Get in touch

More sales from the visitors you already have

I go through your store the way a customer would, and then as a developer. Where it loses buyers I rebuild that part, test it against the current version and keep it only if it wins.

Get in touch
A/B test · product page example
A Current
B Won the test
B won and was built into the theme.
Fabian Genell · Shopify developer

How an A/B test works

  1. Hypothesis I write down what changes, for whom and what we think will happen, before anything gets built.
  2. Build I build the variant as code directly in the theme, so the page loads without flickering.
  3. Split Half the visitors see the current page, half the new one. I run one test at a time on each page.
  4. Result The main measure is purchases per visitor. I also follow average order and returns so nothing else gets worse.
  5. Decision If the new version wins I build it into the theme. Otherwise I remove it the same day.

Where your store loses buyers

The steps are the same in every store. I look at where yours loses the most people and start there. The percentages here are an example.

Continue, step by step
Drop-off
Built
Tap a step to see what happens there.

How it went at three clients

The Turn Me Royal home page
Cart · Plus Turn Me Royal A new cart with upsells, A/B tested against the old one. It raised average order value, and load time went from 9 s to 2–3 s.
The Dailyride home page
Product page · Plus Dailyride Configurator with add-ons, price logic in Shopify Functions and a new menu. The upsells raised average order value.
The Padel Pro Shop home page
Product choice · Plus Padel Pro Shop A quiz that helps the customer pick a racket, built as sections in the theme editor, plus countdowns on metaobjects.
See all cases

The winner is built into the theme

When a test has won I build the variant into the theme as regular code. The store loads faster, and your team can change it in the theme editor. The seconds below are an example.

Added as an app common
theme.js
app-a.js
app-b.js
test-tool.js
jquery.js
Page visible 3.8 s
The script loads for every visitor, even after the test is doneThe page jumps when the variant overwrites the originalYour team cannot change the copy without asking someone
Built into the theme how I do it
theme.liquid
critical css
Page visible 1.2 s
Clean Liquid and code in your theme, no extra script leftThe page is finished right away, average LCP under 1.3 s in the stores I have builtSections and blocks your team changes in the theme editor

Common questions about CRO

How much traffic is needed for A/B tests?
Enough for a test to give a reliable answer within a few weeks. If the store has less traffic I test fewer things.
Which tools do you use?
I use Simple Split Testing for the split itself, Shopify analytics and a custom analytics app for the readout.
Are ads and email included?
No. My work starts once the visitor has landed in the store.

You get a list of what I would test first

Send the store and write what you are wondering about. Afterwards I send you the list, with why I would start there.

  1. 01 Write a few lines about the store.
  2. 02 I go through the buying flow myself.
  3. 03 Afterwards I send you the list.

Send the store address and I will reply within two working days with three things I would test first.

Get in touch →