About
A conversion studio for brands that cannot just discount.
LabsCodex exists because most premium ecommerce brands are optimised on opinion. We replace that with research, experiments run properly, and numbers reported in money.
Why we exist
Almost every premium brand we meet has the same shape of problem. The product is excellent. The photography is excellent. Paid acquisition is well run and getting more expensive every quarter. And the site converts below where it should, for reasons nobody has actually established.
So the team redesigns. A new product page, a new hero, a new checkout flow, each one a considered judgement call by capable people — and none of it measured, so nobody learns anything and the same argument returns next quarter.
We find out why buyers hesitate, test the fixes properly, and tell you what the programme was worth. Including the quarters where the honest answer is less than you hoped.
What we believe
Three opinions that shape every programme
- 01
Research is the product
Anyone can list conversion tactics. The value is knowing which of them applies to your buyers, your category and your price point — and that only comes from research. We spend the first month there, and it is why the test programme performs.
- 02
Most tests lose, and that's fine
Around three in four experiments fail to beat control. That is the normal shape of a healthy programme. The money is not in a high win rate — it is in never shipping the losers, and in learning something either way.
- 03
Premium is a constraint, not a licence
On a considered, high-price purchase you cannot borrow tactics from fast fashion. Urgency banners and discount popups convert a different buyer. We optimise within the brand rather than at its expense.
How we work
One month of research, then a programme that compounds
No tactics before evidence, and no test called before it is ready. The first month establishes where the revenue is leaking; everything after that is a cycle of proving what fixes it.
- 01Weeks 1–3
Research
We interview your team, work through analytics and session replay, run moderated tests with real buyers, and review the funnel end to end. No changes yet — first we find out where the money is leaking, and why.
You receive
A written research report, annotated funnel, and recorded user sessions - 02Week 4
Roadmap
Every finding becomes a testable hypothesis, scored on expected impact, confidence and effort, with a revenue estimate attached. You get a ranked backlog you could hand to any competent team — including your own.
You receive
A scored backlog with revenue estimates — yours to keep either way - 03Ongoing
Experiment
We build, QA and ship tests against that backlog, with the metric and sample size agreed before launch. Results are analysed properly and written up either way, whether the variant won or lost.
You receive
A written result for every test, including the ones that lose - 04Quarterly
Compound
Winners get rolled permanently into the theme, losers get documented so nobody retests them in eighteen months, and the whole programme is reviewed against revenue rather than percentage lifts.
You receive
A quarterly review of programme contribution, net of our fee
Capabilities
What sits behind the work
A deliberately narrow set of things we are genuinely good at, rather than a list of every platform we have logged into.
Research
- Quantitative funnel and cohort analysis
- Moderated user testing and on-site surveys
- Session replay and form analytics
- Heuristic UX, speed and accessibility review
Experimentation
- Hypothesis design and impact scoring
- Server-side and theme-native test builds
- Sequential and Bayesian analysis
- Programme reporting in revenue terms
Engineering
- Shopify Plus themes and Liquid
- Hydrogen and headless storefronts
- Checkout Extensibility and app integration
- Core Web Vitals and analytics instrumentation
How we operate
Four commitments we don't negotiate
They cost us some revenue and win us the clients we want.
Evidence before opinion
We do not ship changes because a competitor did it or because someone senior prefers it. If we cannot point to the research behind a hypothesis, it does not go on the roadmap.
We report the losers
Roughly three in four tests fail to beat control. An agency reporting a 90% win rate is either measuring badly or telling you what you want to hear, and both cost you more than the truth.
Measured in revenue
A lift on a micro-conversion is not a result. We report what the programme contributed in money, net of our fee, and we would rather that conversation be uncomfortable than vague.
You own everything
The research, the test code, the event layer and the analytics all live in your accounts. Ending an engagement with us should never mean losing what you have learned.
Work with us
Start with the research. Decide after that.
Most engagements start with a conversion audit: three to four weeks of research into where revenue leaks in your funnel, and a ranked roadmap of what to test first. The findings are yours to keep, whether you run them with us or not.
- Reply within one business day
- A strategist, not an SDR
- We'll tell you if you're not ready to test