First, compare devices within the same source and campaign
The overall mobile rate mixes device with traffic. If most mobile visits come from social feeds and most desktop visits from search, the gap is largely a difference in who arrives. Compare the conversion rate (CVRShare of visits or clicks that convert.) of phones and desktops within the same source, the same campaign and, where possible, the same landing page. A gap that shrinks sharply once sources are held equal is mostly traffic mix; a gap that persists within one source is a genuine device problem.
Then compare step rates rather than final rates. Lay out the journey from landing to conversion, split each step by device, and look for the step where the mobile rate falls furthest below desktop. A step where both devices lose a similar share has a problem that is not about the device. A step where phones lose far more, such as the first screen, a form, the delivery details or payment, is where the work belongs. If the gap appeared suddenly rather than existing for months, the conversion rate troubleshooting hub routes to the sudden-drop diagnosis, which looks for a release or browser bug first.
Diagnosis table
Diagnosis
| Symptom | Check | Likely cause | Fix |
|---|---|---|---|
| Mobile visitors leave before the first screen is usable; the gap is largest on paid mobile traffic | Field data for loading and responsiveness by device and page | Pages slow on mobile | Fix the heaviest assets and tags on the busiest templates |
| Mobile form starts are healthy but completions lag far behind desktop | Form step rates by device; layout, labels and keyboards on a phone | Forms and layouts designed for desktop | Redesign forms and first screens for the phone |
| The delivery or payment step loses far more mobile shoppers than desktop ones | Step rates by device at delivery and payment; methods offered on phones | Payment and checkout awkward on a phone | Add wallets and lookup; cut typing at checkout |
| Mobile visits are many, mobile sales few, and desktop sales follow mobile visits | Cross-device paths where identity allows; assisted conversions by device | Journeys finished on desktop | Credit mobile's role; make returning on desktop easy |
| Most mobile sessions come from social and display placements | Device share by source; intent of the placements | Low-intent mobile traffic | Judge mobile by source; match pages to intent |
Fix, verify and prevent each cause
Pages that are slow on real phones
Lab tests on a fast laptop hide what a mid-range phone on a mobile network experiences. web.dev describes Core Web Vitals as field measures assessed on real visitors at a percentile that gives weight to the slower experiences, which is exactly where mobile visitors sit.
Fix: read field data for Core Web VitalsGoogle's user experience metrics. by device on the templates that carry most mobile traffic, then fix the causes in the template and the tag configuration as the speed and mobile guide explains; the search section's diagnoses for slow loading and slow responsiveness cover the metric-level causes.
Verify: mobile field measures move into better bands, and the mobile step rate at the first screen improves over the following weeks.
Prevent: set a performance budget per template and test every new tag on a throttled phone before it ships.
Layouts and forms squeezed from a desktop design
A page designed on a large monitor and narrowed for phones pushes the offer below the fold, shrinks buttons, and leaves forms in two columns with placeholder labels.
Fix: design the phone version first: headline, offer and action within the first screen, large separated tap targets, and forms in one column with labels above, the right keyboard for each field and working autofill, as the forms guide sets out; consider a multi-step form that captures contact details first.
Verify: mobile completion at the form and the first-screen action rise towards desktop levels within the same sources.
Prevent: review every new page and form on a phone before a desktop, and make phone screenshots part of design sign-off.
Checkout and payment that demand typing
Address entry, card numbers and account creation are tolerable on a keyboard and painful on a phone, so the delivery and payment steps are where mobile shoppers drop out most.
Fix: offer the digital wallets common on shoppers' phones, add address lookup and autofill attributes, default to guest checkout, and keep the order summary visible at payment, following the checkout optimisation guide.
Verify: the mobile step rates at delivery and payment close much of their gap with desktop, with order value unchanged.
Prevent: test checkout on the most common phones, with the on-screen keyboard open, after every release and whenever a payment method is added or removed.
Journeys that start on a phone and end on a computer
Many people research on a phone and buy later on a laptop, especially for considered purchases. Reports that credit only the last session make mobile look useless while it starts the journeys desktop finishes. The attribution models comparison explains how the choice of model shifts credit between early and late touches.
Fix: read assisted conversions and cross-device paths where the property can join devices, compare an early-touch model with last click to see mobile's share of starts, and make it easy to continue later with saved baskets, emailed quotes and sign-in that carries across devices.
Verify: with AttributionAssigning conversion credit to touchpoints. stated, mobile's contribution to conversions is visible, and the share of saved baskets completed on another device grows.
Prevent: report device performance with a named model and assisted figures, never on last click alone.
Mobile traffic that was never likely to buy
Social feeds, in-app display and video placements reach people on phones in idle moments. Their intent is low whatever device they use, but because they are mostly mobile they drag the mobile rate down. The intent signals guide shows how to read a source's intent from its placement and behaviour.
Fix: judge mobile performance within each source, send low-intent placements to pages with a lighter next step, and exclude placements that bring visits with no engagement at all.
Verify: within search and email, mobile and desktop rates sit closer together than the blended figures suggested, and mobile cost per conversion from paid social improves after exclusions.
Prevent: report conversion by device within source as standard, and review placements on a fixed rhythm.
A prevention routine for the mobile gap
The gap is easiest to manage when it is measured the same way every time and reviewed step by step. Monthly, read conversion by device within each main source, and the step rates by device for the main journey; note the step with the largest device gap and whether it moved since last month. Treat that step as the mobile priority until it closes. After every release, walk the journey on the two or three phones that bring most traffic, over a throttled connection, from landing page to confirmation, including each common payment method. The tutorial on checking conversion paths after a release turns that walk into a script with marked test records and a check at the far end. Before any new page or form is approved, review it on a phone first. Quarterly, check mobile field data for the templates with most paid traffic, review which placements send mobile visits with no engagement, and restate the device report under the agreed attribution model with assisted figures beside last click. Keep the target as the gap at each step, not the overall mobile rate, because the overall rate moves with traffic mix whatever the site does.
An outdoor clothing shop narrows its gap
An outdoor clothing shop converted 3.4% of desktop sessions and 1.1% of mobile ones. Within paid search alone the figures were 3.6% and 2.2%, so much of the headline gap was paid social, which sent 70% of mobile sessions. The device funnel showed the largest mobile loss at delivery details, where 58% of mobile shoppers continued against 81% on desktop: the address form had no lookup and the postcode field opened the full keyboard. Field data also showed the product template's largest element loading slowly on phones because of an uncompressed image carousel. The shop added postcode lookup, a wallet button at the basket and compressed product images. Within search, mobile conversion rose to 2.9%, and the delivery step on mobile reached 74%.
Where speed is the suspected cause, the page speed impact test measures what a faster build is worth in conversions before the work is funded across the site.
If an app is proposed as the answer to a weak mobile checkout, the guide to choosing between an app and the mobile website explains why an app puts an install step in front of the same problem and how to score the options.
Common mistakes
Comparing blended mobile and desktop rates. Judging speed from lab scores on office laptops. Designing on a monitor and checking on a phone afterwards. Leaving wallets out of mobile checkout. Crediting desktop with every sale mobile started. Setting the overall mobile rate as the target.
