Average landing page conversion rates: compare like goals

Average landing page conversion rates should compare the same action against the same eligible traffic, with source and device shown separately. Calculate the overall rate from total converted sessions divided by total sessions. Keep lead submissions, purchases and signups in separate groups so a change in the goal mix does not look like better performance.

Define the action before choosing a benchmark

A landing page is the first page in a visit. A visitor who later opens a product page has not created another landing-page session for that visit. Google's landing-page report documentation defines the landing-page dimension from the first pageview in a session and explains how to break it down by session source and medium.

Write down what counts as success. For a lead page, it might be a completed contact request. For a purchase page, it might be a completed order. For an account page, it might be completed registration. A button click is a different action from a completed request, even when the button begins that request.

An outside benchmark needs the same scrutiny. Unbounce's conversion benchmark report distinguishes medians from averages and explains that its conversion-event comparisons do not measure the value of each conversion. Use that distinction when someone offers a single industry figure: ask which actions, pages and traffic sources it includes.

Covirage is not affiliated with Google or Unbounce.

The business question is usually whether comparable traffic is producing more of the intended outcome. That is a narrower question than whether marketing is generating profitable customers. For the general denominator distinction, see win rate, close rate and conversion rate.

Bring an aggregate export with a declared grain

For this analysis, use one row per reporting period, landing page, goal, source and device. Include eligible sessions and sessions completing that goal. Count each session once within its goal, even if the action repeats. If the export supplies event counts instead, obtain a matching converted-session summary before calling the result a session conversion rate.

Keep the reporting timezone, date window and inclusion rules the same across files. Retain an unknown source or device category rather than silently dropping it. If one session can appear under several goals, keep those goal totals separate; they cannot be added into a distinct-session total without an overlap check.

The following fictional US business uses separate pages for contact requests, purchases and account registrations. The September 2026 export contains mutually exclusive landing-page sessions across these pages. Rates are shown to two decimal places; calculations use the original counts.

Landing-page conversion rate = Sessions completing the defined goal / Eligible landing-page sessions

The desktop search lead rate is 42 / 600 = 7.00%. The mobile search purchase rate is 7 / 700 = 1.00%. Those figures answer different questions despite sharing a traffic source.

Calculate the weighted average and prove the totals

Assume the table starts in row 2, with sessions in column D and converted sessions in column E. The rate in F2 is:

=IF(D2=0,"",E2/D2)

Format the result as a percentage and fill down. A zero denominator returns a blank for investigation; it does not establish a zero conversion rate.

The lead rows occupy rows 2 through 5. Their combined rate is:

=SUM(E2:E5)/SUM(D2:D5)

The unweighted average of the lead percentages is (7 + 4 + 12 + 6) / 4 = 7.25%. It overstates the combined lead rate because it gives the smaller email-desktop row the same weight as the larger search-desktop row.

Reconcile the export before presenting it. Sessions total 1,500 + 1,000 + 500 = 3,000. Converted sessions total 100 + 16 + 70 = 186. Those counts must match the source report under the declared rules. Dividing 186 by 3,000 produces 6.20%, but that mixed-goal rate cannot tell the manager which action improved.

Separate source performance from device mix

For lead generation, search produces (42 + 16) / (600 + 400) = 5.80%. Email produces (24 + 18) / (200 + 300) = 8.40%. Review the device rows before treating this as a reason to move spending.

A comparison using equal desktop and mobile weights holds device mix constant:

These are comparison rates under an assumed traffic mix, not observed results or forecasts. Email remains higher in this example, but the table does not establish that the source caused the difference. Audience intent and the offer can still differ.

Marketing operations should own the definition and export reconciliation. The campaign owner investigates source and device exceptions. Sales operations then checks downstream quality using a separate stage conversion measure. Keep later outcomes, such as quote conversion, tied to their own eligible records rather than blending them into the landing-page denominator.

Where it goes wrong

• Averaging percentages. Sum the matching counts before dividing. Small pages should not receive the same weight as large pages by accident.
• Events divided by sessions. Several events in a visit can inflate a figure intended to measure sessions that converted.
• A changing goal mix. More registration traffic can lift a combined rate while purchase performance is unchanged.
• Overlapping exports. A report total pasted beneath its device rows counts the same traffic again.
• Treating an association as a cause. Aggregate source differences identify questions; they do not establish what would happen after a budget change.

Compare your own landing-page exports in Covirage

Bring an aggregate file or Google Sheet with landing page, goal, source, device, sessions and converted sessions. Covirage maps the columns, and its tools compute and check the grouped rates; the external AI model explains the result with every figure traced to its rows. Create a free Covirage account, then set up your business and bring a sample of your own export.

See sales insights for the wider business context.