Skip to main content
Start for FreeBook a demo

A/B test your store's pages with Replo

Experiments splits visitors between two versions of a page on one link, then ships the version that sells.

Start a testCheck a test result

How a Replo test runs

Every test on a Replo site works the same way.

  1. Pick a goal

    Choose what counts as a win: a visit to a page, or a click on a button.

  2. Split on one link

    Replo picks a version before the page loads, so the URL stays the same and nothing flickers.

  3. Wait for a real answer

    The result reads “Too early” until 100 visitors, 10 conversions and 7 days.

  4. Ship the winner

    The winning version goes live on your next publish, at the same link.

Read the result next to your sales

  • One report per test

    Each test's result lands in Insights, beside your traffic, sales and other reports.

    See Insights

  • A 95% bar to win

    A version is called the winner once it has a 95% chance to beat the original.

    Check the math

  • Heatmaps for each version

    See where shoppers clicked on each version, not only which one sold more.

    See heatmaps

Where your page lives decides the test

Replo has two kinds of pages, and each has its own way to test.

  • Pages in your Shopify theme

    Test in Theme Builder, live today

    Theme Builder, Replo's Shopify page builder, has A/B testing built in for theme pages.

    • Variants and the split set up in one app
    • Choose how much traffic each version gets
    See Theme Builder

  • Sites the Replo agent builds

    Test on one link with Experiments

    Tests on sites Replo builds and hosts run on the page's own link.

    • One URL, no redirect or flicker
    • Results in Insights
    • The winner ships on your next publish
    Start a test

What to test first

Start closest to the purchase, and change one thing per test so you know what moved the number.

  • The headline your ad promised

    Match the hook that got the click, then test it against your current headline.

  • The offer

    A discount, a bundle or free shipping, stated above the fold in plain words.

  • The hero image

    The product in use against the product alone, at the size phones show.

  • The buy box

    Price, reviews and the button, in the order shoppers read them.

Plan and read your next test

Significance calculatorCheck whether a result is real, and how many visitors you need.A/B test ideas for storesTests you can launch on a store this week, from headline to buy box.How to A/B test landing pagesThe method, from hypothesis to calling the winner.HeatmapsSee where shoppers click before you decide what to test.

Before you get started

  • Yes, in two places. Experiments runs A/B tests on the sites Replo builds and hosts, on one link, with results in Insights. For pages in your Shopify theme, Theme Builder has its own A/B testing built in.

  • On one link. Replo picks a version for each visitor before the page loads, so the URL never changes and nothing flickers. You choose the goal, a page visit or a button click, and the result shows each version's chance to win.

  • Until each version has enough visitors to detect the change you care about, and at least a full week, so weekdays and weekends both count. Plan the sample size first with the significance calculator.

  • The parts closest to the purchase: the headline your ad promised, the offer, the hero image and the buy box. Change one thing per test, so you know what moved the number.

  • Yes. VWO and Intelligems still work with Replo: connect either one to your project and run tests there on Replo pages. Experiments runs the same kind of test inside Replo, with results next to your sales.

Keep reading