Test changes with statistical significance
A/B tests, multivariate tests, and robust targeting & exclusion rules. Analyze usage with product analytics and session replay.
boosted community engagement by 40%
"Y Combinator uses PostHog's experiments to try new ideas, which has led to significant improvements."
tests product changes for over 25M users
"Our data scientists are able to rapidly and autonomously iterate on the data models that power our home feed."
increased registrations by 30%
"This experiment cuts drop-off in half – that's a 50% improvement without a single user complaining!"
switched from Mixpanel for a leaner stack
"I feel like, every single week, we discover something new that makes a difference."
Features
Customizable goals
Conversion funnels or trends, secondary metrics, and range for statistical significance
Targeting & exclusion rules
Set criteria for user location, person property, cohort, or group
Recommendations
Automatic suggestions for duration, sample size, and confidence threshold in a winning variant
Built on Feature Flags
All the benefits of feature flags with added functionality around stat-sig experimentsJSON payloads
Modify website content per-variant without additional deploymentsSplit testing
Automatically split traffic between variantsMultivariate testing
Test up to 9 variants against a controlDynamic cohort support
Add new users to an experiment automatically by setting a person property
Answer all of these questions (and more) with PostHog Experiments.
- Does this new onboarding flow increase conversion?
- How does this affect adoption in Europe?
- Will enterprise customers like this new feature?
Usage-based pricing
Use Experiments free. Or enter a credit card for advanced features.
Either way, your first 1,000,000 requests are free – every month.
Free
No credit card required
All other plans
All features, no limitations
Requests
1,000,000/mo
Unlimited
Features
Unlimited feature flags
Boolean feature flags
Multivariate feature flags & experiments
Persist flags across authentication
Test changes without code
Multiple release conditions
Release condition overrides
Flag targeting by groups
Local evaluation & bootstrapping
Flag usage stats
Experiments
Funnel & trend experiments
Secondary experiment metrics
Statistical analysis
Group experiments
Multi-environment support
Data retention
1 year
7 years
Monthly pricing
First 1 million requests
Free
1-2 million
$0.000100/request
2-10 million
$0.000045/request
10-50 million
$0.000025/request
50 million+
$0.000010/request
FAQs
PostHog vs...
So, what's best for you?
Reasons a competitor may be best for you (for now...)
- No-code experiments or CMS capabilities
- You'll still need a designer/engineer to create experiments
- No integration with Google Ads
- PostHog can't run ad experiments, or target users into an experiment based on an ad variant engagement.
Reasons to choose
- Integration with other PostHog products
- Attach surveys to experiments or view replays for a test group. Analyze results beyond your initial hypothesis or goal metric.
- Automated recommendations for sample sizes and runtime
- Automatic significance calculator – to help you figure out the winning variant as quickly as possible
- Robust targeting and exclusion options, including cohorts and location
- Anything you monitor in analytics, you can target in an experiment
Have questions about PostHog?
Ask the community or book a demo.
Featured tutorials
Visit the tutorials section for more.
Running experiments on new users
Optimizing the initial experience of new users is critical for turning them into existing users. Products have a limited amount of time and attention from new users before they leave and churn.
How to set up A/B/n testing
A/B/n testing is like an A/B test where you compare multiple (n) variants instead of just two. It can be especially useful for small but impactful changes where many options are available like copy, styles, or pages.
How to run holdout testing
Holdout testing is a type of A/B testing that measures the long term effects of product changes. In holdout testing, a small group of users is not shown your changes for a long period of time, typically weeks or months after your experiment ends.
How to do A/A testing
An A/A test is the same as an A/B test except both groups receive the same code or components. Teams run A/A tests to ensure their A/B test service, functionality, and implementation work as expected and provides accurate results.
Install & customize
Here are some ways you can fine tune how you implement Experiments.
Explore the docs
Get a more technical overview of how everything works in our docs.
Experiments
Features
Under the hood
Bayesian method
Frequentist method
Meet the team
PostHog works in small teams. The Experiments team is responsible for building Experiments.
Roadmap & changelog
Here’s what the team is up to.
Latest update
Jun 2025
Experiments can now use frequentist statistical methods
You can now create experiments in PostHog which operate using the frequentist statistical approach, rather than Bayesian. If that sounds like white noise to you, what it basically means is: Yay, more maths!
Frequentist statistical models are a more traditional approach for reviewing experiments, and they tend to be preferred in larger companies because data teams like them for their easily validated results. Bayesian is still available as an option if you're an absolute renegade though and we've added the option to set your preference at the organization level too.
Up next
No-code A/B testing
A visual editor for experiments would allow users to test changes to their website / app without having to touch the code.
https://posthog.com/docs/experiments/no-code-web-experiments
Progress
Milestones
Project updates
No updates yet. Engineers are currently hard at work, so check back soon!
Questions?
See more questions (or ask your own!) in our community forums.
- Question / TopicRepliesLast active
Pairs with...
PostHog products are natively designed to be interoperable using Product OS.
Product analytics
Run analysis based on the value of a test, or build a cohort of users from a test variant
Session replay
Watch recordings of users in a variant to discover nuances in why they did or didn’t complete the goal
Feature flags
Make changes to the feature flag the experiment uses - including JSON payload for each variant
This is the call to action.
If nothing else has sold you on PostHog, hopefully these classic marketing tactics will.
PostHog Cloud
Digital download*
Notendorsed
by Kim K
*PostHog is a web product and cannot be installed by CD.
We did once send some customers a floppy disk but it was a Rickroll.