What do you A/B test on Shopify?

Every key page: your homepage, product pages, collections, menu, cart and checkout. That means headlines, offers and bundles, product page sections, layouts, carts and upsells, each one designed from what your customers already say about you.

Which A/B testing tool do you use?

Tests run in Intelligems or Shoplift, each on a 50/50 split. Every variant is built natively in your theme, so the test looks and loads like the rest of your store.

How do you decide if a test won?

We judge every test on revenue per session. It counts conversion rate and average order value together, so a test that gets more orders but smaller ones can’t pass as a win.

How long does a test run?

Long enough for a clear result, and that depends on how much traffic the page gets. When a test ends, the next one is already built and queued, so the page goes straight into it and is never left without a test.

How much traffic do I need?

Enough for tests to reach a clear result in a sensible time. The more sessions a page gets, the faster it can be tested. On the call we look at your store and tell you honestly which pages are worth testing first.

How many tests do you run a month?

There’s no cap. A test is live on every key page 95% of the time, and the next 2 to 3 are always built and waiting behind it.

What happens when a test loses?

It still teaches us something, and that lesson goes into the next design. A losing test tells you what your customers don’t respond to, before you roll it out to everyone.

Shopify A/B testing with a test live on every key page.

We research, design, build and run your tests end to end. 2 to 3 always queued, 10 new Figma designs every Monday, and every test judged on revenue per session.

Book a free call How a test runs

+0%Revenue per session
Animated mechanism section

+0%Revenue per session
Dynamic subscription cart

+0%Revenue per session
Dynamic cart upsells

+0%Revenue per session
Offer builder with free gifts

How a test runs

From your customers’ words to a live test.

The same six steps behind every test, on every key page.

01Start from the data bank.Every test starts with what your customers already say about you: reviews, tickets, Reddit threads, ad comments and your Shopify data.
02Design it in Figma, in your brand.10 new test designs land every Monday.
03Score it for brand before it’s built.Every design gets a balance score and has to land inside its page’s target, so no test ever makes the brand look cheap.
04Build it natively in your theme.No page builders and no apps.
05Run it on a 50/50 split.Tests run in Intelligems or Shoplift, with half your visitors on the current page and half on the new design.
06Judge it, then go straight to the next.Every test is judged on revenue per session.

Revenue per session

The one number we judge by.

Conversion rate can lie. Revenue per session can’t.

Revenue÷Sessions=RPS

Revenue per session lift, winning tests

Animated mechanism section+12%

Dynamic subscription cart+12%

Dynamic cart upsells+11%

Offer builder with free gifts+7%

Every test judged on revenue per session.

Revenue per session counts conversion rate and average order value together. A test that wins more orders but smaller ones can’t pass as a win, and a test that grows the basket gets the credit it’s due.

Revenue divided by the sessions that saw each version
Catches offers and carts that change what people spend
The number behind every result we quote

Read why we use it

Uptime

Never a page without a test.

A page without a live test isn’t learning anything. So the next tests are always built, queued and waiting.

01A test on every key page.Homepage, product pages, collections, menu and cart, each with a test live 95% of the time.
02Design 10 tests a week.10 new test designs in Figma, on your desk every Monday.
03Keep 3 tests queued at all times.The next 2 to 3 tests are built before the live one ends, and the next goes live that same day.

Our work