How to calculate conversion rate. The formula and worked examples.

The short answer. The formula is one line.

Conversion rate = (conversions ÷ total visitors) × 100.

That's it. If 120 people completed a form out of 3,000 who visited the page, your conversion rate is (120 ÷ 3,000) × 100 = 4%.

The maths is not the hard part. The hard part is being precise about what you put into each side of the equation. The formula never changes, but the definitions of "conversions" and "visitors" vary enormously between teams, tools and time periods. If those definitions aren't locked down, the number you produce is not comparable to anything else you calculate.

That's why the four decisions below matter more than the arithmetic. Get those right and the calculation takes about thirty seconds in a spreadsheet or your analytics tool of choice.

Step by step. Four decisions before you do the maths.

Step 1: define your conversion goal

A conversion is a specific, trackable action that indicates a visitor did what you wanted them to do. Common examples include a form submission, a booking confirmation, a purchase, a free trial sign-up, a document download, or a button click that triggers a key event in your product.

The definition needs to be precise and singular per page. "Conversions" is not a useful category on its own. "Demo requests submitted via the contact form on the product landing page" is useful, because it describes exactly one thing you can measure.

Mixing goals in a single rate produces a number you cannot act on. If your landing page has a primary CTA (book a call) and a secondary CTA (download a one-pager), and you count both as conversions, you'll end up with a blended rate that doesn't tell you whether your primary action is working. Track each goal separately.

This definition also needs to be set before measurement begins, not retrofitted to whatever data happens to be available. Defining the goal after the fact introduces selection bias.

Step 2: count the conversions

Once you know what you're measuring, pull the count from your analytics tool (Google Analytics 4 being the most common starting point for web properties) or directly from your CRM, e-commerce platform, or back-end database.

Then cross-reference the two. If GA4 shows 96 form submissions but your CRM received 83 enquiries in the same window, you have a 13% discrepancy, which is over the 10% threshold where you should investigate. The main causes are:

• Confirmation page blocked or not loading. If the event fires on the thank-you page and a percentage of users never reach it because of a slow redirect or a session timeout, those conversions don't get counted.
• Tracking scripts blocked. Browser extensions (ad blockers, privacy tools) prevent the GA4 tag from firing. The user converts but the event is never recorded. Studies suggest ad blockers affect 20-30% of desktop traffic on many B2B sites, which means your GA4 data may structurally undercount.
• Events firing multiple times. If the event trigger isn't de-duplicated, a single conversion can appear as two or three events. This inflates the count in the opposite direction and makes your rate look higher than it is.

A discrepancy below 10% is generally acceptable given the nature of browser-side tracking. Above 10%, fix the tracking before relying on the rate for decisions.

Step 3: define and count visitors

The denominator of the formula has three common forms, and which one you choose changes the number you get:

• Sessions. Each visit to the page or site, including return visits. One person who visits the same page on three separate occasions counts as three sessions. Sessions are the default for website CRO because they measure each opportunity a visitor had to convert during a given campaign window. If you're running a paid campaign and want to know how that campaign's traffic performed, sessions give you the right denominator.
• Unique users. Deduplicates return visits. That same person visiting three times counts as one unique user. This gives you a share-of-individuals metric, which is more appropriate when you're measuring a nurture campaign or a trial-to-paid flow where you care about what proportion of the people in a cohort converted, regardless of how many times they visited.
• Impressions. Used primarily in paid advertising and email, where the denominator is how many times an ad or email was shown or delivered, not how many clicks came through to a page. Click-through rate (CTR) is the version of conversion rate used at the impression level.

For most website CRO work, sessions is the right choice. For email campaign optimisation, unique clicks to the landing page is the right denominator. For ad performance, you're usually working with impressions or clicks depending on whether you're measuring CTR or click-to-action rate.

The critical rule is this: document your choice. Write it down in the same place as your rate, every time. "4% conversion rate (sessions, 28-day window, organic traffic only)" is a number you can compare across time. "4% conversion rate" is ambiguous and will cause arguments three months from now.

Step 4: apply the formula and segment

Now apply the formula: (conversions ÷ visitors) × 100.

Run it at the aggregate level first, then immediately segment. The aggregate rate tells you where you are. The segmented rates tell you what to do about it.

The most useful segments for most businesses are:

• Traffic source (organic search, paid search, paid social, direct, referral, email). This is almost always the most revealing cut because source correlates with intent. Someone who found you by searching "CRM implementation consultant Dubai" is not in the same frame of mind as someone who clicked a retargeting ad while scrolling social media. Their conversion rates will differ substantially, and the fix for underperformance in each case is completely different.
• Device type (desktop, mobile, tablet). Mobile traffic frequently converts at a lower rate than desktop on the same page, and the gap is often a UX problem specific to the mobile layout rather than a demand problem.
• Landing page. If you have multiple entry points to a funnel, each page needs its own rate. The aggregate masks which pages are carrying the load and which are dragging it down.

