#experimentation
12 pieces filed under “experimentation”.
Experiments that don't shift the layout
A/B testing without layout shift: server-side assignment, experiment flags that never leak into the DOM, metric integrity, and the true cost of visual editors.
A/B testing statistics for people who ship
Reading A/B tests honestly: minimum detectable effects, the peeking problem, novelty effects, segment traps, and the decision log that ends the arguments.
Share cards are a growth channel
Every shared link is an ad you didn't design. Running share cards like a channel: title testing, platform caches, and measuring dark-social lift honestly.
A creative-testing cadence that compounds
Creative is paid media's biggest lever and its worst managed. A weekly testing cadence, naming that preserves learnings, and the library that compounds.
Write the measurement plan before you build
Analytics bolted on after the build answers yesterday's questions badly. The measurement plan: events, naming, consent, dashboards — before the first sprint.
Experimenting on pricing pages without lying
You can test framing, proof and order on a pricing page. You can't charge different visitors different prices and keep their trust. Where the line sits.
Audit first, test second: heuristic CRO that finds the free wins
Before your next A/B test, walk the funnel as a stranger. Our heuristic CRO audit method: the friction walkthrough, severity scoring, and fixes that don't need stats.
Designing CRO experiments you can believe
Hypothesis quality, sample-size honesty, the peeking problem, change isolation versus big-bang redesigns, and documenting learnings so your programme compounds.
CRO experiments worth running: a field guide to honest testing
Most test backlogs are junk drawers. How we triage CRO ideas by evidence, opportunity and traffic tier — and sequence experiments so the learnings compound.
Cart cross-sells that don't read as desperation
Cross-sells pay when they answer one complementary question and shut up. Placement ranked by distraction cost, price-band logic, and measuring incrementality.
Landing page testing: velocity over genius
A landing page experiment programme that compounds: message-match first, structural swings second, a backlog scoring model that survives meetings, and kill criteria.
Attribution that admits what it doesn't know
Attribution is always lying a little. Model taxonomy, dark social acceptance, self-reported fields, geo holdouts and decision-making without false precision.