Conversion rate: how to calculate it, and what to do with the result

Start with the question: what counts as a conversion?

Conversion rate tells you what share of the people who reach you do what you wanted them to do. The catch sits in “what you wanted”. A conversion is not always a sale. It might be an order, a quote request, a phone call, a newsletter sign-up or a brochure download. Each of those has its own rate, and blending them into one figure just creates noise.

So the first job is a decision, not a calculation. What is the one action you are measuring on this page, or in this campaign? An online shop usually has a main conversion (a completed order) and a few supporting ones (add to basket, sign-up). A consultancy has the enquiry form or the call. An event landing page has the registration.

If you skip that decision, your conversion rate becomes a number that drifts up and down without meaning anything. Once the action is defined, the same number turns into a working tool: you can follow it over time, split it by traffic source and see where people drop out.

The formula, and a worked example

The arithmetic is trivial. The mistakes usually come from choosing the inputs.

1. Pick the period. For instance the last 30 days. Always compare periods of the same length.
2. Count the conversions. Only the defined action, with no duplicates.
3. Count the denominator. Sessions, users or clicks. Choose one and stick to it.
4. Divide, then multiply by 100. Conversions ÷ denominator × 100.
5. Record the result with its definition. “1.7% — orders ÷ sessions, 30 days” is information; “1.7%” on its own is not.

A hypothetical shop gets 5,000 sessions in a month and 85 orders: 85 ÷ 5,000 × 100 = 1.7%. A consulting firm’s enquiry page receives 450 visitors and 18 completed forms: 18 ÷ 450 × 100 = 4%. These are illustrations of the maths, not benchmarks for your business. To run the numbers on your own data, try our conversion rate calculator.

Which formula for which goal

The same idea adapts to different goals. This table covers the versions you will meet most often.

Notice that an ad platform normally reports conversions against clicks, while your analytics tool reports site-wide conversions against sessions. Both are called “conversion rate”, yet they cannot be compared directly.

Sessions, users or clicks: pick the denominator once

If a visitor comes back three times before buying, that is three sessions and one order. Divided by sessions, the rate looks lower than when divided by users. Neither is wrong, but the choice must stay the same from one month to the next.

Measurement platforms use terms and definitions that change over time. In Google Analytics 4, for example, events you mark as important are now called key events (they used to be called conversions), and reports can show rates per session or per user. Google Ads and Meta Ads count conversions according to their own attribution rules. That is why the figure in analytics rarely matches the ad platform exactly; it does not mean either tool is broken. Check the current definition in the interface and write it down. For help getting the measurement itself right, see our GA4 guide for small businesses.

Micro and macro conversions: reading the funnel

A macro conversion is the end goal, such as an order or enquiry. Micro conversions are the steps on the way: viewing a product, adding it to the basket, starting checkout, scrolling to the form. They matter most when macro conversions are rare, as with expensive services or a new shop.

A hypothetical funnel for 8,000 sessions:

• 2,800 sessions reach a product page (35%);
• 560 add something to the basket (20% of those 2,800);
• 224 start checkout (40% of 560);
• 112 complete the order (50% of 224).

The overall rate is 112 ÷ 8,000 = 1.4%. But the funnel tells you more than that final figure. If half of the people who start checkout leave, that is where to look: surprise delivery costs, forced account creation, a long form or limited payment options. The homepage may be perfectly fine.

Why market averages mislead you

You will often find articles quoting “the average conversion rate” for an industry. Without repeating any specific figure, here is why such averages rarely help you:

• Different definitions. One study counts orders, another enquiries, another any tracked event. The average blends things that are not comparable.
• Different products and prices. A €20 accessory is bought differently from a service that costs thousands. The speed of decision varies.
• Different traffic. People arriving from a search with clear intent behave differently from those who clicked a social post out of curiosity.
• Shifting mix. Even within your own site, the overall rate can fall without anything breaking.

