Conversion Rate Optimisation
Improving what your existing traffic does: research to find the real friction, tests run properly, and honesty about when you do not have the volume to test.
Overview
Conversion work has a reputation for button colours because a lot of it deserves that reputation. Testing arbitrary changes produces a long list of inconclusive results and occasional false positives that get celebrated and never replicate.
The version that works starts with research: session recordings, form analytics, support tickets and a few user tests to find where people actually struggle. Then you test the changes that address what you found, with the sample size worked out before launch rather than the result called when it looks good.
Who this is for
- Companies with healthy traffic and disappointing conversion
- Teams running tests that never reach a conclusion
- Businesses about to buy more traffic rather than convert what they have
How we approach conversion rate optimisation
The specific pieces of work a typical engagement covers. Scope is agreed up front — nothing here is a surprise line item later.
Conversion research
Analytics, session recordings, form analytics and user testing to find the real friction rather than guessing at it.
Hypothesis and prioritisation
A ranked queue of tests, each with a stated hypothesis and an expected effect, so the order is defensible.
Test design
Sample size and duration calculated in advance, with a decision rule agreed before anyone sees the data.
Statistical discipline
Tests run to completion rather than stopped when they look favourable, which is how most false positives are manufactured.
Qualitative testing
Watching people attempt the task, which finds problems no amount of quantitative data will surface.
Documented learning
A record of what was tested and what happened, including the failures, so the programme accumulates knowledge.
From first call to measured result
The same sequence every time, so you always know what happens next.
Research
Find where people struggle, using both behavioural data and watching real users.
Prioritise
Rank hypotheses by expected impact against effort, and agree the queue.
Test
Run each test to its predetermined sample size and call it against the agreed rule.
Implement and learn
Ship the winners, record everything including the losses, and feed it into the next round.
Outcomes, not deliverables
A pile of artefacts isn't progress. These are the changes the work is meant to produce — and what we report against.
More from traffic you already buy
A conversion gain applies to every visitor, which is why it usually beats buying more of them.
Results that hold
Proper sample sizes and stopping rules mean a winning test still wins after it is rolled out.
Research that keeps giving
The friction found in research informs the product and support as well as the tests.
A programme, not a list of tricks
Documented learning compounds, so the tenth round is better targeted than the first.
Common questions about Conversion Rate Optimisation
The things people ask before they get in touch. If yours is not here, ask us directly.
How much traffic do we need for A/B testing?
Enough for a few hundred conversions a month per variant to detect a realistic effect. Below that, tests take months and most finish inconclusive. We will say plainly when you are under that threshold and work through research and established practice instead.
Why do so many winning tests not hold up?
Usually because they were stopped early when they looked good. An A/B test checked daily and called on the first favourable day will produce false positives reliably. Fixing the sample size and stopping rule in advance is most of the remedy.
What should we test first?
Whatever the research says is costing the most, which is rarely the thing anyone expected. Often it is a form field, an unexpected cost, or a missing piece of information rather than anything about layout.
How long does a test need to run?
At least two full weeks regardless of sample size, so weekday and weekend behaviour are both represented, and then until the predetermined sample is reached.
Thinking about Conversion Rate Optimisation?
Tell us what you are trying to change. If we are not the right fit we will say so, and point you somewhere better.
Looking at the wider picture?
Conversion Rate Optimisation usually sits alongside other work in E-commerce & Revenue Enablement. Browse the full area to see what it connects to.