Conversion rate calculator
Divide conversions by the eligible total, then multiply by 100. 20 converted visitors out of 1,000 visitors is 2%. Enter your counts below. Choose whether the numerator counts people or sessions that converted at least once, or repeated conversion events. That choice changes what the answer means.
Calculator
Conversion rate calculator
Enter the number that converted and the number that had the chance to convert.
Use counts from the same population and period. Blank means unknown; zero is a count.
Compare another group
Turn on comparison and enter B counts.
Context: missing. Add definitions below if you want to describe the comparison.
Set a target at the current denominator
Enter a target rate.
Add measurement context
These labels travel with your export. They do not change the arithmetic.
Add the optional definitions before sharing the result so someone else can tell what you counted. You can also compare two rates or calculate how many conversions a target rate requires at your current denominator.
How to calculate conversion rate
Conversion rate (%) = conversions ÷ eligible total × 100. The eligible total is the group that had the opportunity to complete the action you are measuring. For a landing page, that might be sessions on the page. For lead-to-meeting conversion, it is the leads in the defined cohort. For checkout completion, it is the people who reached checkout.
Suppose 1,000 distinct visitors saw a demo page and 20 of those visitors requested a demo. The rate is 20 ÷ 1,000 × 100 = 2%. You can write: “2% of visitors to this demo page requested a demo during the measured period.” That sentence requires a distinct-visitor numerator and denominator. If the 20 is a count of form submissions, including repeat submissions, it answers a different question.
Keep the action, population, and reporting window together. “We converted 2%” leaves the reader guessing. “20 of 1,000 demo-page visitors requested a demo between October 1 and October 14” identifies the calculation. Whether those requests were suitable prospects is a separate measure.
Choose what your numerator and denominator count
Use the first calculator mode when the numerator is a subset of the denominator: users who converted out of all eligible users, sessions with a key event out of all sessions, or leads who attended a meeting out of the eligible leads. Count each denominator entity at most once in the numerator. The result cannot exceed 100%.
Use repeated-event mode when one denominator entity can produce several counted conversions. For example, a person can place several orders or submit several forms. The result is events per 100 denominator units. It may exceed 100%, and it does not tell you the percentage of people who converted.
Share of people who converted
20 people who ordered
÷ 100 eligible people × 100
20% of people
Each person counts once, even if they placed several orders.
Orders, including repeats
150 orders
÷ 100 eligible people × 100
150 orders per 100 people
The order count does not tell you how many different people ordered.
| Question | Numerator | Denominator | What the answer means |
|---|---|---|---|
| What share of visitors requested a demo? | Distinct visitors who requested a demo | Distinct eligible visitors | Percentage of those visitors who requested a demo |
| What share of sessions included a key event? | Sessions with at least one selected key event | Total eligible sessions | Percentage of those sessions with the event |
| What share of leads attended a meeting? | Distinct leads who attended | Distinct leads in the cohort | Lead-to-attended-meeting rate |
| How many orders were placed per customer? | Orders, including repeat orders | Customers in the defined population | Orders per 100 customers |
Google Analytics makes the session and user distinction explicit. Its session key event rate divides sessions containing a key event by total sessions. Its user key event rate divides users who performed a key event by total users. A returning person can have several sessions, so the two rates can differ even when they describe the same website.
An event count is also different from a converted-user count. Google defines event count as the number of times users triggered an event. Its key event counting settings can count once per event or once per session. Counting once per session still does not deduplicate one person across several sessions.
Why a conversion rate can exceed 100%
Imagine 100 customers place 150 orders. The event rate is 150 ÷ 100 × 100 = 150 orders per 100 customers. Those orders could come from one customer or many customers; the order count alone does not establish how many. Select repeated-event mode for the order calculation. Use a distinct count of customers who ordered if you want the share of customers who purchased.
Google Ads defines conversion rate using conversions divided by ad interactions that can be tracked to a conversion in the same period. Google says the rate can exceed 100% when you track multiple conversion actions or count every conversion. Use the eligible interaction count from the same reporting setup when reproducing that metric. Replacing it with website users will produce another metric.
Four worked examples
These examples are fictional. They explain the calculation and its limits; the percentages are not performance benchmarks.
1. A landing page generates requests
A page has 2,000 eligible sessions. In 60 of those sessions, someone submits the demo form. The session conversion rate is 60 ÷ 2,000 × 100 = 3%. The numerator is sessions containing a submission, so two submissions in one session still contribute one converted session.
Now suppose sales accepts 24 distinct people from those requests. Dividing 24 people by 60 sessions would mix units. To calculate request-to-qualified-lead conversion, first identify the distinct people who requested a demo, apply the qualification rule, and count accepted people within that same cohort. Keep the 3% session rate alongside the qualified-person measure in your report.
