Refract Group2026
Glossary
Conversion rate optimisation (CRO)
Conversion rate optimisation (CRO) is the work of raising the share of visitors who take the action a page exists for, such as booking a call or starting a trial, by changing the offer, the page and the form, and measuring the result.
- Form submitted (40)4.0%
- Qualified enquiry (15)1.5%
- Sales opportunity (6)0.6%
Conversion rate optimisation in detail
- Also called
- CRO, Conversion optimisation, Conversion rate optimization
- Topic
- Growth
- Updated
What it is
Conversion rate is the number of conversions divided by the number of visitors or sessions in the same period, as a percentage. Google Analytics 4 calls a conversion a key event and reports the session key event rate: sessions in which a key event happened, divided by all sessions. A page with 1,500 sessions and 30 booked calls converts at 2%.
What counts as a conversion is decided before anything is measured. On a B2B site it is usually an enquiry, a booked call, a demo request or a trial started. A newsletter signup or a whitepaper download is a smaller step and belongs in its own column.
CRO works on what sits between the visit and that action: the offer, how clearly the page explains it, the evidence beside each claim and the form. Changes are split-tested where traffic allows, and compared with the period before where it does not.
Why it matters
Traffic usually costs more than the page it lands on. If a demo page converts at 1% and could convert at 2%, the same ad budget books twice the calls, and every channel pointed at that page gains at once. The work also shows when the page is not the problem, because no rewrite fixes visitors who are the wrong buyers. Measured against the wrong event, CRO does harm, filling the calendar with people sales will never close.
How we do it
Conversion work is part of our growth audits and our website projects. First we agree which action counts, such as a booked call or a demo request, and check that the form can be finished and the event is recorded. Nothing is tested before that. Then we read each page a buyer lands on, on a phone as well, and mark every line where the story gets lost. Each audience gets a task, and we follow it through the live site to where the route breaks.
You get the fixes ranked by effort and effect, with the first ones drafted. Repairs are kept apart from ideas that need a test, and each idea names the measure that would show it worked.
An example
Say a B2B analytics product gets 1,500 sessions a month on its demo page and 30 booked calls. The team wants to test a green button. Reading the page first shows the headline never says who the product is for, and the form asks for budget before an email address. The rewrite names the buyer and the job, puts a customer’s numbers beside the main claim and cuts the form from nine fields to four. Booked calls over the next eight weeks are compared with the eight before, with ad targeting left alone.
Common mistakes
- Counting the easiest event. A rate built on newsletter signups or button clicks can rise while booked calls fall, so count the action the business needs.
- Calling a result on a few conversions. With a handful of conversions a period the rate swings by chance, so a jump from 10 calls to 13 tells you little until it holds over a longer run.
- Starting with the button. A visitor weighs whether they understood the offer and believed it long before they notice the button’s colour, so work on the offer and the page first.
- Ignoring what changed upstream. A new campaign or a change in targeting can move the rate while the page stays identical, so compare like audiences and periods.
Conversion rate optimisation vs A/B testing
Conversion rate optimisation is the whole job of raising a page’s conversion rate, and A/B testing is one method within it: two versions shown to split traffic at the same time, then compared.
| Conversion rate optimisation | A/B testing | |
|---|---|---|
| Scope | Offer, page, form and measurement | One comparison between two versions |
| Traffic needed | Any; low traffic uses before and after | Enough conversions per version to judge |
| Answers | What stops buyers acting | Which of two versions did better |
Questions
What people ask about conversion rate optimisation.
Divide the number of conversions by the number of visitors or sessions in the same period and multiply by 100. A demo page with 1,500 sessions and 30 booked calls has a conversion rate of 2%. Choose one way to count, per session or per user, once a person or every time, and keep to it.
There is no single benchmark for a B2B conversion rate, because it depends on what you count and who arrives. Nielsen Norman Group puts typical rates between 1% and 10%. The useful comparison is with your own page in an earlier period, measured the same way, and with what happens to those enquiries afterwards.
Conversion rate optimisation has no fixed traffic threshold: it depends on the starting rate and the difference you need to see. A site with a few conversions a month gets more from clear fixes to the offer, the page and the form, judged over a longer period, than from split tests that cannot reach a result.
Where this shows up in our work
Related terms
Working on this now?
Tell us what you’re building and where it’s stuck. We reply within 12 hours.
Most teams start with a call, then a sprint or an audit.
The people you’d work with, from the first call to launch