Chapter 8
Conversion rate: Why the average deceives
Table of contents
A fashion store shows a conversion rate of 1.7 percent over twelve months – unremarkable, right in the middle of the industry range. Then the managing director asks for mobile and desktop separately: desktop 3.8 percent, mobile 0.9 percent. The error in the mobile checkout hit exactly the device that 72 percent of all sessions came from. The average had hidden it for twelve months.
What the conversion rate measures
Orders divided by sessions, times 100. A store with 30,000 sessions and 600 orders has a conversion rate of 2.0 percent. The session-based rate is standard and comparable with industry figures. A word of caution on GA4: the "session key event rate" shown there counts sessions with any key event, not orders. For the conversion rate, you need to select the purchase event specifically.
Important for interpretation: purchases and sessions must come from the same measurement basis, i.e. both from your web analytics. Anyone who divides store orders by GA4 sessions gets an artificially high rate (worked example in Chapter 3). The orders from the store are used to check how complete purchase measurement is. Read the conversion rate primarily as a trend and as a comparison between segments.
Three splits that turn the average into a basis for decisions
By device: mobile and desktop separately. Mobile almost always converts worse. If the mobile rate is below half the desktop rate, that is a sign of a structural problem in many assortments.
By source: direct and email typically convert several times better than social advertising. An average hides these differences completely. If the share of weakly converting sources rises, the average falls even though the store has not got worse.
By visit and customer status: returning visitors convert far more often than new ones. A rising share of new visitors pushes the rate down without anything being broken. Note: "returning" in GA4 only means someone has been there before, not that they have already bought. If you reliably recognize existing customers, also evaluate by customer status.
How the mix shifts the average
Desktop converts unchanged at 4 percent, mobile at 1 percent. If the mobile share rises from 50 to 80 percent, the overall rate falls from 2.5 to 1.6 percent. No segment has got worse – only the weighting has shifted. Anyone who now rebuilds the checkout is repairing something that is not broken.
When mobile is the problem
Example | Desktop | Mobile | Mobile traffic | Finding |
|---|---|---|---|---|
Fashion store A | 3.8% | 0.9% | 72% | Checkout form breaks on iOS |
Beauty store B | 4.2% | 2.1% | 68% | Mobile loading time over five seconds |
Furniture store C | 1.1% | 0.3% | 54% | PayPal and Klarna unusable on mobile |
Sports store D | 2.8% | 1.9% | 70% | Healthy: mobile at 68% of desktop |
What a weak mobile rate costs is shown by a simple model calculation: with 100,000 sessions a month and a 72 percent mobile share, improving the mobile conversion rate from 0.9 to 1.5 percent brings 432 additional orders. At a €100 net order value, that is €43,200 additional monthly revenue – with unchanged traffic and without spending a single euro more on advertising.
With technical and design problems, these causes are particularly common: slow mobile pages, forms that are tedious on a smartphone, missing express payment methods such as Apple Pay and Google Pay, too many checkout steps, poor readability.
Benchmarks by industry
Industry | Healthy conversion rate |
|---|---|
Consumables (beauty, supplements, food) | 2.0 to 4.0% |
Fashion and accessories | 1.2 to 2.8% |
Electronics and technology | 1.0 to 2.5% |
Furniture and home | 0.4 to 1.2% |
Luxury and jewelry | 0.3 to 1.0% |
B2B stores | 0.8 to 3.0% |
Sports and outdoor | 1.2 to 2.5% |
An industry figure is a rough anchor. The development of your own, cleanly split rates over time is more meaningful.
The levers with the biggest impact
Product pages and checkout that are fast to use on mobile – tested under realistic conditions, not just on the office Wi-Fi
Apple Pay, Google Pay and PayPal as express options
A checkout with as few steps as possible and working guest checkout
Trust elements on the product page and in the checkout: delivery time, returns, reviews
Forms that work with autofill on a smartphone
A higher conversion rate is only progress, however, if more contribution margin remains after discounts, returns and additional costs. Anyone who buys the rate with permanent discounts improves the ratio and worsens the result.
More on optimizing the checkout in Chapter 10.
Where your visitors actually get stuck is revealed by a customer journey analysis. The implementation in your store is described on the page Shopware and Shopify development.
What you can check today
Do you know your conversion rate for mobile and desktop separately?
How large is the gap between the two – is mobile below half of desktop?
Do you know your conversion rate for each important traffic source?
Are Apple Pay and Google Pay active in your mobile checkout?
Have you run a test on a critical checkout page in the last six months?
When did you last go through your own store on a smartphone as a first-time buyer?
Frequently asked questions
Depending on the industry, between about 0.4 percent (furniture, luxury) and 4 percent (consumables). More important than the industry figure is the development of your own rate, split by device and source.
Because mobile structurally converts less easily – but usually not as much less as many stores show. If mobile is below half of desktop, a close analysis of loading time, forms and payment methods pays off.
Per session. That is the usual standard and makes your figure comparable with industry values.
It changes the measurement basis. Declined consent makes sessions and purchases disappear from GA4 to different degrees. If the banner changes, the measured rate can shift without the store having changed.
Often because of a changed visitor mix. More mobile visitors, more new visitors or more social traffic push the average down without the store having got worse. That is why you should always split by source and customer type.
)