2. Leads progress to attended meetings
A campaign produces 80 distinct eligible leads. After the agreed follow-up window, 12 of them attend a meeting. Lead-to-attended-meeting conversion is 12 ÷ 80 × 100 = 15%. A booked meeting that was cancelled does not belong in an attended-meeting numerator.
If the campaign is still producing leads and the newest people have only had a day to respond, label the result with its outcome cutoff. Compare it with another cohort only after considering how long each group had to reach the outcome. This calculator divides the supplied counts; it cannot establish that both groups had an equal opportunity to convert.
3. A rate rises from 2% to 3%
Group A has 20 converted visitors out of 1,000. Group B has 36 out of 1,200. Their rates are 2% and 3%. Subtract the rates to get +1 percentage point. Divide that change by the original 2% rate to get +50% relative change.
Both statements describe the same change. “Up 1%” is ambiguous. Use the full label when you share it. If Group A starts at zero, relative change is undefined because it requires division by zero; the percentage-point difference still describes the absolute change.
The arithmetic does not establish why the rate changed. Different audiences, offers, sources, time periods, or follow-up windows can change the result. An A/B-test winner also requires an experiment design and uncertainty analysis. The comparison here does not perform that analysis.
4. Small and large groups have different rates
One group has 10 conversions from 100 eligible people: 10%. Another has 20 from 1,000: 2%. Averaging the two percentages gives 6%, but that gives the small group the same weight as the large one.
If the groups are disjoint and use the same definitions, combine the counts: (10 + 20) ÷ (100 + 1,000) × 100 = 2.73%. The combined result gives each eligible person equal weight. Do not combine the counts when the same people appear in both groups or the groups measure different actions. Keep those rates separate until the overlap or definition is resolved.
Calculate the count needed for a target rate
Open the target option and enter the rate you want at the current denominator. With 101 eligible people, a 2.5% target needs 2.525 conversions mathematically. Counts are whole numbers, so you need 3 converted people, producing 3 ÷ 101 × 100 = 2.97%. Two would fall short.
If you already have one conversion, the additional count is two. If you already meet the target, the additional count is zero. The calculation holds the denominator fixed. If more people enter the population, recalculate. It does not predict that an ad or page change will produce the additional conversions.
Missing data, zero results, and inconsistent counts
A blank count means you have not supplied it. It stays Unknown. A measured zero conversion count with a positive denominator produces 0%. Those are different situations: “no conversions recorded among 500 people” is a result; “we do not have the conversion count” is a data gap.
A zero denominator produces an undefined rate. This applies to both 0 ÷ 0 and a positive numerator divided by zero. Check whether the population really had no eligible entities, whether the export is incomplete, or whether the numerator belongs to another group. The calculator keeps the zero visible and does not turn it into an invented percentage.
In subset mode, more converters than eligible entities is inconsistent. Review duplicates, filters, attribution, and the reporting window. If the numerator genuinely counts repeat events, switch modes and describe that metric accurately. The switch changes the interpretation; it does not verify your source data.
Save a rate someone else can reproduce
Add the numerator definition, denominator definition, population, window, and source in the optional context fields. Then copy the memo or download the JSON record. The memo carries the counts, arithmetic, supplied context, and missing information. JSON lets you reopen the same inputs later; the calculator recomputes the result when you import it.
Use the blank record for a clean start or inspect the worked record. The worked memo shows how to describe the calculation. If you want an agent to review an export, give it the standalone review prompt and the full JSON file. The prompt checks the calculation and definitions without filling gaps with invented facts.
For a comparison, use the same action definition and population unit, and explain the source and window for each group. Matching labels alone do not establish that two audiences or observation periods are comparable. Resolve those questions before changing a budget or declaring an improvement.
What is a good conversion rate?
Use a baseline for the same action, audience, source, and measurement method. A low-friction download and an enterprise demo request ask for different commitments. A session rate and a user rate count different populations. Comparing either with an unlabeled industry average can send you toward the wrong change.
Review the business outcome alongside the rate. More form fills can raise a page rate while creating more unsuitable leads. A smaller number of requests can still produce more accepted opportunities. Keep counts and qualification definitions visible, then decide whether to improve the offer, audience, page, form, or follow-up. The landing-page examples show ways to make the offer and next step easier to understand.
Use CTR for clicks divided by impressions, cost per lead for spend divided by leads, and customer acquisition cost for acquisition cost divided by customers. Each answers a different question. If your team needs help connecting its ads, destination, and CRM outcomes, book a Demand Gen Roast.