62% of visitors leave while reading the headline. It names the category (“optimization”) but never says what the visitor gets.
Data-driven website optimization for modern teams.
Let your visitors pick the better page.
Let your visitors pick the better page.
3.9% sign-up rate
Data-driven website optimization for modern teams.
Variant A, “Data-driven website optimization for modern teams.” 2.9% retired
2.9% sign-up rate
This page is testing its own headline. 1 dot = 20 visitors. Simulated data.
+33%more sign-ups with B
2.9% → 3.9%
2.9% → 3.9%
Forma shows two versions of a headline, a button or a whole page to real traffic and tells you which one people prefer. The headline above just won that test.
Observe. See where people give up.
Forma follows how visitors move through a page and marks the place where most of them leave. On this sample shop it is not the price. It is the very first sentence.
62%of visitors leave while reading the headline
Headline62%
Product picture9%
Features11%
Plans13%
- visitarrived from an ad · /
- clickopened FAQ · “Do I need a developer?”
- leaveleft at the headline · 0:04
- scrollread to the plans · 0:41
- visitarrived from a newsletter · /
- leaveleft at the headline · 0:06
- visitarrived from search · /pricing
Simulated sessions. Each block shows the share of visitors who leave there.
Hypothesize. Get one change worth trying.
Forma does not hand you forty charts. It reads the drop-off, compares it with what usually works on pages like yours, and suggests one change to test next, with the reason written out.
Every change to a website starts as a guess. Forma turns the guess into a test, the test into a number, and the number into a decision you can ship on Monday.
Test. Real traffic, two versions, one winner.
Half the visitors see the old page, half see the new one. Forma counts what each half does and calls the winner only when the difference is real, not luck. Here is a test on a sample tea shop. Click a button to add your own visit.
Fernhill Tea Club
Three loose-leaf teas a month, $18. Skip or cancel any time.
850 visitors4.6% subscribe
Fernhill Tea Club
Three loose-leaf teas a month, $18. Skip or cancel any time.
860 visitors6.5% subscribe
96% chance B is better · ships at 95% sure · B ships to every visitor
Simulated visitors, sample shopShip. Keep the winner. Queue the next guess.
When a test is decided, Forma switches every visitor to the winning version and moves the next idea up the queue. Losing tests are kept too, so nobody runs the same bad idea twice.
- more sign-ups from one headline
- +33%
- days from idea to decision
- 9
- lines of code to ship the winner
- 0
Experiments · sample-shop.examplesign-up rate over the test, %
| Test | Status | Trend | Result |
|---|---|---|---|
| Homepage headlineB: “Let your visitors pick the better page.” | shipped | +33% sign-ups | |
| Pricing page buttonB: “Try it free for 14 days” | running | 71% chance B wins | |
| Checkout trust badgesB: badges above the card form | kept A | −2% orders · not shipped | |
| Welcome email subjectB: “Your first test is ready” | queued | — | starts after the button test |
Notes from people who ship this way
We argued about the homepage for a month. The test took nine days and settled it. Nobody argues anymore, we just test.
The best part is when it tells you to keep the old version. It saved us from shipping a redesign that would have cost us orders.
I am not a developer. I changed a button, waited a week, and had a number I could show my boss.
Fictional people and quotes, written for this concept.
And back to Observe. The next test starts where this one ended.
Pricing. Pay for the traffic you test.
A test needs visitors to reach a clear answer, so plans follow your monthly traffic. Move the slider to your numbers.
That is enough traffic to decide about 20 tests a month.
Growth
$39/ month
$468 billed once a year
For teams that want a test running every week.
- 3 websites
- 10 tests running at once
- Daily notes and alerts
- Page and button tests
- Priority support
14 days free, no card needed.
| Plan | Visitors a month | Tests at once | Price |
|---|---|---|---|
| Starter | up to 10,000 | 2 | $15 / mo |
| Growth | up to 100,000 | 10 | $39 / mo |
| Scale | up to 500,000 | 40 | $79 / mo |
Questions. Short answers.
Until the result is clear, usually one to three weeks. Forma watches the numbers and calls the winner once it is 95% sure, so you never stop a test too early or let it run forever.
No. You paste one script tag once, then pick the text, button or image to change right on your page. Bigger page-level tests can use a second URL.
The script loads after your page and weighs about 9 KB. Visitors see their version on the first paint, without a flash of the old one.
Then Forma says so and suggests keeping the version that is simpler to maintain. A “no difference” result is still a decision you can stop worrying about.
No. Forma is a concept project made for a design portfolio. The product, the people, the prices and every number on this page are fictional.
Your next test is one headline away.
Paste one script tag, pick what to change, and let your visitors decide. The first 14 days are free.