Stripe can’t A/B test prices by itself: its built‑in A/B testing only compares payment methods.
- A price test needs a second Stripe price, shown at random to part of your new visitors.
- The winner is the price with more revenue per visitor, not more sales.
- PricingFast does both on your own pricing page. Your coding agent installs it.
What Stripe tests today
Stripe’s A/B testing is built for payment methods: “to measure the impact of offering new payment methods to a percentage of customers”, such as buy now, pay later options.[1] It works in the Dashboard, scores tests by revenue per checkout session, and has no option to compare two prices.
Stripe does publish advice on running pricing experiments, but leaves the experiment itself to you.[2]
Three ways to test a Stripe price
| Build it yourself | General A/B tool | PricingFast | |
|---|---|---|---|
| Test price in Stripe | Create by hand | Create by hand | Created for you |
| Random split of new visitors | Your code | Feature flag | Built in |
| Right price at checkout | Your code | Your code | Built in |
| Revenue per visitor | Your queries | Join Stripe data to visitors | Built in |
| When to stop | Your call | Tool’s stats | 95% confidence, 30 sales |
| If the service is down | n/a | Depends on your code | Your prices within 1.5 s |
Test price in Stripe
- Build it yourself
- Create by hand
- General A/B tool
- Create by hand
- PricingFast
- Created for you
Random split of new visitors
- Build it yourself
- Your code
- General A/B tool
- Feature flag
- PricingFast
- Built in
Right price at checkout
- Build it yourself
- Your code
- General A/B tool
- Your code
- PricingFast
- Built in
Revenue per visitor
- Build it yourself
- Your queries
- General A/B tool
- Join Stripe data to visitors
- PricingFast
- Built in
When to stop
- Build it yourself
- Your call
- General A/B tool
- Tool’s stats
- PricingFast
- 95% confidence, 30 sales
If the service is down
- Build it yourself
- n/a
- General A/B tool
- Depends on your code
- PricingFast
- Your prices within 1.5 s
Building it yourself is a fine choice for a team with time and a data person. A general tool like PostHog can measure Stripe revenue per variant once you link Stripe customers to visitors in its data warehouse.[3] We compare the two in detail in PricingFast vs PostHog.
How it works with PricingFast
- 1
Connect Stripe and say what to try
Connect with OAuth, pick a plan and the price to test, for example €69 to €89.
- 2
Your coding agent installs it
From one prompt or through the PricingFast MCP server. It adds a script, a
data-pf-priceattribute on your prices and aPRICINGFAST_ENABLEDoff switch, on a branch you review. - 3
Rehearse on localhost
See both prices, check that your page still works without PricingFast, and make a test purchase. Visitors on your live site see nothing new yet.
- 4
Press Launch
Choose 10, 25 or 50% of new visitors for the test price. Only you can launch, from the dashboard. The verdict appears when one price is ahead with 95% confidence.
Full details are in the install guide.
What changes in your Stripe account
| What happens | |
|---|---|
| Today’s prices | Untouched |
| Test prices | Added as normal prices on your products, under your plan names |
| Checkout | Runs through Stripe as usual; card details never reach us |
| Existing subscriptions | Never changed |
| When the test ends | The price you don’t keep is archived; its buyers keep their plan |
Today’s prices
Untouched
Test prices
Added as normal prices on your products, under your plan names
Checkout
Runs through Stripe as usual; card details never reach us
Existing subscriptions
Never changed
When the test ends
The price you don’t keep is archived; its buyers keep their plan
Example: a test that loses
Not every test finds a better price, and that’s a result too. RunStack, a visual LLM workflow builder, tests €89 instead of €69 on its Pro plan.
The test price earns 20% less per visitor.
| Price | Visitors | Sales | Per visitor |
|---|---|---|---|
| €69/moToday’s price | 6,000 | 168 | €1.93 |
| €89/moTest price | 6,000 | 104 | €1.54 |
Verdict reached. Far fewer teams buy at €89, so each visitor brings less. RunStack keeps €69. The €89 price is archived, and the teams who bought it keep their plan.
Questions
Can Stripe A/B test prices?
Not on its own. Stripe's built-in A/B testing measures the impact of offering new payment methods to a share of customers. To test a price you need a second Stripe price, a way to show it to part of your visitors, and a comparison of revenue per visitor.
What does PricingFast change in my Stripe account?
It adds your test prices as normal prices on your existing products, under your plan names. Today's prices aren't touched. When the test ends, the price you don't keep is archived: nobody new can buy it, and subscribers already on it keep their plan.
Does PricingFast see my customers' card details?
No. Checkout runs through Stripe as usual and card details never reach PricingFast. You connect Stripe with OAuth.
What happens to my checkout if PricingFast is down?
Your page shows its own prices and your checkout keeps working. The PricingFast script loads in the background; if it doesn't answer within 1.5 seconds, your page falls back to today's prices.
Will my existing subscribers be moved to the test price?
No. PricingFast never changes existing subscriptions. Only new checkouts can see a test price.
Sources
- [1]A/B testing a payment method, Stripe documentation
- [2]Pricing experiments, Stripe resources
- [3]Experiments with data warehouse metrics, PostHog documentation