog can replace multiple tools, such as Mixpanel for product analytics, Hotjar for session replay and surveys, and LaunchDarkly for A/B tests and feature flags.

*You can't run A/B tests using Mixpanel, only analyse the results of tests created using third-party integrations.

Product analytics

There are two key differences here:

1. PostHog supports autocapture, which makes it much faster to implement as you don't have to manually instrument every event. Don't want autocapture? Just turn it off and rely on one of our many SDKs – we offer the best of both worlds.

2. PostHog has a SQL-based query builder for power users, which enables you to generate custom data analyses using PostHog event data. In future, our query builder will let you analyze data as you would in a business intelligence tool, like Metabase, inside PostHog.

Feature flags

Feature flags make it easy to roll out features to specific users or groups, and safely test in production. Our feature flags are also tightly integrated with other tools, meaning you can target session replays, surveys and other features using existing feature flags. See benefits of feature flags for more.

A/B testing

PostHog's A/B testing are billed with feature flags. You can't run A/B tests using Mixpanel, but you can analyze the results of experiments created using third-party A/B testing tools.

Session replay

Replays enable you watch how users experience your app, diagnose issues, improve support, and understand real user behavior in a way raw data can't. You get 5,000 recordings for free every month, so you'll pay nothing if you remain below this limit.

Note: Mixpanel's session replay tool is currently a closed, invite-only beta for customers on its Enterprise plan.

Surveys

Surveys are great for tracking customer satisfaction scores like Net Promoter Score (NPS), customer satisfaction (CSAT), and customer effort score (CES). We include survey templates for all these and more. Read our guide comparing NPS, CSAT, and CES to see which survey type you should use and when.

Data integrations

PostHog and Mixpanel diverge here:

1. PostHog has a built-in data warehouse, so you can import and query data directly from our other sources like Stripe and Zendesk, or from an existing warehouse.

2. Mixpanel relies on data warehouse connectors to combine customer and product data. This means you don't import data directly from third-party tools. You need to get them into a third-party warehouse and then import that data into Mixpanel for analysis.

See our docs for full lists of destinations and data warehouse connectors.

Security and compliance

Frequently asked questions

Can I migrate my data from Mixpanel to PostHog?

Yes. See our Mixpanel to PostHog migration guide.

Can PostHog replace Google Analytics?

Yes. PostHog can replace Google Analytics for many use cases – our marketing team uses PostHog, for example. You can also integrate PostHog into your website using Google Tag Manager.

See our comparison of PostHog and Google Analytics 4 and An intro to PostHog for Google Analytics users for more.

How much does PostHog cost?

Every PostHog user gets a generous amount of free usage each month:

You'll never pay anything if you stay within these limits and you can set billing limits to avoid surprise bills.

A full breakdown and pricing calculator is on our pricing page.

Can I use PostHog with a CDP? (Segment, Rudderstack, etc.)

Yes. See Using PostHog with a CDP in our docs.

What about ad blockers?

We recommend all users deploy a reverse proxy, which enables you send events to PostHog Cloud using your own domain. Events sent from your own domain and are less likely to be intercepted by tracking blockers, ensuring you capture the best data possible.

We make this super easy by offering a managed reverse proxy – a no code solution that takes minutes to setup. We also have reverse proxy setup guides for AWS Cloudfront, Caddy, Cloudflare, Netlify, Vercel, and more in our docs.

How do PostHog and Mixpanel compare to Amplitude?

Amplitude is closer to Mixpanel than PostHog – it doesn't support features like session replays or surveys, and doesn't offer autocapture. Read our PostHog vs Amplitude comparison for information.

Septemeber 2024 Update: We built our own, dedicated web analytics tool. Now PostHog is an even better alternative to Google Analytics 4!

How is PostHog different?

1. We're an all-in-one platform

PostHog combines product analytics and web analytics with session replay, feature flags, A/B testing, surveys, and a baked in data warehouse into one tightly integrated platform. Everything you need from a single app with a single contract. A genuine single source of truth for your product and customer data.

2. It's built for developers

This means you get support from the engineers who actually build the product, extensively documented APIs, and a SQL query builder, so you can analyze data how you want. We're open source, so you can inspect our source code and vote on what we should build via our public roadmap.

3. Transparent and cheap pricing (forever)

We default to charging as little as possible while still making a profit – we also have a generous free tier on all our products. We can do this because we're efficient. We don't splurge on massive sales teams, or D-list comedians to host an annual convention you'll never attend. Want to know how much we'll charge? See our pricing calculator.

Comparing PostHog and Google Analytics

Analytics

Features like group analytics, which enables you analyze behavior at an account of company level, make PostHog ideal for B2B SaaS products.

You can also go deeper on user behavior by utilizing heatmaps, scrollmaps, formulas, and the custom SQL insights.

Platform

When you choose PostHog, you get more than analytics.

Integrations

It's hard to import data into Google Analytics because:

1. The data type and format you're allowed to import is restrictive.
2. You either need to constantly upload CSV files manually, or set up an SFTP server to automatically do this for you.

In contrast, PostHog is built to be your single source of truth, so it's simple to import data from other sources using our built-in data warehouse, or send PostHog data to other tools using our realtime destinations.

Below is a comparison of some of the most popular apps – see our data pipeline and warehouse docs for a complete list of integrations.

Security and compliance

PostHog makes GDPR compliance easy by letting you choose where your data is hosted: EU or US. Google also offers various privacy controls, but you can't choose where your data is stored.

Frequently asked questions

Does PostHog have reports, dimensions, and other GA4 features?

Yes, PostHog has much of the same functionality as Google Analytics, but we different terminology. Here’s a quick comparison of the two:

See our guide to PostHog for Google Analytics users for more help on making the switch.

How much do PostHog and Google Analytics cost?

Google Analytics has a free tier that's sufficient for most startups. Larger enterprises with higher data collection and export requirements will need to upgrade to GA360, which starts at $50,000 per year.

A good rule of thumb is that if you're exporting more than 1 million events per day, or querying data with more than 10 million events, you'll need to upgrade to GA360. See Google's breakdown for more.

PostHog has transparent pricing based on the usage. It's free to get started and completely free for the first 1 million analytics events. After this free monthly allowance you're charged based on usage – we charge progressively less per event the more you use.

We recommend trying our pricing calculator to estimate your costs.

Does PostHog offer a free trial?

With PostHog, it's free to get started, with a generous monthly allowance of events, replays, and feature flag requests. If you’re within this allowance, PostHog is free to use forever.

Can I migrate from Google Analytics to PostHog?

Yes, see our Google Analytics to PostHog migration guide for more details.

Does PostHog block bots by default?

Yes. See the full blocklist in our docs.

What about ad blockers?

We recommend all users deploy a reverse proxy, which enables you send events to PostHog Cloud using your own domain.

Events sent from your own domain and are less likely to be intercepted by tracking blockers, ensuring you capture the best data possible. We have reverse proxy setup guides for AWS Cloudfront, Caddy, Cloudflare, Netlify, Vercel, and more, in our docs.

Are there discounts for non-profits and startups?

Yes, PostHog offers both. Non-profit organizations can contact our team and are usually eligible for a 50% discount, while startups can sign up for $50,000 of free credit (and a host of other perks) in the PostHog for Startups program.

This year, we headed to Mykonos – not to rave, but to code. And to enjoy a cocktail or two in the sun, of course.

Our all-company offsites are a mix of socializing, group activities, strategic sessions, Post-its, workshops, more Post-its, and, the star of the show, the annual PostHog hackathon.

Everyone in the company has to pitch a couple of ideas, then we all vote on our favorites, assemble teams, and have a little over a day to go from pitch to demo.

In some cases, our hackathon projects are ready to ship right away. Some have even become core parts of our product – session replay started as a hackathon project, as did our data warehouse beta.

Here’s what we built in Mykonos.

Our own programming language

What’s cooler than having your own query language? Having your own general purpose programming language, of course! That’s literally what we built: PostHog’s first ever programming language, Hog. (Earlier versions of this name included Hög and Höge, but turns out we're anti-umlaut.)

Things moved fast the week after Mykonos: we split up the existing product analytics team and built a new team to productize Hog. We plan to use Hog to build our CDP and messaging products, and might even pivot the entire company around it.

Why? Because, in the past, we’ve sometimes struggled to build UI fast enough for users to take advantage of new products that are ready to ship. Hog will mean anyone can drop down into hacker mode and get things done. We move fast. You move fast. Everyone is happy.

