A Conversion Rate Optimisation Checklist That Is Not Just Button Colours
CRO has a reputation problem, mostly earned by a decade of button-colour case studies. Done properly it is the highest-leverage work in marketing, because a conversion improvement multiplies the return on every channel at once.
Find the leak before choosing the fix
Build a funnel view with the drop-off at each step: landing, product or service page, cart or enquiry, checkout or submission, completion. One step is always disproportionately bad.
Fixing the worst step first is not just efficient, it is the only way to know whether anything worked. Improving a step that was already performing gives you a change too small to measure.
Watch sessions before you form a theory
Analytics tells you where people leave. It never tells you why. An hour of session recordings on the leaking step will generate more real hypotheses than a week of dashboard analysis.
You are looking for hesitation: repeated scrolling between two sections, rage clicks on something that is not a link, form fields abandoned mid-entry, people opening shipping information and never coming back.
Prioritise with a simple score
Every team generates more test ideas than it can run, so the constraint is choosing. Score each idea from one to five on three axes and rank by the total:
- Impact — how much of the funnel does this touch, and how bad is the step now.
- Confidence — is there evidence for this, or is it a preference.
- Effort — how much engineering and design time it costs.
- Anything scoring low on confidence needs research before it needs a test.
Respect sample size or do not test at all
Most A/B tests reported as wins are noise. At 2% conversion and a target of detecting a 20% relative improvement, you need roughly 4,000 visitors per variant. Below that, you are looking at random variation and calling it insight.
If your traffic cannot support proper testing — and for many businesses it genuinely cannot — do not run A/B tests. Make well-reasoned changes based on research, measure before and after over equal periods, and accept the lower certainty. That is more honest than a test that cannot conclude.
Start with the changes that usually work
While building a testing programme, there are changes with a strong track record across most sites. They are not glamorous and they are worth doing before anything clever:
Reduce form fields to what you genuinely need. Show total cost including shipping early rather than at the final step. Add the guarantee or returns policy next to the buy button. Make the primary action visually dominant over everything else. Remove navigation from checkout. Show a progress indicator on multi-step forms.
Frequently asked
What is a good conversion rate?
It depends entirely on traffic source and category. Indian D2C ecommerce commonly sits between 1% and 2.5%; high-intent branded search can exceed 8%; cold paid social often runs under 1%. Compare against your own past performance and against the same traffic source, never against a published average.
How long should an A/B test run?
At minimum two full business cycles — usually two weeks — so weekday and weekend behaviour are both represented, and until each variant reaches the required sample size. Stopping early because a variant is ahead is the most common way to record a win that does not exist.