How conversion rate optimisation works
It is a loop rather than a project. You measure what people currently do, find where they stop, form an explanation of why, change one thing to address it, and measure again. The evidence comes from several places: analytics for where the drop-off sits, recordings and heatmaps for what happens on the page itself, form analytics for the field that loses people, and — most usefully and least often — conversations with customers and with whoever answers the phone about what nearly stopped them buying.
The change is then validated. With enough traffic that means a controlled test. Without it, it means shipping a well-reasoned improvement, keeping other changes out of the way, and watching the trend over a longer period before deciding it worked.
Why conversion rate optimisation matters
Buying more traffic has a rising cost, because the cheapest audiences are exhausted first and each additional visitor is dearer than the last. Improving the rate at which existing visitors act does not behave that way: the gain applies to every channel at once, and it does not stop when the ad budget does.
The second benefit is what you learn. Working out why people hesitate produces better ad copy, better sales conversations and a clearer statement of what the business actually sells — which often turns out to be worth more than the lift itself.
Common mistakes with conversion rate optimisation
Treating it as a list of tricks is the first: adding a countdown, a popup and a badge because they worked on somebody else’s site, without knowing what stops your own visitors. The second is redesigning everything at once, which moves the number without teaching you anything and leaves you unable to undo the part that did harm.
The third is optimising the wrong page. Effort spent on a page with little traffic or little commercial value cannot pay back, however clever the change. The fourth is measuring the enquiry and never the outcome, so a change that produces more forms and fewer customers is filed as a success.
How to act on it
Start where the money and the traffic already are: the main service pages, the checkout, the enquiry form. Rank the problems you find by how many people they affect and how hard they are to fix, and take the cheap, wide ones first — loading speed on a mid-range phone, a form asking for more than you need, an unclear next step, a price question the page never answers.
Remove the friction you can already see before reaching for A/B testing, which needs volume to reach a conclusion and is wasted on problems that are obvious to any stranger using the site. Keep a written log of every change and its date, so that a shift in performance months later can be traced to a cause. Run as a continuing programme of conversion rate optimisation, the loop keeps improving; run as an occasional burst of redesign, it mostly produces opinions.