It was just another PostHog hackathon, no big deal. Stay tuned for the aftermath.

RealTimeHog 3000

Seeing people using your product live boosts dopamine levels. Probably. PostHog does that, but right now we keep you waiting a bit.

Before a PostHog event is available for querying, it goes through our ingestion pipeline, where users are identified and the events themselves enriched. The process takes seconds, rarely a minute – a delay imperceptible in analytical queries, but a dopamine decrease in the live view.

The solution? Introducing RealTimeHog 3000, a livestream service powering our first truly real-time view of user activity. The process is simple:

1. Consume the raw events from the same Kafka topic as our ingestion service.
2. Stream them to you ASAP using server-sent events.

The lack of person data is a fair trade-off, because it doesn’t get more live than this. User activity appears within milliseconds of happening on the other side of the world.

Speaking of data being processed live, perhaps at some point you’ve wondered if PostHog actually scales, or just talks the talk.

To dispel any doubts, we developed one extra feature: an anonymized stream of all events being captured globally, with only the geolocation included.

Millions of events per minute, and they look great on a 3D globe on PostHog’s website, where each event is an arc from the user’s location to that of the relevant data center. Global scale, visualized for your pleasure.

MykoLogs

MykoLogs is a logging product that integrates with the existing PostHog SDKs, bringing backend logs straight into a shiny new product on PostHog.

The best part? You can link backend logs to session recordings through the user’s session ID – letting you debug what was happening on the backend during your user's API requests. MykoLogs plays nicely with all other PostHog products, meaning logs and session replay are now BFFs. Debugging has never been this breezy!

It's internal-only for now, but could one day be made public.

The presidential briefing

As PostHog grows as a company, keeping track of everything that’s happening will get harder. Yes, we write everything down but that creates a lot of reading, and a lot of noise. The solution? An AI-generated briefing, tailored to each individual team member and their interests.

The presidential briefing was built by scraping PRs and issues from GitHub, along with Slack messages, and then training an LLM to understand what’s interesting and important. The bot then produces a pithy briefing that removes the noise and gives people just the information they need.

While just a proof-of-concept for now, if we were to ship it we’d want to add more data sources and build it using Llama to avoid the need to send any data to external services.

10x terms

Are you fed up with lawyers making everything so hard to understand? Are you fed up with those nerds in Brussels making us sign DPAs for everything? Not anymore! On PostHog.com, we’ve made all the legal stuff fun – and kept the lawyers happy.

First up, we summarized our terms and privacy policy in plain English. You can still read the long, legal-y version, but it’s now way easier to understand what it actually means. And we didn’t stop there.

Then we took on our data processing agreement, or DPA, to create a generator that makes this hugely exciting task even more fun. You can quickly populate your own form, select the data region, and, if you want, add some pizazz with fairy tale or Taylor Swift mode. DPA? Try DPYAY!

ZenHog

Our support flow currently uses Zendesk and that goes through email. This causes three problems:

1. Emails sometimes bounce
2. There are long delays in checking emails
3. It’s really clunky to add other team members via CC.

It’s not the most optimal flow. So, imagine if instead users could view and respond to their open support tickets without ever having to leave PostHog? Yhat's exactly what we built.

But that's not all, if you're using Zendesk then you could, in the future, add this view to your customer-facing website with just a few clicks. Like the sound of this project? Got to our public roadmap, search for 'Customer support product', and vote for it.

The referral scheme

Everyone loves a pyramid scheme, right?!

Wait, no, we mean a referral program. Ben, Raquel, and Joe worked together to build a referral product right into PostHog. This means we can offer sweet merch, platform credits, good vibes, and other things to loyal PostHog users who lure in their friends and family.

What’s more, as the referral program product is built right into PostHog, you can build your own for your customers. The system is hooked up to Zapier to automate the process for redeeming codes making the whole thing a doddle.

It’s not shipped yet, but we’re close.

Managed reverse proxy

Everyone loves ad-blockers. But, for a lot of our customers, they stop data from reaching PostHog.

You can already deploy a reverse proxy to PostHog Cloud to get around this, but it’s a somewhat convoluted process that requires you to jump through 16 hoops and login to AWS. Our docs on this are great, but Frank decided to build a better solution.

During the hackathon, he built the reverse proxy functionality right into PostHog. The option is tucked away in the PostHog settings.

Simply add in any domain you control and the system will spit out a CNAME that you then need to set in your DNS provider. Wait a few seconds for the update to happen and voila, the reverse proxy is live.

A/B TestHog

Want to know how to improve your website but don’t know where to start? You need A/B TestHog. Enter a website URL, click ‘Analyze’ and an ingenious generative AI system will give you a bunch of recommendations for what A/B tests you might run to take your website to the next level.

These are all expertly authored by an AI, and include the goal metrics, secondary metrics, and guardrail metrics and detailed instructions of what to change for your test.

HERMES

At PostHog, we love speaking to our users. Maybe a bit too much. Right now, our master customer interviews doc is 382 pages long and contains almost 200 user interviews. It’s a great resource, but it’s getting a bit unwieldy.

But we've now entered a bold new era of feedback management at PostHog thanks to HERMES, or... Holistic Evaluation Repository for Managing Enhancements and Suggestions.

This is effectively a database of user interviews, showing who was interviewed, who they work for, what they do, how much they pay for PostHog, the products they talked to us about, and an AI-generated summary of our user interview notes. The database is searchable and you can easily add new interviews in a couple of clicks.

The database is linked right into PostHog, making it easy to see associated user and organization profiles. It’s also hooked into Vitaly, our customer success tool, to automatically pull in more customer and business information.

HERMES also uses ChatGPT to generate a summary of the features requested during the interview based on the human-authored interview notes. This makes it easy to share actionable feedback from users directly with the product team responsible for that feature.

As part of the project, we also revamped our system for categorizing and tracking feature requests from customers, making it easier for us to see what people want and prioritize the most important product work.

Data crunching

PostHog crunches a lot of data, especially on very complext queries. To help users better understand the hard work we’re doing when they make a query, we built a loading bar that includes live data on how much data we're crunching (database rows and data volume) and CPU usage we're deploying to generate an answer for you.

If you feel the need for a quick distraction while you wait, you could also check out Hedgehog mode 2.0, which also shipped recently.

CLI

We build products for engineers, so there’s nothing better than bringing PostHog closer to their natural environment: the terminal.

The PostHog CLI is a command line that allows users to do a few things that are normally tucked away in the PostHog app: creating, reading, updating, deleting, and enabling or disabling feature flags, for example. The PostHog CLI authentication flow is also seamless as it spins up a new browser and allows you to log in with your SSO instead of copying and pasting tokens manually.

In the future, the CLI could be expanded with more features such as creating surveys, and events, installing SDKs automatically, uploading debug symbols, using the CLI as a package, or even as a GitHub action. And, even better, you could do all of that with natural language, no need to memorize all commands by heart.

It's an internal tool for now, but could be made public if people ask for it enough...

4 years at PostHog

We’re still a young company, but some of our wonderful team members have now been with us for four years or more. So we want to celebrate them. Coua and Kendal came up with a great anniversary gift scheme, meaning our longest-serving colleagues get something special to celebrate.

This year, Marius, Eric, James G, Lottie, Charles, and Michael all celebrate four years at PostHog and will get to pick between a fancy luggage set (handy for traveling to all-team offsites and other PostHog meet-ups), or a James Hawkins-approved coffee machine.

Forbidden secret project

We have a Slack channel called #do-more-weird for odd, fun ideas, and this hackathon project belongs there.

Head to the careers page and click on James Hawkins' face for an inspirational surprise...

Passed in 1996, HIPAA (Health Insurance Portability and Accountability Act) defines the legal requirements for securing and handling health information, and the severe penalties for failing to do so.

Data protected under HIPAA is called Protected Health Information (PHI), or ePHI if it is digitized. It includes any data that can be used to identify the past, current or future health status of an individual.

This includes test results and diagnoses, but also birthdays, ethnicity, gender and other information. Even an IP address can be considered ePHI under HIPAA.

While similar in some respects to the EU's General Data Protection Regulation (GDPR), HIPAA applies specifically to companies handling the PHI of US-based customers. Companies that also need to comply with the GDPR should see our guide to GDPR-compliant analytics.

There are two ways to be HIPAA-compliant while using analytics tools:

1. Self-host your analytics, so data remains totally within your control.
2. Sign a Business Associate Agreement (BAA) with a third-party analytics tool.

