Conversion rate

Google Ads says the campaign converts at 6%. GA4 says 2.4%. Your store's dashboard says 1.9%. Same traffic, same week. Each tool divides by something different, and counts a different thing as a conversion.

The action, and what you divide by, are both choices. Google Ads divides conversions by ad interactions, such as clicks. GA4 reports session key event rate, the share of sessions with a key event, and user key event rate, the share of users. (GA4 called key events "conversions" until 2024.) An online shop's dashboard usually divides orders by sessions.

Conversion rate sits between Click-Through Rate (CTR), which gets people to the page, and Cost Per Acquisition (CPA), which tells you what each conversion cost. For a journey with several steps, Funnel Analysis shows where people drop out, and Landing Page Performance Analysis shows which pages convert.

Formula: Conversion rate = conversions ÷ visits × 100

One week, three conversion rates

A week of paid search for an online shop. The figures are invented to show the arithmetic, but every line is a choice real tools make.

Google Ads counts purchases it can link to a click within its window, even days later. GA4 counts sessions, and a returning visitor makes more than one. Your store has its own idea of where a session came from. None of them is wrong; they answer different questions.

The four choices inside the number

1. What counts as a conversion. A purchase, an add to basket, a sign-up. A rate on smaller actions is always higher, so say which.
2. What you divide by. Clicks, sessions or users. Per user comes out higher than per session for the same traffic, because one person can visit several times.
3. Once or every time. Someone who buys twice in a session: one conversion or two? GA4's session key event rate counts the session once.
4. Which traffic. All of it, or one channel. Returning customers convert far better than new visitors, so a change in the mix moves the rate.

Averaging rates is the quiet mistake

A landing page with 100 sessions and 10 orders converts at 10%. Another with 4,900 sessions and 98 orders converts at 2%. The average of the two rates is 6%. The real rate is 108 ÷ 5,000 = 2.2%.

If you sell through a sales team, the form fill isn't the conversion that pays. The lead that becomes a customer is, often months later, and the rate depends on which step you divide by.

Formula: Lead conversion rate = leads that reached the next stage ÷ leads that entered this stage × 100

Lead, MQL, SQL, customer

HubSpot's default lifecycle stages give the usual sequence. A lead is a contact who has converted on your website, or interacted in some other way beyond subscribing to a newsletter or blog. A marketing qualified lead (MQL) is one "that your marketing team has qualified as ready for the sales team". A sales qualified lead (SQL) is one "that your sales team has qualified as a potential customer". A customer has at least one closed deal. Your CRM may use other names. What matters is that the definitions are written down and both teams count the same thing.

Each step has its own rate, and lead to customer is the steps multiplied together.

Lead conversion rate by source

One quarter's leads, followed for six months so most of them had time to close. The figures are invented to show the arithmetic.

Organic: 30% × 40% × 25% = 3%. Paid search brings the most leads and converts the fewest of them; referral is the other way round. Neither fact says where to spend the next pound. For that, put cost next to it: Customer Acquisition Cost (CAC) by source.

People search for lead conversion rate statistics by source, and we don't publish any. Published figures mix different definitions of a lead, count over different windows and come from sales cycles of different lengths. Your own rates by source, measured the same way each quarter, are the comparison that holds.

Form conversion rate

Formula: Form conversion rate = form submissions ÷ visits to the page with the form × 100

It's the top of the same funnel, with the same choices about what you divide by. Judge a form on what happens next as well. One that converts 12% of visitors into leads that never reach MQL is worse than one at 4% whose leads do.

Three traps

1. Attribution. Which source gets the lead? The ad platform credits its own click. HubSpot's Original Traffic Source credits the first known visit, so a lead who found you through organic search and came back through an ad counts as organic. Pick one rule and print it next to the number.
2. The time lag. B2B leads can take months to close. Divide this month's customers by this month's leads and you mix cohorts, and a source that has just grown looks worse than it is. Follow a cohort instead: leads created in one period, counted when they reach each stage. HubSpot records a "Date entered" property for every lifecycle stage, which is what makes that possible.
3. Duplicates. One person fills in two forms, or comes back through another channel. HubSpot deduplicates contacts by email address, so a second submission updates the same contact. A spreadsheet of form submissions doesn't, and the same person counts twice. Count people, not submissions, or say which you counted.

Ask an AI for your conversion rate from an export, and the risk isn't the division. It's what it divides by. A GA4 export might offer sessions, users and engaged sessions; a Google Ads export offers clicks and interactions. It will pick one, usually without saying which.

A two-minute check

1. "What did you count as a conversion, and from which column?"
2. "What did you divide by: clicks, sessions or users?"
3. "Did you divide totals, or average rates across rows?"
4. "Is this one channel, or all traffic?"

The monthly channel report in Claude for marketing pulls GA4 sessions and key events beside each ad platform's own conversions, with every attribution setting printed, so you can see which tool each number came from.

A good conversion rate is one that makes the traffic pay. Work it out from what you pay for a visit:

Formula: Break-even conversion rate = cost per click ÷ gross profit per conversion

At a £1.20 cost per click and £30 of gross profit per order, 4% of clicks need to buy before the ads pay for themselves.

We don't publish average conversion rates by industry. Published tables mix purchases with sign-ups, sessions with users and paid traffic with all traffic, so one study's 3% and another's aren't measuring the same thing.

Better comparisons: the same page over time, new against returning visitors, and one channel against another inside the same tool.

Check tracking first. A broken tag or a new consent banner lowers the rate in one tool and not in your order system. Compare the two before anything else.

The traffic changed. A broad campaign, a new social channel or a viral post brings people with less intent. Look at conversion rate by channel, not for the whole site. A falling Click-Through Rate (CTR) on the ads is often the first sign.

The page changed. A redesign, slower loading, a new pop-up. Compare the rate before and after the release date.

The offer changed. Price, delivery cost, stock. These show up in the last steps of the funnel: people add to basket, then leave.

More of the traffic is on mobile. Mobile visitors usually convert at a lower rate than desktop, so a shift towards mobile lowers the blended rate.

Find the step where people leave. Funnel Analysis shows whether the problem is the landing page, the basket or the checkout. Fix the biggest drop first.

Match the page to the ad. The headline should repeat the promise that earned the click.

Remove steps and surprises. Forced account creation and delivery costs that appear at checkout are classic reasons for abandoned baskets.

Speed up the page, especially on mobile.

Test one change at a time, and judge it on total conversions ÷ total visits over the test, with enough traffic to tell a real difference from noise.

To see the rate by campaign or page from your own data, connect Google Ads or Google Analytics to Count.