How long a test has to run before you can believe it
Sample size, duration, cycle effects, and what it really costs to call a winner a week too early.
The Visiopt Knowledge Center
A working library
for people who
optimize for a living.
Guides, frameworks, and original research on conversion optimization, testing, tracking, and customer psychology - written to be referenced, not scrolled.
How long should a test run?Which attribution model do I trust?What should I test first?Why isn't my page converting?Finding what blocks a decision on a page, and fixing it in the right order.
Designing tests that answer a question, and reading results without wishful thinking.
Why people hesitate, what earns trust, and how belief forms before a purchase.
Measuring what actually happened when the platforms disagree with each other.
Sample size, duration, cycle effects, and what it really costs to call a winner a week too early.
Where common attribution models disagree, why they disagree, and which one to trust for which decision.
Ranking ideas by reach, effort, and how much they actually change the decision a visitor is making.
What happened when a page that had already failed three A/B tests was tested as combinations instead.
A primary measure, defined before launch, is what stops a test from being reinterpreted after the fact.
Everything published here is written to stay useful — maintained, dated, and revised when the ground shifts. No news, no roundups, no reposted commentary. Original work on optimization from a team that does it at scale, organized so a question leads to an answer rather than an archive.
Research, frameworks, and case studies, sent when they're published.
No digest, no newsletter filler. Unsubscribe anytime.
See what is driving conversions, test what matters, and understand which changes are actually improving revenue.