What is a Business Associate Agreement (BAA)?

Some services enable HIPAA compliance through the creation of a Business Associate Agreement (US Department of Health). This is a contract with a service provider to ensure that they are jointly compliant and liable for services they provide.

It's worth noting that, because BAAs expose third-parties to increased risk and scrutiny, they are often an expensive option and/or require users to purchase a higher tier of license.

Some analytics tools, such as Google Analytics, don't offer BAAs and are therefore not HIPAA-compliant.

The best HIPAA-compliant analytics tools

1. PostHog

Overview

• Product analytics: ✔
• Web analytics: ✔
• Session replay: ✔
• Feature flags: ✔
• A/B testing: ✔
• Surveys: ✔
• Customer data platform: ✔

PostHog is an open-source platform that combines product analytics, web analytics, session replay, feature flags, A/B testing, and user surveys into one platform. It gives you every tool you need to understand user behavior and, unlike typical analytics tools that rely on third-party integrations, all these tools work together seamlessly.

Being an all-in-one platform has two further benefits:

1. PostHog can replace multiple individual products – e.g. Mixpanel for product analytics, LaunchDarkly for feature management, etc.
2. You only need to sign one BAA to get all these tools, reducing legal complexity and risk.

PostHog and HIPAA compliance

• Self-hostable: ✔
• BAA available: ✔

A BAA is available on PostHog's Teams plan, which also includes priority support and generous free usage limits for all tools – e.g. 1 million free analytics events every month. You can also self-host the open-source edition of PostHog, but this isn't recommended as it's provided without guarantee or support.

2. Mixpanel

Overview

• Product analytics: ✔
• Web analytics: ✔
• Session replay: ✖
• Feature flags: ✖
• A/B testing: ✖
• Surveys: ✖
• Customer data platform: ✖

Mixpanel is a pure product analytics tool for analyzing user funnels and behavior. It doesn't offer built-in tools for things like session replay, feature management, A/B testing, or user surveys, but it does have extensive integrations with third-party tools for these.

Mixpanel and HIPAA compliance

• Self-hostable: ✖
• BAA available: ✔

A BAA is available on Mixpanel's Growth plan, which starts at $20 per month for 10,000 events. You'll need to sign separate BAAs with any other analytics you wish to integrate with Mixpanel.

See also: The most popular Mixpanel alternatives

3. Countly

Overview

• Product analytics: ✔
• Web analytics: ✔
• Session replay: ✖
• Feature flags: ✖
• A/B testing: ✔
• Surveys: ✔
• Customer data platform: ✖

Countly is an analytics platform for mobile, web, and desktop applications that also offers add-ons for remote configuration, A/B testing, and user surveys. Support for app crash and error reports, and push notifications, makes it particularly well-suited to mobile app analytics.

Countly and HIPAA compliance

• Self-hostable: ✔
• BAA available: ✖

Countly doesn't offer a BAA for HIPAA compliance on its hosted cloud, but it does offer the option to either:

1. Self-host the product on your own.
2. Deploy it to a private cloud managed by Countly.

This makes Countly a good option if you'd prefer to self-host your analytics.

4. Freshpaint

• Product analytics: ✖
• Web analytics: ✖
• Session replay: ✖
• Feature flags: ✖
• A/B testing: ✖
• Surveys: ✖
• Customer data platform: ✔

Overview

Freshpaint isn't an analytics tool per se, it's more of an analytics event tracker and customer data platform (CDP) that's specifically designed for healthcare companies.

Freshpaint sits between data sources (e.g data warehouses) and third-party data destinations and ensures no PHI is passed between them. This means you can continue to use non-HIPAA compliant tools, such as Google Analytics, safe in the knowledge you're not accidentally passing PHI into them.

Freshpaint and HIPAA compliance

• Self-hostable: ✖
• BAA available: ✔

Freshpaint is a cloud-only product specifically designed for healthcare companies, so offers a BAA for HIPAA compliance.

5. Piwik PRO

• Product analytics: ✖
• Web analytics: ✔
• Session replay: ✖
• Feature flags: ✖
• A/B testing: ✖
• Surveys: ✖
• Customer data platform: ✔

Overview

Piwik PRO is a commercial analytics and customer data platform spun out of the open-source analytics tool, Matomo. As such, it's more a web analytics tool than other options in this list, though you can use it on mobile and web apps. Because it's based on Europe, Piwik PRO is popular among companies also seeking GDPR compliance – it has a built-in compliance manager to assist with this, too.

PiwikPRO and HIPAA compliance

• Self-hostable: ✔
• BAA available: ✔

PiwikPRO offers HIPAA compliance as part of its PRO Enterprise plan, either by signing a BAA or by self-hosting, giving you maximum flexibility.

6. Amplitude

Overview

• Product analytics: ✔
• Web analytics: ✖
• Session replay: ✔
• Feature flags: ✔
• A/B testing: ✔
• Surveys: ✖
• Customer data platform: ✔

Amplitude sits somewhere between PostHog and Mixpanel. It's a product analytics tool at its core, but also has extra features such as session replay, feature flags, and A/B testing. It also has anomaly detection, which will automatically flag when certain metrics fall outside expected trends, and creating insights based on natural language questions like "signups in the last 30 days."

Amplitude and HIPAA compliance

• Self-hostable: ✖
• BAA available: ✔

Amplitude offers a BAA, but doesn't stipulate the minimum terms for signing one on its website. You can also use its product analytics tool on top of a Snowflake data warehouse, which may be an option for HIPAA compliance if you're already storing analytics data in Snowflake.

See also: The most popular Amplitude alternatives

7. Heap

• Product analytics: ✔
• Web analytics: ✖
• Session replay: ✔
• Feature flags: ✖
• A/B testing: ✖
• Surveys: ✖
• Customer data platform: ✖

Overview

Heap is a product analytics tool that, like PostHog, Mixpanel and Amplitude, helps you understand user behavior using trends and funnel analysis. It also offers a basic session replay tool. Like Amplitude, it caters more to non-technical users, such as marketing and e-commerce teams focused on optimizing conversions.

Heap and HIPAA compliance

• Self-hostable: ✖
• BAA available: ✔

Heap offers a BAA, but only on its Pro and Premier plans, not on its self-serve Growth plan. It's not available as a self-hosted product.

See also: The most popular Heap alternatives

FAQ

Who does HIPAA apply to?

HIPAA applies to "covered entities," such as healthcare providers who transmit any health information in electronic form, health plans, and healthcare clearinghouses. Mobile apps fall under HIPAA if they store protected health information (PHI), and share it with any covered entity.

HIPAA also applies to "business associates," which, according to the US Department of Health and Human Services, are "a subcontractor that creates, receives, maintains, or transmits protected health information on behalf of another business associate."

Under HIPAA, the analytics tools in this guide would all be considered business associates.

What is PHI (Protected Health Information)?

Protected Health Information (PHI) is any information about health status, provision of healthcare, or payment for healthcare that can be linked to an individual.

This includes medical records, laboratory results, billing information, and any other information that identifies an individual and relates to their past, present, or future physical or mental health condition, treatment, or payment for healthcare services.

Is self-hosting analytics better than signing a BAA?

There's no objective correct answer here. In theory, self-hosting is preferable as it means you don't share any data with third-parties (business associates), and thus you don't need to sign a BAA.

But self-hosting also presents additional risks. You're wholly liable for ensuring your analytics infrastructure is secure, which can be challenging if you don't have the internal expertise to manage this. If this is the case, it may be better to rely on a HIPAA-compliant business associate who has experience hosting analytics at scale.

1. They're popular A/B testing tools
2. None of them are HIPAA-compliant

And, while it's tempting to live without A/B testing for your healthcare product, doing so is like trying to navigate an ocean by sailing roughly in the correct direction. You'll probably arrive somewhere, but it won't be where you intended.

What you need for HIPAA compliance

You need to comply the Privacy Rule and the Security Rule. Breaching either can result in hefty financial penalties, but for the sake of this guide we're mostly interested in how the Privacy Rule impacts analytics and A/B testing.

There are three ways to comply with the Privacy rule when adopting analytics and testing tools:

1. Anonymize all PHI and identifiers: There are two so-called "De-identification Standards" – "Expert Determination," where an expert verifies that data isn't personally identifiable, and "Safe Harbor" where all 18 types of identifier are removed. The former is preferable simply because applying the Safe Harbor approach can render data effectively useless for analytical purposes.