A hypothetical example of that last point. In month one, paid social brings 3,000 sessions converting at 0.8% (24 orders) and search brings 1,000 sessions at 4% (40 orders). That is 64 orders from 4,000 sessions, or 1.6%. In month two you raise the social budget: 8,000 social sessions at the same 0.8% (64 orders) plus the same 1,000 search sessions (40 orders). Now you have 104 orders from 9,000 sessions, or about 1.16%. Orders went up by more than half, yet the overall rate dropped, because you added more low-intent traffic. Nothing is broken. The lesson is to split the rate by source, device and page type.

Devices show the same effect. If desktop has 2,000 sessions at 2.5% (50 orders) and mobile has 6,000 at 1.0% (60 orders), the combined rate is 110 ÷ 8,000, roughly 1.4%. That blended figure hides the fact that mobile converts at well under half the desktop rate, which may be your biggest opportunity.

Is your sample big enough?

A rate based on a handful of conversions is unstable. If you get 4 orders one week and 7 the next, the gap may be pure chance rather than a trend. As a rough guide, you want at least a few hundred sessions and a few dozen conversions in the segment you are analysing before you make big decisions on it. The smaller the difference between two numbers, the more data you need to trust it.

On low-traffic sites, measure over longer periods, group data into broad segments (all mobile traffic, for instance) and lean on micro conversions. When you compare two versions of a page, do not stop the test after two days because one version seems to be ahead. The rules are in our guide to A/B testing your website.

From percentage to decision

A conversion rate is a starting point, not a score to chase. A simple routine:

• Segment it: by traffic source, device, landing page, new versus returning visitors.
• Find the biggest leak: the funnel step where most people leave.
• Write a hypothesis: “the form has too many fields on mobile”, not “we want more”.
• Change one thing and measure.
• Tie the rate to money: cost per conversion and revenue per visitor.

The link to your budget is direct. Cost per conversion is cost per click divided by conversion rate. At €0.80 per click and a 2% rate, a conversion costs €40; at 4% it costs €20. A better rate lets you either pay more per click or get the same results for less money. We explain how these figures feed into a spending plan in our guide to setting an online advertising budget.

Common mistakes

• Benchmarking against a number you read somewhere. Compare yourself with last month and with the same season last year.
• Changing the definition mid-period. If you add a new tracked event, note the date, or your charts will mislead.
• Counting duplicates. A thank-you page that reloads can record the same order twice.
• Leaving in bots and internal traffic. Filter out your own team’s visits and obviously fake traffic.
• Judging by one week. Holidays, promotions and even the weather can change behaviour.
• Chasing rate without value. A higher rate won through heavy discounting can bring less profit.
• Trusting broken tracking. If tags are misconfigured, every calculation sits on a faulty base.

Key terms

Conversion: the action you want, defined by you. Session: one visit to your site, which can include several pages. User: a person (in practice a browser or device) who makes one or more sessions. Denominator: the figure you divide conversions by. Attribution: the rule that credits a conversion to a channel or campaign. Cost per acquisition (CPA): spend divided by the number of conversions. Revenue per session: total revenue divided by sessions.

How we work with conversion data

When we review a site, we begin with the conversion definition and with the measurement itself: what is counted, what is missed, what is counted twice. Only then do we split the rate by source, device and page and look for the step with the biggest loss. Our recommendations are hypotheses to test, never promises of uplift. We cannot guarantee a particular percentage, because the outcome depends on offer, price and traffic quality.

If you would like help with this process, our conversion rate optimisation page explains what we do and how we work.

In short

• Conversion rate = conversions ÷ denominator × 100. Choose both and keep them fixed.
• Calculate separate rates per goal: orders, leads, sign-ups, add to basket.
• Split by source, device and page; the overall rate can hide problems or fall for harmless reasons.
• Ignore “industry averages” and compare with your own history.
• Turn the rate into action: find the biggest leak, form a hypothesis, test, then look at cost per conversion.