Test it.
Keep what
wins.
Run A/B tests on any Funnel step and see which version earns more. On Scale, hold back a control group to prove the revenue a Funnel adds.
14 days of Traction · No card required

Less guessing,
more proof.
1 · Vary
Add a variant.
Test up to four versions: Control, Variant B and more. Traffic splits evenly unless you change it.
2 · Hold back
Keep a control group.
Hold back 1 to 50% of eligible accounts (10% by default). They see nothing at all.
3 · Decide
Pick the winner.
Results count paid retention from first assignment, and wait for a minimum sample before any call.
A/B tests
Variants on any step.
Change the headline, the offer or the format and split traffic between versions. Every variant reports the same Stripe-verified results.
- Up to 4 variants per Funnel
- Even traffic split by default
- The same verified results for every variant

Control groups
Proof, not a guess.
A control group sees nothing. Comparing it with everyone who saw the Funnel shows the revenue the Funnel added, not revenue that would have come anyway.
- Hold back 10% by default, from 1 to 50%
- Split by account or by anonymous visitor
- Plain verdicts, with a warning when it’s too early

Plans
A/B tests start on Traction.
Control groups that prove added revenue are on Scale.
Compare plans →Builder
Not included
—
Launch
Not included
—
Traction
Included
A/B tests
Scale
Included
A/B tests + control groups
Pairs well with.
All featuresFunnels
Turn a moment into the right offer, in your app.
Act
See FunnelsRevenue attribution
Every dollar confirmed by Stripe.
Prove
See Revenue attributionAI Funnel Copilot
Draft and edit Funnels by describing them.
Act
See Copilot
Questions,
answered.
How long should a test run?
Until both groups reach the minimum sample. The report says when a result is still early, so you don’t pick a winner too soon.
What’s the difference between an A/B test and a control group?
An A/B test compares versions of a Funnel. A control group sees no Funnel at all, which shows how much revenue the Funnel adds on its own.
What does a test measure?
Paid retention from each account’s first assignment, counted from Stripe, with a minimum sample per group before any verdict.
Can I test a whole Funnel or one step?
Variants are set per step, so you can change one step’s headline, offer or format while the rest of the Funnel stays the same.
Which plans include experiments?
A/B tests start on Traction. Control groups are on Scale.