2. Sign a BAA with a third-party tool: You must sign a Business Associate Agreement (BAA) with any third-party platform that handles your protected health information (PHI). This can mean signing multiple agreements, though, such as one with your analytical partner, but also any tools you use for importing and exporting data from your data warehouse.

3. Self-host and keep control of all your data: The less common is to self-host tools for analytics and experimentation on your own infrastructure. This reduces the number of BAAs and general legal wrangling needed to generate user insights. The only downside is you'll need the expertise to manage self-hosted instances, or third-party support to do so, and you are wholly liable for any security breaches.

These are the broad principles, but please consult an expert before making any final decision on how to implement tools in compliance with HIPAA.

The best HIPAA-compliant A/B testing tools

1. PostHog

Features

• Product analytics: ✔
• Web analytics: ✔
• Session replay: ✔
• Feature flags: ✔
• A/B testing: ✔
• Surveys: ✔
• Self-hostable: ✔
• BAA available: ✔

Summary

PostHog is an open source all-in-one platform that combines A/B testing with product analytics, session replay, feature management, and user surveys – everything you need to understand user behavior. All these tools are seamlessly integrated and, because you get everything in one, you only need to sign one BAA for all your analytics needs.

PostHog offers a BAA on its Teams plan, which starts at $450 and includes generous monthly free allowances, such as 1 million analytics events every month. You can also self-host the open-source edition for free, though this isn't recommended as it's provided without support or guarantee.

2. Kameleoon

Features

• Product analytics: ✖
• Web analytics: ✖
• Session replay: ✖
• Feature flags: ✔
• A/B testing: ✔
• Surveys: ✖
• Self-hostable: ✖
• BAA available: ✔

Summary

Kameleoon is an A/B testing and personalization platform. It supports A/B and multivariate testing, and feature flags for managing the rollout of new features and running tests. In addition to testing, it has a real-time personalization engine that's particularly useful for e-commerce. It doesn't have any deeper analytics features, so you'll need to run it alongside another HIPAA-compliant analytics tool to gather deeper user behavior data.

Kameleoon doesn't publish pricing publicly, but conversion optimization consultants BrillMark reports pricing starts at $35,000 per year and scales based on traffic volume, making it a premium option.

3. VWO

Features

• Product analytics: ✖
• Web analytics: ✔
• Session replay: ✔
• Feature flags: ✔
• A/B testing: ✔
• Surveys: ✔
• Self-hostable: ✖
• BAA available: ✔

Summary

VWO is best known as an A/B testing platform for e-commerce websites and mobile apps, though it also offers basic session replay and analytics tools as part of its myriad pricing tiers. A/B testing features include support for multi-armed bandit, a visual editor, and advanced targeting options, such as targeting based on screen resolution.

Unlike most tools in this list, VWO charges separately for website and mobile apps based on monthly tracked users (MTUs) for both, so it can get expensive quickly if you need both. VWO offers a BAA when you pay for the Security Plus Add-on ($529 per month).

4. LaunchDarkly

Features

• Product analytics: ✖
• Web analytics: ✖
• Session replay: ✖
• Feature flags: ✔
• A/B testing: ✔
• Surveys: ✖
• Self-hostable: ✖
• BAA available: ✔

Summary

LaunchDarkly is primarily a feature management platform for controlling what users see and when, and managing the rollout of new features. However, it also offers an experimentation suite, albeit as a paid add-on.

As a tool designed for engineers, LaunchDarkly supports running experiments on the front and back end. This enables engineers to run experiments to measure the performance impact of API and infrastructure changes, for example.

FAQ

Who does HIPAA apply to?

HIPAA applies to "covered entities," such as healthcare providers who transmit any health information in electronic form, health plans, and healthcare clearinghouses. Mobile apps fall under HIPAA if they store protected health information (PHI), and share it with any covered entity.

HIPAA also applies to "business associates," which, according to the US Department of Health and Human Services, are "a subcontractor that creates, receives, maintains, or transmits protected health information on behalf of another business associate."

Under HIPAA, the A/B testing tools in this guide would all be considered business associates.

What is PHI (Protected Health Information)?

Protected Health Information (PHI) is any information about health status, provision of healthcare, or payment for healthcare that can be linked to an individual.

This includes medical records, laboratory results, billing information, and any other information that identifies an individual and relates to their past, present, or future physical or mental health condition, treatment, or payment for healthcare services.

Is self-hosting better than signing a BAA?

There's no objective correct answer here. In theory, self-hosting is preferable as it means you don't share any data with third-parties (business associates), and thus you don't need to sign a BAA.

But self-hosting also presents additional risks. You're wholly liable for ensuring your A/B testing infrastructure is secure, which can be challenging if you don't have the internal expertise to manage this. If this is the case, it may be better to rely on a HIPAA-compliant business associate who has experience hosting analytics at scale.

Updated July 26, 2024: Added information about Onboarding 6.0

Marketers believe onboarding emails are a powerful tool for driving engagement. Everyone else thinks they're annoying and ineffective. The truth, as always, is a little more complicated.

People, and developers in particular, hate onboarding campaigns because the bad ones are so memorable. And how would you know what "good" looks like without signing up and risking a flood of tedious outreach emails from sales reps?

We don’t do those sort of outreach emails. This is what we do instead, and how we got there over 5 major iterations. We think it works and we've included performance data so you can judge for yourself.

Onboarding 1.0: The one where it started costing lots

Back in 2022, our onboarding flow was powered by Mailchimp and was very simple. In fact, nobody actually knew what it looked like. So, I mapped it in an RFC.

We didn’t do welcome emails for new users, and even our monthly changelog emails were only sent to 660 opted-in users. I decided to change this and imported 30,000+ users into that email list.

The move was covered by our T&Cs and, because it was one email a month about actual product updates, we figured nobody would mind. The good news was that we were right and didn't get a single complaint and only a handful of unsubscribes.

The bad news is that it sent our Mailchimp costs skyrocketing. This is because Mailchimp's pricing is based partly on how many emails you send each month, and it has low limits that make it entirely unsuited for a complex onboarding campaign. We started searching for a platform with more predictable and scalable pricing.

Onboarding 2.0: The one where we chose a new email tool

Towards the end of 2022, we finally moved off of Mailchimp. We looked at tools such as Braze, Hubspot, and Customer.io, ultimately choosing the latter because it had more transparent pricing and we had an existing app for pushing data to it. Signing the deal was arduous – Customer.io wanted almost ten separate meetings before they took our money, but at least they didn't send any awful sales outreach emails!

I’d observed from other companies that we needed to at least add a welcome email for new users, and I wanted to add checks after 24, 96, and 168 hours to see if users had ingested events. If yes, we’d send them some basic usage advice. If no, we’d offer them help. This was the very least I felt we could do.

This didn’t perform as well as you’d think, however, mainly because the ingestion checks were unreliable. Users kept complaining about getting the wrong emails, so we threw this all out and moved to a new plan.

💡 Tip: Something we learned for our welcome email was not to send it immediately after sign-up. The email is meant to give users initial direction, but first they need to set up ingestion – we added a one hour delay so that users can get data to work with first.

Onboarding 3.0: The one where we started tracking data

By Q3 2023, we were adding more and more new products to PostHog. We wanted to feature these in our flow, but we were also worried about becoming spammy. Thankfully, now that we had Customer.io in place, we were able to judge this based on the data.

Our conversion event was a user log-in within one week of opening an email. We hadn’t yet thought of anything better.

• Open rate: 56%
• Click-through-rate: 7.4%
• Conversion rate: 3%
• Unsubscribe rate: 0.6%

Based on this high open rate and low unsubscribe rate we felt that users didn't mind getting the emails, and the click-through rate suggested they liked what they read. So, we continued adding more messages.

What does good look like for an email onboarding flow? Benchmarks vary wildly by industry, implementation, and product. As a rule though, the guideline benchmarks I use from my experience in other startups are:

• Open rate: Anything above 40% is OK. 50% is the goal.
• Click-through-rate: Anything above 4% is OK. 6% is the goal.
• Conversion rate: Anything above 3% is OK. 5% is the goal.
• Unsubscribe rate: Anything below 1% is OK. 0% is the goal.

Onboarding 3.1: The one where we added personalization

Data is important, but I strongly believe it doesn’t tell the whole story. You also need feedback and intuition. Most of the feedback on our emails came in 3.1, when we started sending personalized recommendations to users from my personal email.

