Přejít na obsah
Peak Atlas

Peak Pages · A/B testing · Demo data

Two variants. One reliable answer.
Instead of a gut feeling.

Test page variants against each other. Each visitor always sees the same variant. Significance and the decision run through Peak Experiments, and you promote the winner yourself.

Prepare an A/B test in Peak Pages

Demo data · fictional numbers · no real customer results

DEMO DATA / A/B TEST

One example value.
The fitting next step.

Peak Pages / Demo testDEMO DATA

01 / Test state

How far along is the test? (runtime 0 to 10 of 10)

02 / What it meansA: 9 leads from 80 visits, B: 12 leads from 80 visits. Not significant.

Keep it running · no winner

Switch between an early and a late test state. All numbers are demo data.
Demo data · no real test

One example. Two paths.
One setting
changes the page.Compare both paths below ↓

Two test states · demo data

How far along is the test? (runtime 0 to 10 of 10)

Demo data: invented numbers. In the product Peak Experiments calculates significance, not this page.

2/ 10Early state
Early: keep it runningEXAMPLE PATH
A: 9 leads from 80 visits, B: 12 leads from 80 visits. Not significant.
Then

The test is complete. You read the result in Peak Experiments.

Keep it running · no winner

10/ 10Late state
Late: read the resultEXAMPLE PATH
The test is complete. You read the result in Peak Experiments.

You decide · publish the better variant yourself

Trust needs clarity

Honest about tests.

01 / The existing evidence

Demo data

The numbers above are invented and show no real customer and no real winner.

Peak Pages / Demo test statesDEMO DATA
  1. 2/10
    A: 9 leads from 80 visits, B: 12 leads from 80 visits. Not significant.Keep it running · no winner
  2. 10/10
    The test is complete. You read the result in Peak Experiments.You decide · publish the better variant yourself

02 / Your check

No winner without significance

Peak Pages does not declare a winner just because one variant is ahead. The result comes from Peak Experiments.

4 / 1010 / 10

Compare both paths and their reasons.

03 / The clear limit

You decide

Peak does not apply even a significant result on its own. You publish the winner.

DemoNothing published

The next step

Improve what demonstrably works.

Start Peak Pages, add a second variant and let Peak Experiments check the result.

Prepare an A/B test in Peak Pages
Peak Pages · A/B testing · Demo dataOne change.

Demo data · fictional numbers · no real customer results

Frequently asked questions

Does a visitor always see the same variant?

Yes. The server assigns each visitor to one variant for good.

Who decides on the winner?

You do. Peak Experiments calculates significance; nothing is applied on its own.

Are the test numbers here real?

No, they are demo data with no real customer and no real winner.

More in Peak Pages