Conversion Rate
Conversion rate is the percentage of people who complete a desired action, such as submitting a form or finishing a quiz, out of everyone who had the opportunity to do so.
Key takeaways
- A conversion rate is meaningless until its numerator, denominator and time window are stated.
- Narrower targeting usually lifts the rate while cutting the absolute number of conversions.
- Track a chain of stage rates instead of one headline funnel number.
- Small samples make the rate swing on a few conversions; wait for volume.
- Published industry benchmarks rarely share your denominator, so they cannot validate your number.
In depth
A conversion rate is a fraction with two deliberate choices behind it. The numerator counts a specific completed action within a defined window; the denominator counts the population that had a real chance to take it. Change either and the number moves without anything about the funnel changing. Sessions, unique visitors and eligible sessions all produce different rates from identical behaviour. This is why a rate travels with its definition attached: the action, the population, the time window and whether repeat visits are counted once or every time.
Two forces move the rate in opposite directions: how motivated the arriving traffic is, and how much the page asks of it. Tightening targeting raises the rate while shrinking volume; broadening reach does the reverse. Asking for fewer fields lifts completion but hands sales a thinner record. Discounts and free offers raise the rate and lower the average value of what converts. Every gain has a side that must be checked, which is why a rate is judged next to volume and downstream quality rather than alone.
In practice teams keep a small set of named rates rather than one headline number: click-to-page, page-to-start, start-to-finish, finish-to-capture. Each is watched by source and by device, because a mobile rate and a desktop rate rarely match and averaging them hides both. A quiz funnel makes these breakpoints explicit, since a scorecard already records who started, who answered the last question and who handed over contact details. Reporting the chain rather than the endpoint shows which link, not which campaign, is doing the damage.
The rate stops being informative at low volume, where a handful of extra conversions swings it by several points and the swing is noise. It also misleads for long or multi-visit purchases, because the visit that converts is rarely the visit that persuaded. Comparing your rate to a published industry figure is close to meaningless when the other party defined the denominator differently. And a rate cannot tell you whether the converting audience was worth having; only revenue or qualification data answers that.
Example in practice
How to measure it
Start with the pair of counts, not the percentage: completed actions over eligible visitors in the same window. Store both, because the percentage alone hides whether a move came from more conversions or fewer visitors. Split the result by traffic source, device and landing page, and report the segment rates next to their volumes so a 40 percent rate on thirty visitors is visibly fragile.
Then pair the rate with a quality signal from further down: the share of converters that reach a qualifying score, book a call, or become revenue. Read the two together. Rate up and quality flat is a real win; rate up and quality down means you bought completions with looser criteria. Trend both over rolling weeks rather than comparing single days, which are dominated by weekday effects.
Common mistakes
The most common error is removing form fields until the rate climbs, then handing sales a list of first names and free email addresses. The rate improved; the pipeline did not. Before cutting a field, check whether sales actually uses it. If they do, move it after the value has been delivered rather than deleting it, so the ask lands when the visitor already has a reason to answer.
The second is calling a test after a few days because one variant is ahead. Early leads reverse constantly, and traffic mix shifts by weekday. Decide the sample size and the run length before starting, then leave the test alone. A related error is comparing this month's rate to last month's when a new channel entered the mix; segment by source first, or you will credit a design change for a traffic change.
Frequently asked questions
How is conversion rate different from click-through rate?
Click-through rate measures how many people clicked an ad or link, while conversion rate measures how many completed the final desired action. A high click-through rate with a low conversion rate usually signals a mismatch between the promise and the destination.
Why can a higher conversion rate sometimes hurt?
If you optimize purely for completions, you can attract many low-quality leads who never buy. Pairing conversion rate with lead-quality scoring keeps the metric honest by rewarding conversions that actually move toward revenue.
What counts as the denominator in a conversion rate?
The population that had a genuine opportunity to take the action, counted the same way every period. For a landing page that is usually unique visitors who loaded the page; for a checkout it is sessions that reached the cart. Pick one basis, write it down, and never switch it mid-comparison, because a denominator change looks exactly like a performance change.
What is a good conversion rate?
There is no portable answer, because the number depends on the offer, the price, the traffic source and how the denominator was defined. A better target is your own trailing figure: beat last quarter's rate on the same definition and the same segments. Use outside numbers only as a sanity check on order of magnitude, never as a goal.
How much traffic do I need before the rate is reliable?
Enough that a few extra conversions do not move the percentage much. A practical rule of thumb is to look at the raw conversion count rather than the visitor count: below a few dozen conversions per variant, the rate is mostly noise. If the page converts rarely, lengthen the measurement window instead of drawing conclusions from a short one.
Why did my conversion rate drop when traffic went up?
Usually because the new traffic is less qualified, not because the page got worse. Scaling a campaign reaches people further from the original audience, and the rate falls even as total conversions rise. Segment by source and campaign before reacting. If the rate held steady inside every segment, nothing broke; the mix changed.
Should I optimise for conversion rate or for lead quality?
For both, measured together. Set the rate as the target and a quality floor as the constraint: the share of converters reaching a qualifying score, or cost per qualified lead, must not degrade. A change that lifts the rate and breaks the floor is rejected. This keeps the team from winning the metric and losing the pipeline.
How do I calculate a conversion rate across several steps?
Multiply the step rates rather than averaging them. If eighty percent of visitors start a quiz, sixty percent of starters finish and half of finishers submit contact details, the end-to-end rate is the product of the three, not their mean. Keeping the steps separate also shows which multiplication term is dragging the result down.