The aggregate rate across your whole site is interesting but less useful than knowing that your paid social traffic converts at 1.2% while your organic blog traffic converts at 4.8%. The first number gives you a health indicator; the segmented numbers give you a prioritised action list.

A worked example.

Here's how this plays out with real numbers. Consider a UK-based B2B software company running a demo-request landing page for their product.

Over a four-week period, the page received 2,400 sessions. GA4 recorded 96 conversion events (demo request form submissions, confirmed against CRM records). The top-level calculation:

(96 ÷ 2,400) × 100 = 4.0%

That number looks reasonable at face value. But when you break it down by traffic source, the picture changes considerably:

• Organic search: 900 sessions, 54 conversions = 6.0%
• Paid search: 1,100 sessions, 33 conversions = 3.0%
• Direct / referral: 400 sessions, 9 conversions = 2.25%

The aggregate 4.0% masks a significant problem. Paid search, which is the most expensive traffic source by a considerable margin, is converting at half the rate of organic search. For every £1,000 spent bringing paid search visitors to that page, you're getting half the demos you'd get from the same number of organic visitors.

The next action is now clear: investigate why paid search traffic is underperforming. Is there a message-match problem between the ad copy and the landing page? Are paid search visitors landing with different intent? Is the audience targeting pulling in prospects who are earlier in their research? These are all answerable questions, but you'd never know to ask them without the segmented rates.

That's the real value of calculating conversion rate correctly: not the headline percentage, but what the segmented data points you towards.

Common mistakes. And how to avoid them.

Mixing sessions and users

Switching between sessions and unique users from one reporting period to the next makes period-on-period comparisons invalid. If January's rate was calculated on sessions and February's on unique users, you're not measuring the same thing, and any apparent change in performance is noise rather than signal.

Set the definition once, at the start of a campaign or measurement period, write it down in your reporting template, and check it every single time you pull numbers. If you're handing a report to someone else, include the definition in the report. Don't make them guess.

Counting the wrong event

Form submissions and form confirmations are not the same thing. A form submission event fires when someone clicks the submit button. A confirmation event fires when the thank-you page loads, or when a success state appears in the form UI.

If your tracking fires on the button click, you're counting every submission attempt, including the ones that fail due to validation errors, network timeouts, or duplicate submissions. That inflates your conversion count and makes your rate look higher than it actually is.

Track the confirmation page view or the thank-you state. In GA4, this typically means either a page_view event on the /thank-you URL or a custom event that fires only after a successful server response. Test it in DebugView before trusting the data in production.

Aggregating too broadly

A site-wide conversion rate is not an actionable number. It tells you that some percentage of all your visitors across all your pages did something you wanted them to do. But which pages? From which sources? On which devices? For which conversion goals?

A site-wide rate can tell you whether your overall performance is trending up or down over time. It cannot tell you what to fix. For that, you need rates at the page level, source level, and device level at minimum.

If you're reporting conversion rate to a team or to a client, include the segmented breakdown alongside the top-level number. The headline rate is context; the segments are the story.

Not segmenting by device

Mobile and desktop visitors behave differently and convert at different rates, often for reasons that have nothing to do with traffic quality. Research by the Baymard Institute shows that mobile cart abandonment in e-commerce runs 10 to 15 percentage points higher than desktop, even when the mobile experience is considered acceptable by standard usability testing.

If your page is converting at 5.0% on desktop and 1.8% on mobile, you don't have a demand problem or a targeting problem. You have a mobile experience problem specifically. The fix is a UX audit of the mobile flow, not a revision of your ad targeting or your offer.

Run the device-type segment on every page where conversion rate matters. The gap between mobile and desktop performance is almost always larger than teams expect, and it's consistently one of the fastest wins available once it's identified and addressed directly.

From calculation to action.

The formula takes seconds. The discipline required to get the inputs right, the definitions consistent, and the output properly segmented takes a little longer. But once you've built that habit into how you report on performance, you have something genuinely useful: a baseline that tells you where you are, segments that tell you what to do next, and a methodology you can repeat every period without renegotiating what the number means.

That's the foundation. The next stage is building a structured programme around it. Our guide to conversion and funnel optimisation covers how to move from a calculated rate to a tested hypothesis and a measured improvement. If you're already at that stage and want structured support, the CRO consulting and support page outlines what that looks like in practice.