hello@craftberry.co
EU, EST 2015©
craftberry is a research-led Shopify A/B testing agency. We study your customers and store data to find what stops people buying, then design, build and run the AB tests in-house.
Every test is measured on revenue per visitor, and every testing program is backed by a performance guarantee.
Shopify partner trusted by great clients
Ongoing testing programs
5
Concluded tests
50
Winning tests
30
Win rate
60%
Enhance
+21% conversion from a single A/B test for a leading Tesla accessories brand
Migration to Shopify and scalable rebuild for a globally recognised smart automotive brand
Mi Amante
11.1% onversion uplift
How Mi Amante increased conversion by 11.1% with a data-led CRO program
Moon Nude
14% Conversion rate increase
Five CRO tests that increased conversion and revenue per visitor for Moon Nude
carista
29% Lower Bouce Rate
A custom Shopify build for Carista - unifying two web properties into one global storefront.
Shelly
500K+ customers moved to Shopify Plus for one of Europe’s standout smart-home unicorns
Transforming an IoT unicorn with a seamless migration from Shopware to Shopify Plus
Project 1 of 5: Enhance
4.8 on Clutch from 28 verified reviews.
“All of our deliverables were provided on time unless otherwise communicated, and their responses were prompt.”
“They don't feel like an agency, but rather part of a team that is just as invested in our company's success as we are.”
“They recognized that our business was unique and offered bespoke solutions rather than one-size-fits-all features.”
4x
Our target return in added revenue (a target, not part of the guarantee)
10%
Typical revenue-per-visitor lift we guarantee, agreed with you before the first test
90 days
From your first test launch
What we guarantee on every testing program
Before the first test, we agree your baseline revenue per visitor and a guaranteed lift, typically 10%. Intelligems measures it against a control group running at the same time, so seasonality doesn't skew the result.
If we haven't reached it 90 days after the first launch, we keep testing without a retainer fee until we do.
The guarantee requires stable traffic, winning variants shipped on schedule and a retainer of at least 90 days.
Brands that have tested before often tell us the same story: the tests were small, the results were hard to read, and the winners never made it into the theme. These are the parts we run differently.
Test ideas from your own customers
We start from your own customers: about 70% of our testing time goes into research, and every hypothesis links to the evidence behind it.
See how Signalveritas worksEvery test chosen on evidence
We rank every idea with PXL, so tests backed by research and quick to build run first.
Discuss your testing programBuilt and judged by one team
Our designers and developers build and QA every variant, and we judge results on revenue per visitor, split by device.
Discuss your testing programWinners shipped to your live theme
We build each winner into your live theme with the developers who coded the test. If you only need a testing tool, an app may be enough.
See our Shopify developmentWe reconcile orders and revenue across GA4, Shopify Analytics and Intelligems, so results can be trusted before anything launches.
Funnels, heatmaps, session recordings, polls, surveys, reviews and support chats show where and why shoppers drop off.
Each hypothesis names the problem, the change and the metric it should move, linked to its evidence.
PXL ranks ideas on how visible the change is, where it runs and how much evidence backs it, against build effort.
A UX designer creates the variant, our developers code it in your theme, and QA checks it across devices.
We call a winner at 95% confidence in Intelligems, after at least two full business cycles.
Winners go into your live theme, and every result feeds the test log behind the next hypothesis.
Discuss your testing programOne team for the whole testing cycle.
Winning tests from our late 2025 and 2026 programs, grouped by store area. Lifts measured in Intelligems against the control.
Every test we run starts from evidence Signalveritas finds across your reviews, session recordings, analytics and surveys.
Shopify's built-in split testing is called Rollouts. It runs Experiments from the admin on the Grow plan or higher and is a low-cost way to start AB testing simple theme and offer changes.
Rollouts can test changes to:
Rollouts tests whatever your team picks. Our tests start from customer research, and if Rollouts can handle an idea, we'll say so.
Liquid templates can't be changed in a rollout, so variants like a new product finder or cart drawer need our developers.
We run them in Intelligems and read the results on profit per visitor.
The developers who coded the test build the winner into your live theme.
Testing
Our testing platform and partner, for code-level, price, shipping and offer tests, with revenue and profit per visitor for each variant.
Research
Signalveritas
Our in-house research tool, combining up to eight sources per store, from reviews and surveys to analytics.
Behavior
Microsoft Clarity · Hotjar · Mouseflow
Heatmaps and recordings in the tool your store already uses. We watch them during research and while a variant is live.
Analytics · GA4 & Shopify Analytics
Our cross-check on every result, by device, channel and customer type, reconciled with Intelligems.
We help brands grow through connected strategy, technology, creative, and marketing.
Discuss your testing programShare your Shopify store and monthly order volume. We'll reply within one business day on whether an A/B testing program fits your store and what the next steps could be.
Shopify partner trusted by great clients
