Sample size and duration fixed before launch
Calculated from your baseline rate and the smallest effect worth detecting. Stopping when a result looks good is the single largest source of fake wins in this discipline.
CRO & Landing Pages
A/B Testing Programme
A running experimentation programme: hypotheses from the audit, sample sizes calculated before launch, and results recorded whichever way they go.
Free proposal · No contracts over 30 days · Reply within one business day
- Confidence threshold, fixed before launch
- 95%Confidence threshold, fixed before launch
- Full business cycles minimum per test
- 2Full business cycles minimum per test
- Results documented, including the losses
- 100%Results documented, including the losses
The problem
Most testing programmes produce false wins. The test is stopped the day it looks significant, four metrics are checked until one moves, and the "27% uplift" does not survive into the quarterly numbers.
How we do it
What the work actually involves
One primary metric, declared in advance
Secondary metrics are monitored for harm, not for a win. Checking enough numbers guarantees one of them will look significant by chance.
Minimum two full business cycles
Traffic behaves differently by day of week. A test run Tuesday to Friday measures Tuesday-to-Friday visitors, not your customers.
Losers documented as carefully as winners
A negative result is information about your customers and it is worth as much as a win. Without a record, teams re-test the same idea every eighteen months.
Included
What you actually receive
- Prioritised hypothesis backlog
- Power calculation and duration per test
- Variant design and build
- QA across devices before launch
- Analysis and readout with the decision recorded
- Cumulative learning log
Tools we use
- GA4
- VWO
- Optimizely
- Statsig
- Hotjar
Questions
Any question in your mind?
The questions we get asked most, answered without the sales gloss. If yours is not here, ask us directly — you will get a straight answer either way.
Still have questions?
Book 30 minutes with a strategist. No deck, no pitch — just answers about your situation.
Book a callUsually two to six weeks. It is determined by your traffic and conversion volume, not by preference — a test that reaches significance in four days almost certainly had a sample size problem rather than an enormous effect.
Looking for pricing, contract terms or timelines?
Read all FAQsKeep reading
More cro & landing pages work
- Conversion Rate AuditWhere visitors leave and why, from recordings, form analytics and funnel data — ranked by revenue at stake.Read
- Landing Page DesignCampaign landing pages built to convert paid traffic, with message match to the ad and a test plan from day one.Read
- GA4 Setup & TrackingAnalytics you can act on: events named to a plan, conversions defined as revenue, consent handled properly.Read
Ready when you are
Ready to Build Your Next Big Project?
Tell us what you are trying to grow. You will get a straight answer on whether we can help, and what it would cost, within one business day.