How this works is simple: when a user signs up, they can optionally tell us what their role is. If their role_at_organization = engineer we trigger an email from me telling them about our Product for Engineers newsletter.

This email performed very well, with a 68% open rate and a 16% CTR. Two emails into the flow and developers were already opting in to more emails!

Best of all, though, because it was personalized and came from me directly, it also earned a steady trickle of replies. I responded in kind and was able to feed in further improvements, including adding another option for role_at_organization = founder.

Onboarding 3.3: The one where we added experiments

In 3.3, we started running a series of tests and experiments. We removed the founder targeted email, which invited teams to join PostHog for Startups, because it was too successful and cannibalised revenue.

Most of the tests we ran proved inconclusive, or outright failed. Subtle changes to subject lines or body copy had no noticeable impact. We learned that users did slightly prefer Lottie’s hedgehog art over screenshots of the app, but the difference was a negligible 1-2% in CTR.

One successful experiment we did roll out was adding more personalized emails for non-technical users. If role_at_organization = marketing, sales or product and a user had not created an action within 12 hours, I’d send them another personal tip. These emails averaged a steady 6% CTR and 54% open rate, so we kept them in, but interestingly I didn't see many replies on these messages.

Onboarding 4.0: The one where we filtered by ICP

4.0 was a massive step up in complexity as a result of our growing number of features. By late 2023 the flow had 28 separate emails, of which users would get a maximum of 6 over several weeks. We still worried this would be too many, but the low (0.4%) unsubscribe rate suggested otherwise.

By this point we’d also changed our conversion event to mean a user enabling billing for a product within one week of opening an email - a far better indicator of if the emails are delivering value than simply measuring logins.

• Open rate: 52%
• Click-through-rate: 4%
• Unsubscribe rate: 0.4%
• Conversion rate: 6%

Around this point we also started filtering out teams in our ideal customer profile (ICP) using a scoring system implemented in Hubspot. If a user scored over a certain threshold they’d be moved out of the usual flow so our customer success team could reach out directly. We’ve since removed this check because it felt too salesy and didn’t perform well, however.

Onboarding 5.0: The gosh-thats-complicated one

The next version was where things got out of hand. It contained over 53 emails and 38 timed events and was, quite simply, a pain in the ass to explain to everyone internally.

The goal of this flow was to focus on improving activation, and build on some parallel improvements to our in-app onboarding flow. When a user signed up to PostHog we asked them to select a product to setup first: analytics, replays, flags, or surveys. Onboarding 5.0 builds off that selection by encouraging users to use that product, and then showing how other features can combine with it.

The hope was that this would improve in-app activation, which we roughly define as:

• Product analytics: Events ingested, 1 insight created, 1 dashboard created
• Session replay: 5 recordings viewed
• Feature flags: 1 feature flag created
• Surveys: 1 survey launched

To avoid annoying users, we added a time block based on our activation criteria – if a user hasn’t activated after two emails, we stop sending them emails for six weeks. We also use an unsubscribe notice that tells users how many emails they can typically expect: one email per week, for six weeks.

And as for performance?

• Open rate: 50%
• Click-through-rate: 2.4%
• Conversion rate: 3.5%
• Unsubscribe rate: 0.4%

Onboarding 6.0: The current one

The next (and current) version is where we decided to simplify things. We stripped out a bunch of the flows which were adding maintenance overheads and bought everything down to a far simpler, linear flow where every email was determined by a simple user choice. Already using session replays? Fine, we'll email you about something else then.

It's also where, following some feedback, I decided to more directly embrace the PostHog tone of voice and make everything a little bit more sassy and anarchic. I added things like calling out product names, or the irony of emailing users about an email. I also made liberal use of memes from Lottie.

The biggest change, however, was changing just over 50% of the emails to come from me. As in, directly from my email address. Users can (and do) respond to them, and a growing amount of my time is now spent chatting with users over email - time well spent, in other words.

You'd think doing things like actively goading users into unsubscribing would be bad for conversion rates and subscriptions. You'd be wrong.

• Open rate: 52%
• Click-through-rate: 3.1%
• Conversion rate: 2.2% (with a new conversion event of activating any single product)
• Unsubscribe rate: 0.3%

What we learned

Across the ~five~ six major versions (and many other smaller iterations) of our email onboarding, there are a few definite things we've learned.

Email works for engineers too

As I said at the start, the conventional wisdom is that engineers are particularly skeptical of any marketing activity, and likely to unsubscribe to emails. However, our unsubscribe rates have been consistently low enough to disprove this. If the emails add value for engineers and are well-written for that audience, they can still be effective.

You can send more emails than you think

One of our consistent worries has been that we send too many emails. Again though, the low unsubscribe rate suggests that users aren't getting annoyed by this. The current email flow gives users approximately six emails, yet we've not heard any complaints about this.

We think the reason for this is...

1. We space emails out quite conservatively. Users will typically only get one email per week, and we are defensive about deviating from this schedule.

2. We tell users how many emails they'll get. It's added to the welcome email, right next to the unsubscribe button. This means users know we value their time and can unsubscribe before they get annoyed.

3. Our emails are actually worth reading. We don't send those awful outreach emails from sales reps, we don't run limited-time discounts, and we don't beg users to pay more. We just tell them how to get more out of PostHog.

Users prefer fun visual art to screenshots

I mean, who wouldn't want to look at a cute hedgehog instead of a graph?

We've run tests on this several times and Lottie's art is victorious by a small margin every time.

Personalization matters

Personalization doesn't mean calling users by their first name using a liquid tag and pretending to be best buds. It means providing the right content, at the right time, in the right way.

We do this in several ways in the latest version – delaying our welcome email, tailoring the emails users receive based on their choices, and sending newsletter invites from my personal address. If nothing else, the replies I frequently receive from users are a testament to how effective this is.

Now, though, we’re closing that Slack group in favor of creating a better, more scalable, and helpful PostHog community here on our site. In this post, I’ll explain why we’re doing this, and what comes next.

TL;DR: We’re closing the Slack group on Jan 24 2024 and inviting members to join the 1,500 users who have already moved to the new community forum. As thanks you'll receive a unique community achievement when you create an account. This decision only impacts the public Slack group, not private Slack channels for users who pay for additional support.

Why we're changing the way we do community

PostHog has grown incredibly fast over the last four years, and throughout that time the public Slack group has been the central hub of our community. It's been a place where we chat to users, listen to feature requests, answer questions, and solicit feedback.

However, with over 5,000 members in the community, it's become clear we've outgrown Slack as a platform. Messages quickly disappear from the chat history, it is disconnected from our main support flow, and useful solutions aren't searchable on our site or via Google.

We've explored a few things to fix this, including joining a paid plan ($7.25+ per user, per month) and building our own AI bot, but earlier this year we realized a new approach was needed.

So, we decided to build something better.

Introducing the PostHog community forum

Rather than using an off-the-shelf forum platform, like vBulletin or phpBB, we decided to create our own forum using Strapi as a headless CMS. The forum has already been live for several months while we worked out the kinks, and has over 1,500 active members.

The new forum provides a dedicated area where you can post questions for both the PostHog team and the wider community. Anyone can respond, and answers can be selected as the preferred solution to help guide other users in the future. All of this content is connected to our main support flow, lives permanently on our website, and is visible to search engines, which is enough on its own to make the forums a better place for seeking community support.

We've also integrated the forum into other parts of the site. You can, for example, post questions when browsing the PostHog docs, and we'll automatically aggregate them into sortable categories. This is especially helpful if you're following a guide and something isn't clear enough for you - just pop a question on it and we'll take a look.

Profiles are also a big part of the forum; a place where you can add info, track discussions you're involved in, and display achievements earned through the community. In fact, we're giving everyone who joins the PostHog community from the public Slack a unique achievement to say thanks for the support!

You can even open your profile to function as an Ask Me Anything (AMA) - something James, Cory, and myself have already done.

Ready to join the community? Create an account today and, if you've previously signed up to the public Slack group, you'll get a unique achievement to display on your community profile!

What happens next?

One option for the public Slack would have been to just leaving it running in parallel to the forums, but that would leave Slack users in limbo. Instead, we'll close the Slack in favor of the PostHog community forum. Currently you need to create an account to sign-up to the community, but soon we'll merge this with your normal PostHog account.

On January 12, we'll archive all channels in the public PostHog Slack, so that no new discussion or replies can be posted. This will give you chance to move on-going conversations to a new location, such as the PostHog community, without losing anything.

A week later, on January 22, we'll close the Slack group permanently and delete all existing content.

