Stage 2 · Testing and CRO
More tests that move take rate and revenue per visitor.
CRO used to spend most of its time on design and development. Our engineers and PMD Cortex AI have sped that part up, so our time goes on shipping more and better tests, pages and offers that let you scale paid spend.
- 01Research
- 02Strategy
- 03Copy and design
- 04Build and QA
- 05Test
- 06Implement
Where the time goes
Less time building. More time testing.
Our engineers each run a bench of AI bots that handle the design and build legwork. An engineer checks every output, so the hours you pay for go on deciding what to test and shipping it.
How CRO used to run
Most of the month went on making things
- Most of the hours spent designing and developing each new page or variant
- Few tests live at once, because building was the bottleneck
- Results judged on conversion rate alone
- Learnings kept in slide decks and people's heads
How we run it now
Most of the month goes on what to test
- Bots draft the design and build work, engineers review and ship it
- More tests, pages and offers live in the same time
- Every test judged on subscription take rate and revenue per visitor
- Every result feeds the CRO Brain inside PMD Cortex AI
The numbers we answer to
Two metrics decide if a test wins.
Subscription take rate
The share of first orders that start a subscription. Lift it and each customer you pay to acquire is worth more over their lifetime, so you can afford to bid harder on ads.
Revenue per visitor
Conversion rate and order value in one number. A test that lifts conversion but cuts order value does not count as a win here.
Real tests
Control against winner, side by side.
Four tests from our case studies. Pick one to see what changed and why it won.
Natrava: subscription upgrade in the cart
The upgrade block spells out the saving, free shipping on the first subscription order and that customers can modify, skip or cancel, with one Upgrade and Save action.
52,421 visitors over two weeks, about 85% reported probability of being best.
Read the Natrava case study
MyoMaster: a cart for a four-figure basket
On a four-figure basket the cart is a second product page. Clearer product cards, delivery messaging for each product and the saving made visible before checkout.
Gymproluxe: make the value impossible to miss
Instead of a list of free gifts, the buy box explains what each bonus is worth and how it helps a beginner get started, answering the price worry customers raised in surveys.
Avantera: show the saving in money
The one-time price now sits next to the subscription price, so the saving is visible at the buying decision. It set the pattern for how savings are shown across the whole funnel.
More test results
One test can move the whole number.
Single tests from our case studies, rounded and checked against each study.
Where your budget goes
One budget, six kinds of work.
What your testing programme pays for, in the order a test gets made.
One programme budget pays for all six. Our build bots now do most of the design and development legwork, so more of your budget goes on the research, copy and tests that grow revenue. We agree the split with you in your audit.
Your pod
Who runs your tests.
One pod on your account, from the first piece of research to the rollout of every winner.
Paddy McLarnonFounder. Takes the first call with every brand and stays close to your strategy.
-
CRO strategist
Finds where your funnel loses money and decides what we test next.
Kirsty, Veselina, Sasha, Ammiel
-
Copywriter
Writes the angles, headlines and offers we put in front of your customers.
-
Designer
Designs every page and test variant in Figma, built around your brand.
Joshua, Yana, Oksana, Mirian
-
Shopify developer
Builds the pages and tests on your theme, with our build bots doing the legwork.
Haziq, Yuri, Rahul, Anil
-
QA specialist
Checks every build on every device before it reaches your customers.
Ali, Claudiu
-
Project manager
Your day to day contact. Runs the plan, the queue and the weekly updates.
Isabel
Plus Paul and Omer, who build and run the bots inside PMD Cortex AI, so your pod spends its time on output, not legwork.
Our own testing app
We built AB Genius to run tests inside Shopify.
Price, theme, template and checkout tests inside your Shopify store, reported on revenue and profit. Our pods use it on client stores, alongside Intelligems where a brand already runs it.
- Tests read on revenue per visitor and profit, not clicks
- Built and run by our own team
- Every result, win or loss, goes back into the CRO Brain
What feeds the CRO Brain
Every test starts with what customers say, see and click.
One test cycle
How a test goes from idea to implemented.
- 01
Research
Surveys, reviews, support tickets, cancellation reasons, heatmaps and recordings.
- 02
Strategy
A hypothesis tied to take rate or revenue per visitor, ranked by what it could be worth.
- 03
Copy and design
The variant, written in your customers' words and designed in your brand system.
- 04
Build and QA
Built by our bots, checked by an engineer on every device before launch.
- 05
Test
Run in AB Genius or Intelligems until the result is clear enough to act on.
- 06
Implement
Winners go live, and the learning goes back into the CRO Brain for the next test.