Conversion rate is the share of visitors who do the thing you wanted. Simple to calculate and remarkably easy to misuse, because the number means nothing without knowing what you counted.
The calculation
Conversions divided by visitors, times one hundred. Fifty enquiries from two thousand visitors is 2.5 percent.
The complications are in the definitions. Visitors or sessions? Unique users or all visits? And what counts as a conversion — a form submission, a phone call, a purchase, a newsletter signup?
Two sites reporting the same rate may be measuring entirely different things. Comparing them is meaningless.
Why benchmarks mislead
You will find articles stating an average conversion rate for your industry. Treat them with suspicion.
The number depends on traffic mix. A site whose visitors come mostly from branded search converts far better than one running broad display campaigns — same site, same quality, different denominator.
It also depends on what you count. A site counting newsletter signups will always look better than one counting only purchases.
Your own trend over time is the only comparison that means anything.
What actually moves it
Measure the right thing first
Before optimising, confirm the number is real. A conversion recorded on a button click rather than a confirmed submission overstates results, and every decision after that is corrupted.
Match message to expectation
If someone arrives from an ad promising a specific thing and the page discusses something else, they leave. This is the largest and most common loss.
Remove friction
Every form field, every extra step, every required account costs completions. Ask only for what you genuinely need.
Answer the objection
People hesitate for specific reasons: unclear price, unknown company, uncertainty about what happens next. Addressing these directly converts better than any button colour.
Make it work
A surprising share of conversion problems are defects. A form failing on one browser, a phone number that is not tappable, a checkout breaking at a particular step.
This site is an example: its own contact form once returned success while sending nothing, because a hidden anti-spam field named website was being autofilled by browsers. The conversion rate looked fine. Every enquiry was being discarded.
Micro and macro conversions
The macro conversion is the goal — a sale or an enquiry. Micro conversions are steps toward it: a pricing page view, a video watched, a document downloaded.
Tracking micro conversions helps diagnose where a funnel loses people, especially at volumes too low for reliable testing.
Is a higher rate always better?
No. Weakening qualification raises the rate and fills the pipeline with people who never buy. Quality matters more than percentage.
How much traffic do we need to A/B test?
More than most sites have. Reaching significance needs meaningful volume per variant. Below that, testing produces noise read as insight.
What should we fix first?
Verify the measurement, then fix anything that is actually broken, then improve message match. In that order.