Important: Private Slack channels for paying users will continue to function as normal via Slack Connect, and we’ll continue to handle the majority of customer support via the in-app help. Obviously, the GitHub repos will also continue to function as normal and you can add comments or submit there too.

Thank you to everyone who participated in the Slack group over the last four years. Your support and feedback has helped us immensely, and we’ve deeply enjoyed speaking with you. We very much hope that all of you will join us in the new community soon!

1. LogRocket is a frontend monitoring tool that helps developers detects and solve issues.

2. PostHog goes beyond just frontend monitoring by integrating powerful analytics, session replay, feature management, A/B testing tools into one platform.

How is PostHog different?

1. PostHog replaces multiple tools

LogRocket is a narrow tool. It does one thing – and it does it well. PostHog includes many of the same tools, such as session replay, error tracking, and console logs, but augments them with powerful product and data tools, such as product analytics, feature flags, A/B testing, and surveys. It's everything you need to both squash bugs and understand every aspect of user behavior.

2. PostHog is open source and transparent

PostHog is built with transparency at its core. Not only do we work in the open and give full access to our source code, we also enable developers to build their own integrations, open their own PRs, or give feedback on our roadmap. PostHog's open app framework makes it easy to integrate internal tools, an advantage closed-source products like LogRocket can't offer.

3. PostHog constantly ships new features

We ship weirdly fast. We update our changelog every week and often there’s even more features in beta testing. And that’s without counting all the apps and integrations submitted by our community! We work hard to keep PostHog on the cutting edge and we’re transparent about the ways we do that.

Comparing PostHog and LogRocket

• Feature flags: PostHog offers robust, multivariate feature flags which support JSON payloads. This enables you to push real-time changes to your product without needing to redeploy. Visit our feature flag page for more information. LogRocket doesn’t have any in-built feature flag functions.

• Experiments: PostHog offers multivariate experimentation, which enables you to test changes and discover statistically relevant insights. Visit the experimentation page for more information. LogRocket doesn’t have any in-built experimentation features.

Product analytics

While LogRocket offers some product analytics features, it isn’t primarily intended as an analytics platform and lacks many things product managers and engineers require. PostHog is a more capable product analytics platform, offering advanced features such as correlation analysis, custom formulas, and a flexible SQL query editor.

• Correlation analysis: This feature enables you to automatically find correlated events or properties which affect the conversion rate of users within a funnel. LogRocket doesn’t offer any such automated correlation discovery, meaning users must search for correlating factors manually and without assistance.

Further reading: Comparing the most popular LogRocket alternatives

Session replay

LogRocket and PostHog both offer robust, full-featured session replay tools that outperform even dedicated tools, such as Hotjar.

• Mobile app recordings: Session replay for mobile apps is currently under development – see our GitHub mega issue for more.

Heatmaps and click tracking

Tracking & SDKs

Tl;dr: Both support a broad range of tracking options and libraries, and manual event instrumentation, as well as autocapture.

Library support

PostHog supports a wide range of client and server libraries, but not all features are equally available across all of them. We recommend using PostHog's JavaScript snippet to enjoy all our features. See our client library documentation for more information.

Data pipelines

PostHog offers more than 50 integrations and apps, while LogRocket offers slightly fewer. As an open-source application, PostHog welcomes contributions from the community. If an integration you need isn’t available, it’s possible to create it!

Popular integrations

Below, we've listed a few of the most popular integrations used across PostHog and LogRocket. See our app directory for a full list.

Compliance

Frequently asked questions

Got another question? You can ask the PostHog team anything you want!

How much does PostHog cost?

PostHog has transparent pricing based on usage with a generous free tier:

We charge progressively less the more you use once you exceed your monthly free allowance. Volume, non-profit and startup discounts are also available.

How can I estimate my usage?

The easiest way is to sign up to PostHog, integrate our snippet, then check the projection on your billing page after a few days. Alternatively, you can guesstimate by multiplying your current monthly active users by an estimate of events generated per user – 50 to 100 per user is a good starting point. See Estimating usage & costs in our docs for more.

Does PostHog block bots by default?

Yes. See the full blocklist in our docs.

What about ad blockers?

We recommend all users deploy a reverse proxy, which enables you send events to PostHog Cloud using your own domain. Events sent from your own domain and are less likely to be intercepted by tracking blockers, ensuring you capture the best data possible. We have reverse proxy setup guides for AWS Cloudfront, Caddy, Cloudflare, Netlify, Vercel, and more, in our docs.

Can I use PostHog with a CDP? (Segment, Rudderstack, etc.)

Yes. See Using PostHog with a CDP in our docs.

Suffice to say that the new version has a lot to offer. It’s got dark mode! It’s got data density! It’s got side panels! It’s got... notebooks?

Notebooks is a brand new feature, available to all users for free. Think of it as an ever-present place to collect data from within PostHog, explore it, add context, and even share with others.

Notebooks help in a wide variety of typical tasks, or workflows.

Investigating a bug report? Drag and drop session replays into a scratchpad and watch them as normal, or add timestamped comments to break things down.

Researching a new idea? Collect insights and add them to your proposal seamlessly, alongside survey results or cohorts.

Planning a launch? Embed the feature flags, events, persons, or cohorts you’ll need to deploy changes and track success.

Or, obviously, you can use a notebook as just a daily scratchpad. This is actually the default behavior, whereby a freshly opened notebook will persist only in the browser until you save it.

There’s no limit to how many notebooks you can create, or how you can share them within your organization. Colleagues can even collaborate with you by adding to a notebook, though we block multiplayer editing to stop things getting too messy.

We’ve been testing notebooks internally and with some users for the past few weeks, and have found they’ve already become indispensable to the way we work. We’ve created notebooks for running engineering sprints, tracking marketing activities, investigating errors, and even monitoring feature adoption.

Best of all, we’re just getting started. We have a lot more planned for notebooks, including more templates, more sharing options, and built-in cheatsheets to help you get even more from PostHog. We’d love to know any feedback or ideas you have — or you can just jump straight in and start creating your own notebooks immediately.

While it's tempting to build your own tools, there's no shortage to feature flag services with decent free tiers, or open-source feature flag tools you can self-host. These are the best and most popular right now.

The best open-source feature flag tools

Changelog:

• Mar 15, 2023: Removed DevCycle, added Flipt. Updated copy and screenshots.
• Sep 14, 2023: Removed Harness, added Flipper and FeatBit, added SDKs
• Dec 11, 2023: Added FeatureHub

1. PostHog

• License: MIT
• GitHub stars: 14.8k
• PostHog on GitHub

What is PostHog?

PostHog is a popular alternative to LaunchDarkly and analytics tools like Amplitude.

It's an all-in-one platform, which means it offers product analytics, session replay, A/B testing, and user surveys, in addition to feature flags.

PostHog's feature flags enable conditional rollouts, multivariate flags, JSON payloads, automated usage reports, local evaluation, and more. They power its A/B testing features, so you can test product changes for their impact on key metrics.

PostHog also offers an early access feature management feature, which enables users to self-onboard to beta features behind feature flags.

Supported libraries

• Client-side: JavaScript, React Native, React, Android, iOS, Flutter, Remix, Angular, Next.js, Svelte.

• Server-side: Node.js, Python, PHP, Ruby, Go.

You can also evaluate feature flags using PostHog's API from any language that can make HTTP requests.

How much does it cost?

There are no limits on feature flag usage in the open-source edition, though A/B testing isn't available. PostHog's Cloud edition includes 1 million API requests per month for free – see the feature flag pricing page for more info.

2. Unleash

• License: Apache
• GitHub stars: 9k
• Unleash on GitHub

What is Unleash?

Unleash is open-source feature flag service with a public roadmap and a strong community. Official SDK support is comprehensive and there's a large selection of community maintained SDKs for less popular frameworks, too.

Supported SDKs

• Client-side: JavaScript, Android, Flutter, iOS, React, Svelte, Vue.

• Server-side: Go, Java, Node.js, PHP, Python, Ruby, Rust, .NET.

• Community maintained: Angular, Clojure, C++, ColdBox, Dart, Elixir, Haskell, Laravel, NestJS, React Native, Solid.

How much does it cost?

Unleash is free for two environments when you self-host. A managed version is also available at cost if you want to add team members or access additional features like advanced segmentation or custom activation strategies.

3. GrowthBook

• License: MIT
• GitHub stars: 5.2k
• GrowthBook on GitHub

What is GrowthBook?

