Skip to main content

Your First A/B Experiment

You can set up and launch an A/B test in under 5 minutes. No developer needed, no code to write.

Step 1 — Create an experiment​

  1. Go to A/B Tests in your dashboard
  2. Click New experiment
  3. Give it a name like "Homepage CTA — Start free vs Get started"
  4. Enter the URL of the page you want to test (e.g. /pricing or /)

Step 2 — Make your change in the visual editor​

Click Open visual editor. Your site loads inside Datsnip with a click-to-edit overlay.

  • Click any element — headline, button, image, paragraph
  • Edit it — change the text, color, size, or style
  • Save the change

You don't need to know CSS or HTML. Just click and type.

Don't know what to test first? Start with your main call-to-action button. Change the text.

  • "Sign up free" → "Start my free trial"
  • "Get started" → "Try it free — no card needed"
  • "Buy now" → "Add to bag"

Small wording changes on your primary CTA consistently produce 10–25% lifts.

Step 3 — Choose a goal​

What counts as a win? Pick the action you care about:

  • Click — visitor clicks a specific element (your CTA button, a link, a form submit)
  • Page visit — visitor reaches a specific page (your thank-you page, dashboard, checkout confirmation)
  • Purchase / custom action — visitor completes a custom event you track

For your first experiment, pick the simplest one: click on your CTA button.

Step 4 — Launch​

Click Start experiment. Immediately, Datsnip splits your visitors 50/50:

  • Half see your original page (Control)
  • Half see your changed version (Variant A)

The same visitor always sees the same version — no jumping.

Step 5 — Read your results​

Come back in a few days. Go to your experiment → Results.

You'll see:

  • P(win) — the probability your variant is actually better. Once this hits 95%+, you have a winner.
  • Uplift — how much better the variant performed (e.g. "+18%")
  • Visitors — a progress bar showing how close you are to a reliable result
How long to run it?

At minimum, 7 days. This captures a full week of traffic — weekdays and weekends have different conversion patterns. Datsnip will warn you if you try to call a winner too early.

What to do when you have a winner​

If Variant A wins, you have two options:

  1. Roll Out Winner — Datsnip applies the winning change as a permanent page fix, so you don't need to touch your CMS
  2. Update your site manually — edit the original in your CMS and stop the experiment

Either way: ship it, then run the next experiment.

Ideas for your first few experiments​

  • Headline — benefit-focused vs feature-focused copy
  • CTA button — text, color, or size
  • Hero image — product shot vs lifestyle photo
  • Pricing page — annual vs monthly default, which plan is highlighted
  • Form — fewer fields vs more fields

More experiment ideas by industry →