Glossary

What is Experiment Design?

The structured planning of A/B tests and experiments, including hypothesis formation, variable selection, sample size calculation, and success criteria definition.

In Depth

Understanding the Details

Experiment design is what separates rigorous testing from random guessing. Before running a test, proper design requires: a clear hypothesis (what you believe will happen and why), defined variables (what exactly you're changing), success metrics (what improvement you're measuring), sample size calculation (how many visitors you need), test duration (how long to run), and decision criteria (what result means you implement the change). Without this structure, teams run tests that are too small, measure the wrong things, or declare winners based on noise. The best experiment programmes maintain a backlog of prioritised test ideas, document results systematically, and build institutional knowledge about what drives conversions.

Examples

How It Works in Practice

Structured hypothesis

Hypothesis: Changing the CTA from 'Start Free Trial' to 'See It In Action' will increase demo requests by 15% because it reduces commitment anxiety.

Pre-registered analysis

Before the test launches, the team documents the primary metric, minimum detectable effect, sample size, and test duration — preventing post-hoc rationalisation.

Test prioritisation

A PIE framework (Potential, Importance, Ease) ranks 20 test ideas, ensuring the team works on the highest-impact experiments first.

Importance

Why It Matters

Rigorous experiment design maximises the learning from each test. In a world where traffic and time are finite, testing the right things in the right way matters enormously.

Misconceptions

What People Often Get Wrong

Just testing anything is valuable. Actually, poorly designed tests waste traffic and can lead to incorrect conclusions.

Experiment design is overkill for marketing tests. Actually, even simple A/B tests benefit from clear hypotheses and sample size planning.

Failed tests are wasted. Actually, tests that disprove hypotheses are equally valuable — they prevent implementing changes that don't work.

Our Approach

How We Handle Experiment Design

We approach testing with structured experiment design, ensuring every test has a clear hypothesis, appropriate statistical power, and documented learnings.

FAQ

Common Questions

Need Help With Experiment Design?

If you'd like to discuss how experiment design applies to your business, we're happy to explain further.