GrowthBook is a warehouse-native, open-source feature flagging and experimentation platform built to help companies make better data-driven decisions. Its flag tool features advanced targeting, gradual rollouts, and experiments.

It's warehouse native with support for most major SQL data sources, and there's a code-free visual experiment editor that can used be less technical end users.

Supported SDKs

• Client-side: JavaScript, React, Android (Kotlin), iOS (Swift), Flutter, Vue

• Server-side: PHP, Ruby, Python, Go, Java, C#.

How much does it cost?

Growthbook’s open source self-hosted plan is free forever, with unlimited flags and experiments. GrowthBook also offers a hosted version with pricing based on the number of users.

4. Flagsmith

• License: BSD 3-clause
• GitHub stars: 3.4k
• Flagsmith on GitHub

What is Flagsmith?

Flagsmith is a feature flag tool that makes it easy to test and deploy new functional and visual changes to users without pushing updates to code.

All flags in Flagsmith are capable of being configured for cross-platform remote configuration, so you can alter an app in real-time without having to wait for app store approval.

It supports local evaluation, scheduling, and paid plans offer an Edge API that automatically routes requests to the closest data center to the user.

Supported SDKs

• Client-side: JavaScript, React Native, React, Android, iOS, Next.js.

• Server-side: Node.js, Python, PHP, Ruby, Go, .NET, Ruby, Rust, Elixir.

How much does it cost?

Flagsmith is free up to 50,000 requests per month when you choose the cloud feature flag service. In addition to open source, Flagsmith also offers self-hosted and private cloud options.

5. Flipt

• License: GPL-3.0
• GitHub stars: 3k
• Flipt on GitHub

What is Flipt?

Flipt is a popular open-source option for feature flagging. It's 100% open source with no paid editions. It supports all basic feature flag use cases, such as rolling out features based on segments (e.g. new users, email, country etc.), multiple variations, and percentage thresholds for rollouts. It also supports storing your flags in Git, so it can be used without a database.

Supported SDKs

• Client-side: JavaScript/TypeScript, React, Android (Kotlin).

• Server-side: C#, .NET, C++, Dart, Go, Java, Node, PHP, Python, Ruby.

How much does it cost?

Nothing. There is no paid version of Flipt. It's distributed under a GPL-3.0 license.

6. Flipper

• License: MIT
• GitHub stars: 3.4k
• Flipper on GitHub

Flipper is a rebuilt version of the feature flag platform used by GitHub, maintained by a small team of ex-GitHub engineers. It supports all the common use cases for feature flags, such as boolean flags, group targeted flags, and targeting by actors or a percentage of actors. It currently only supports Ruby applications, though support for other languages is planned.

Supported SDKs

Ruby only

How much does it cost?

In addition to the free open-source edition, Flipper also offers a cloud-hosted version with a 14-day trial. It's priced at $20 per seat with unlimited flags and projects.

7. FeatBit

• License: MIT
• GitHub stars: 1.1k
• FeatBit on GitHub

What is FeatBit?

FeatBit is an open-source feature flag and A/B testing tool that's available in Standard (for small businesses) and Professional (for enterprises) flavors – the Standard edition is easier deploy due to fewer dependencies. It supports custom user segments, rolling out by percentages, and feature flag scheduling. A/B testing data can also be exported to tools like Datadog, Grafana, and Amplitude.

Supported SDKs

• Client-side: JavaScript, React, Vue.

• Server-side: Python, PHP, Ruby, Go, .NET, Java.

How much does it cost?

All features are completely free. FeatBit monetizes its platform by offering premium support packages, which start at $399 per month for dedicated support with a 2-hour reply SLA during business hours.

8. FeatureHub

• License: Apache 2.0
• GitHub Stars: 276
• Feature Hub on GitHub

What is FeatureHub?

FeatureHub is an open-source feature flag tool that offers all the basics you need, including multivariate flags, client and server-side evaluation, and percentage rollouts.

The open-source version has no usage limits – it even supports SAML and social sign-in via Microsoft, Google and GitHub. There's also a Google Analytics integration, though this is the only analytics integration it offers at present.

Supported SDKs

• Client-side: JavaScript, React, Flutter, iOS (Swift), Android, Angular.

• Server-side: Node, C#, Dart, Go, Python, Ruby, Java.

How much does it cost?

FeatureHub offers a hosted option in addition to its open-source edition, though both version offer the same features. The hosted version is $4.99 per user, per month and ~$0.33-0.38 per 10,000 API requests.

Further reading

• Feature flag best practices and tips (with examples)
• Should you build or buy? Feature flags as a service, explained
• Feature flags vs configuration: Which should you choose?

1. PostHog

• Free tier: ✔
• Open source: ✔
• Transparent pricing: ✔
• Includes product analytics: ✔
• Get started without talking to sales: ✔
• Best for: Engineering and product teams

What is PostHog?

PostHog (that's us!) is an all-in-one platform built to help engineers create better products. It includes A/B testing and a whole bunch more, such as product analytics, session replays, feature flags, and surveys.

It's designed for product-minded engineers, growth teams, and product managers who need to move fast and iterate based on reliable, actionable insights.

Which platforms does PostHog support?

• iOS: ✔
• Android: ✔
• React Native: ✔
• Flutter: ✔
• Web: ✔
• Server: ✔

How much does PostHog cost?

PostHog has transparent pricing based on the usage. It’s free to get started and completely free for the first 1 million A/B testing requests. After this free monthly allowance, you'll pay $0.0001/request, and requests cost progressively less the more you use. You can also set billing limits to ensure you don't get surprise bills.

While PostHog is open source, its A/B testing features aren't included in the open source release.

Further reading: New to A/B testing? Read a software engineer's guide to A/B testing and our guide to common A/B testing mistakes.

2. VWO

• Free tier: ✖
• Open source: ✖
• Transparent pricing: ✔
• Includes product analytics: ✔
• Get started without talking to sales: ✔
• Best for: Non-technical teams

What is VWO?

VWO is a testing platform that primarily targets large enterprises. Its experimentation platform includes support for A/B and multivariate tests and a visual editor. It's useful for non-technical users who need an easy-to-use interface to test in-app messaging and UI copy.

It also offers product analytics, so you don't need a separate platform or integration to link your analytics data to your A/B test results.

Which platforms does VWO support?

• iOS: ✔
• Android: ✔
• Flutter: ✔
• React Native: ✔
• Web: ✔
• Server: ✔

How much does VWO cost?

VWO mobile app A/B tests starts at $1,526 per month for up to 10k users. This is billed annually, so it means an upfront cost of $18,312. Also note that this price is only for mobile A/B tests. If you want to include analytics in your package, prices start at an additional $15,936 per year.

Lastly, if you also want to run server-side tests, you also need to purchases a separate package. Prices for this start at $28,164 per year.

3. Optimizely

• Free tier: ✖
• Open source: ✖
• Transparent pricing: ✖
• Includes product analytics: ✖
• Get started without talking to sales: ✖
• Best for: Marketing teams

What is Optimizely?

Optimizely is an all-in-one system for marketing that includes A/B testing and more – like content management, campaign planning, asset management, and checkout customizations.

It's built for marketers in large enterprises looking to optimize their content, apps, and e-commerce experiences.

Which platforms does Optimizely support?

• iOS: ✔
• Android: ✔
• Flutter: ✔
• React Native: ✔
• Web: ✔
• Server: ✔

How much does Optimizely cost?

Optimizely's pricing is sales-driven and completely opaque, so it's not clear how much it costs. However, according to Splitbase, Optimizely will cost you at least $36,000 per year.

4. Firebase

• Free tier: ✔
• Open source: ✖
• Transparent pricing: ✔
• Includes product analytics: ✖
• Get started without talking to sales: ✔
• Best for: Apps already using Firebase's other features

What is Firebase A/B testing?

Firebase A/B testing is built specifically for mobile apps and is completely free. It includes an easy to use interface for running experiments on Firebase's other app features, such as cloud messaging and in-app messaging.

It's important to note that using A/B testing on Firebase requires installing Google Analytics in your app. It's also not possible to run A/B tests on your web app or server.

Which platforms does Firebase support?

• iOS: ✔
• Android: ✔
• Flutter: ✔
• React Native: ✖
• Web: ✖
• Server: ✖

How much does Firebase A/B testing cost?

A/B testing with Firebase is completely free.

5. Kameleoon

• Free tier: ✖
• Open source: ✖
• Transparent pricing: ✖
• Includes product analytics: ✖
• Get started without talking to sales: ✖
• Best for: E-commerce and healthcare companies

What is Kameleoon?

