Conversion rate is the percentage of visits, visitors or ad clicks that end in a chosen action, such as an enquiry, a booking or a sale. If 1,200 sessions on your site produce 30 enquiries, the conversion rate is 30 divided by 1,200, which is 2.5%.
How conversion rate works
Every conversion rate has two parts, and both need defining before the number means anything:
- The top figure: which action counts as a conversion. A purchase, a form submission, a phone call, a brochure download?
- The bottom figure: what you divide by. Sessions, users, ad clicks or visitors to one particular page?
Different tools choose differently. In GA4, actions you mark as important are called key events (Google renamed them from conversions in 2024), and GA4 reports a session key event rate and a user key event rate. Google Ads divides conversions by ad interactions, usually clicks. Meta reports results against clicks or impressions depending on the column. Your shop platform may divide orders by its own count of visits. So the same week can show four different conversion rates, all correct by their own definitions.
On UK sites there is another wrinkle. With a PECR-compliant cookie banner, analytics only fully records visitors who accept cookies, plus whatever modelled data your tools add. Your conversion rate is calculated on a partial sample, and if the banner, the consent setting or the modelling changes, the rate can move without any real change in behaviour.
Why it matters
Conversion rate connects traffic to outcomes. Combined with cost per click, it sets what each lead or sale costs you. At £2 a click and a 4% conversion rate, each enquiry costs £50; lift the rate to 5% and the same enquiry costs £40, with no change to the ads. That is why improving the page can matter as much as improving the campaign.
It also helps you find problems. A page with plenty of traffic and a far lower rate than similar pages is usually the best place to start looking.
Common mistakes
- Comparing rates from different tools as if they measured the same thing.
- Blending all traffic together. Branded searches convert far better than cold social traffic, so a sitewide average hides what each channel is doing.
- Counting small actions as conversions, such as page scrolls or video plays, which inflates the rate. Keep these as micro-conversions, reported separately.
- Chasing the rate instead of the result. Removing all qualifying questions from a form may raise the rate while filling your inbox with leads you cannot help.
- Reading too much into small numbers. Three enquiries from 60 visits is 5%; four would be nearly 7%. That is noise, not a trend.
- Borrowing benchmarks from other industries, price points or countries.
How to act on it
Write down which actions count as conversions and which denominator you use, then stick to it so month-on-month comparisons are fair. Split the rate by channel, device and landing page; the overall figure rarely tells you what to fix. Note in your reports when the consent banner or tracking set-up changes.
Check lead quality alongside the rate. A rate that rises while sales stay flat means more enquiries of the wrong kind. Once the measurement is sound, the structured way to improve the number is conversion rate optimisation.
In performance marketing, I judge paid channels on the cost of each real lead or sale, so conversion rate is one of the first numbers I check, and one of the first I question.