Kameleoon is built specifically for mobile apps. It includes an easy to use interface for running experiments on push notifications and in-app messaging.

In addition to testing, it offers a real-time personalization engine that's particularly useful for e-commerce apps. It's also HIPAA-compliant, which makes Kameleoon a viable option for healthcare apps.

Which platforms does Kameleoon support?

• iOS: ✔
• Android: ✔
• Flutter: ✔
• React Native: ✔
• Web: ✔
• Server: ✔

How much does Kameleoon cost?

Kameleoon doesn't publish pricing publicly, but conversion optimization consultants BrillMark state pricing starts at $35,000 per year and scales based on traffic volume.

6. Statsig

• Free tier: ✔
• Open source:✖
• Transparent pricing: Partial
• Includes product analytics: ✔
• Get started without talking to sales: ✔
• Best for: Product teams in high-growth startups

What is Statsig?

Statsig is platform built for feature management and experimentation. It offers feature flags, A/B testing, and product analytics.

It offers advanced experimentation techniques, such as multi-armed bandit experiments and holdout testing, making it well suited for growth teams in startups.

Which platforms does Statsig support?

• iOS: ✔
• Android: ✔
• Flutter: ✔
• React Native: ✔
• Web: ✔
• Server: ✔

How much does Statsig cost?

Statsig has a free tier with an allowance of 1 million events. After that, you'll pay $150 per month for up to 5 million events and $50 for every 1 million events thereafter. They also offer enterprise discounts for large volume customers.

7. LaunchDarkly

• Free tier: ✔
• Open source: ✖
• Transparent pricing: Partial
• Includes product analytics: ✖
• Get started without talking to sales: ✖
• Best for: Engineering and product teams

What is LaunchDarkly?

LaunchDarkly is built for enterprises wanting to follow software development best practices. This means A/B testing changes, managing features, de-risking releases, and coordinating deploys.

The people who find LaunchDarkly most useful are engineering managers, site reliability engineers, and product managers.

Which platforms does LaunchDarkly support?

• iOS: ✔
• Android: ✔
• Flutter: ✔
• React Native: ✔
• Web: ✔
• Server: ✔

How much does LaunchDarkly cost?

Experimentation is available on LaunchDarkly's Pro plan, which costs $20 per seat, per month, but experiments are priced separately to this. As per LaunchDarkly's docs:

"Experiments are billed monthly based on the number of unique user keys in each experiment. To estimate your Experimentation billing costs, multiply the number of experiments you run each month by the number of unique users in your experiment audience."

8. Taplytics

• Free tier: ✖
• Open source: ✖
• Transparent pricing: ✖
• Includes product analytics: ✖
• Get started without talking to sales: ✖
• Best for: Product and marketing teams

What is Taplytics?

Taplytics is an A/B testing platform focusing on mobile. Its no-code experiment builder makes it easy for non-technical users to create and run A/B tests.

Which platforms does Taplytics support?

• iOS: ✔
• Android: ✔
• Flutter: ✖
• React Native: ✔
• Web: ✔
• Server: ✔

How much does Taplytics cost?

Taplytics don't share their pricing publicly, so it's not clear how much it costs. You need to talk to sales first.

9. Apptimize

• Free tier: ✖
• Open source: ✖
• Transparent pricing: ✖
• Includes product analytics: ✖
• Get started without talking to sales: ✖
• Best for: Product teams in large enterprisess

What is Apptimize?

Similar to Taplytics, Apptimize focsuses on A/B tests on mobile. It doesn't have a no-code builder like Taplytics has, which makes it less user friendly for non-technical users.

Which platforms does Apptimize support?

• iOS: ✔
• Android: ✔
• Flutter: ✔
• React Native: ✔
• Web: ✔
• Server: ✔

How much does Apptimize cost?

Apptimize don't share their pricing publicly. You need to ask sales for a custom quote.

Is PostHog right for you?

Here's the (short) sales pitch.

We're biased, obviously, but we think you'll love PostHog if:

• You value transparency (we're open source and open core)
• You want tools to ship, track, and analyze new features – like A/B testing, feature flags, and session replays
• You want try before you buy (we're self-serve with a generous free tier)

Check out our product pages and read our docs to learn more.

Further reading

• The 6 best free and open-source A/B testing tools
• 7 best session replay tools for mobile apps (iOS & Android)
• The 7 best mobile app analytics tools

This guide only includes open-source analytics tools that:

1. Enable users to visualize and analyze behavioral data for products or websites. This includes tools that actively capture data as well as dashboarding tools that connect to existing data sources, but not data pipeline platforms like Snowplow or Rudderstack.

2. Are actively developed or maintained by a company or community. This is subjectively judged based on recently merged pull requests, commits to their repos, and the responsiveness of the maintainer to review requests.

3. Are self-hostable. Tools that are "open core" but don't offer a free, self-hostable deployment option are not included.

Who maintains this guide?

This guide is maintained by PostHog, an open-source analytics platform. You can suggest additions to this guide by opening an issue on our website repo. We update the guide ~3 times per year.

We maintain separate guides on open-source feature flag tools and open-source AB testing products.

Changelog

• Mar 14, 2022: Added Umami; updates to copy and screenshots
• Jun 16, 2022: Added Fathom Lite, GrowthBook, OpenReplay
• Aug 24, 2022: Removed Fathom Lite (dormant), added Apache Superset, Redash
• Sep 13, 2023: Added Ackee and Open Web Analytics
• Nov 22, 2023: Added inclusion criteria, removed Ackee (dormant)

The best open source analytics tools

1. PostHog

• GitHub Stars: 14.4k
• Category: Product and website analytics
• Alternative to: Mixpanel, Amplitude, Hotjar, FullStory, LaunchDarkly
• Language(s): TypeScript (53%) and Python (42%)
• PostHog on GitHub

What is PostHog?

PostHog is an all-in-one platform designed to give engineers and product managers a complete view of user behavior. It's built on the highly-scalable ClickHouse OLAP database.

Features include a comprehensive product analytics suite (e.g. trend analysis, dashboards, funnels, path analysis, retention tracking etc.), feature flags for feature control, experimentation (A/B and multivariate testing), session recording, heatmaps and more.

It’s entirely self-serve, with the ability to autocapture event information out of the box. PostHog can also be self-hosted on your existing infrastructure using Docker Compose, making it ideal if you don't want to share data with third parties.

It's a good option if you're after an open-source session recording tool, too.

Who is PostHog for?

PostHog is useful for engineering, data science, and product teams. As an all-in-one platform, it can replace multiple tools (e.g. Mixpanel, LaunchDarkly, Hotjar) so you can simplify your tech stack. PostHog customers include the likes of Hasura, Netdata, and Phantom.

Strengths

• All-in-one analytics suite
• Easy-to-use, no SQL required
• Self-hosting and cloud-hosting available
• Feature flags for safely deploying changes
• Multivariate experimentation suite for testing ideas
• Session recording for gathering insights
• Event pipelines to integrate with data warehouses
• Built on the ultra-fast ClickHouse OLAP database

License and monetization

PostHog Open Source is free to use for life and distributed under an MIT license. The open-source version includes the core product analytics, feature flag, and session recording features, but has a one project limit. There's also a FOSS edition that removes all proprietary code.

Subscribing to PostHog Cloud removes the project limit and adds numerous paid-only features, including experimentation, correlation analysis, group analytics for tracking organizations, and advanced cohorts. PostHog Cloud is free up to 1 million events per month.

2. Matomo

• GitHub Stars: 18.4k
• Category: Website analytics
• Alternative to: Google Analytics
• Languages: PHP (78%), JavaScript (8.1%)
• Matomo on GitHub

What is Matomo?

Matomo is an open-source platform for website analytics. It's one of the most popular open-source alternatives to Google Analytics 4 for website owners and marketing teams. It even offers a data import tool for Google Analytics, so you can bring your data with you. You can run it on-premise or use Matomo's own cloud hosting service.

Who is Matomo for?

Matomo is intended for marketing and website teams looking to track content performance and marketing attribution. It's suitable for both SMBs and enterprises – the European Union runs a version of Matomo on all its websites.

Strengths

• Google Analytics data importer
• Easy to set up on WordPress and WooCommerce
• Marketing metrics and attribution
• Integrations with popular off-the-shelf CMS
• Built-in GDPR and tag managers

Open source license and monetization