adlaneSign in Practical guide
Google Ads device performance: compare mobile and desktop fairly
Device segmentation can reveal a different buying context or a task that breaks on one screen size. It can also reflect nothing more than a different mix of searches and campaigns. Compare devices within a coherent scope, inspect the experience on the weaker device, and only then consider any change to delivery.
By Adlane · Updated
Why Google Ads device performance differs before anything is wrong
A device report at account level blends every campaign together. If branded search traffic arrives mostly from mobile and broad research queries arrive mostly from desktop, the devices will look different even when each campaign performs identically on both. The gap belongs to the campaign mix, not to the device.
Query intent, ad position and time of day also shift by device. People research on a phone during the day and complete a purchase or enquiry from a desktop in the evening. Attributed conversions describe where the last reported interaction happened; they do not measure incremental sales and they do not describe the whole decision process.
Compare devices inside one campaign with enough activity
Choose a single campaign, or a group of campaigns that share an offer and a bidding approach, with enough spend and conversions for the comparison to mean something. Use the same complete dates, the same account timezone and the same conversion actions for every device. Record the figures in a worksheet before forming a view.
| Check | What to record | Why it matters |
|---|---|---|
| Scope | Campaign or campaign group, dates, timezone, conversion actions | Mixed scopes compare audiences, not devices |
| Volume | Spend, clicks and conversions per device | Small counts make cost ratios unstable |
| Intent | Top search terms and landing pages per device | Different queries explain many gaps |
| Experience | The task completed on each device | A broken form hides behind a cost ratio |
| Context | Bid strategy, device adjustments, recent changes | Explains what delivery is already doing |
Illustrative example: mobile research, desktop enquiry
Illustrative example: a consultancy campaign spends an invented 1,200 on mobile and 900 on desktop over four complete weeks. Mobile records three enquiries and desktop records nine. The simple cost ratio is 400 per enquiry on mobile and 100 on desktop, which looks like a reason to pull mobile spend.
The search terms tell a different story. Mobile clicks come from general service queries while desktop clicks come mostly from the brand name. Many desktop enquirers first met the brand on a phone. Reducing mobile delivery could reduce the desktop enquiries that the report credits to desktop. The ratio is real; the conclusion drawn from it is not supported.
Small conversion counts make device cost ratios unstable
Three conversions against nine is a difference that one or two extra enquiries would erase. Before treating a device gap as a pattern, ask how many additional outcomes would be needed to reverse it. If the answer is a handful, the evidence is still thin, and the right move is to keep observing with the same definitions.
A device with zero conversions has an undefined cost per conversion, not an infinitely bad one. Report the spend and the zero count together. Recent clicks have also had less time to convert, so treat the newest days as provisional rather than as proof that one device has stopped working.
Test the task on the weaker device before changing delivery
Open the landing page on the weaker device and complete the action the ad promises. Look for forms that need a desktop keyboard, pricing hidden by a layout that collapses on a small screen, phone numbers that are not tappable, or a handoff that asks a mobile visitor to finish later on a computer. A confirmed experience fault is fixed on the page, not with a bid.
If the experience is sound and the gap persists across mature periods, review any delivery change in the context of the campaign’s bidding strategy. Automated bidding already treats device as a signal, and device adjustments behave differently depending on the strategy in use. Read the current Google guidance for that strategy and write down the expected effect before proposing anything.
Run the device performance review in Adlane
Connect the Google Ads account for the business you are reviewing. Ask a bounded question with dates and a goal, for example: “Compare mobile and desktop for the Consulting Search campaign over the last four complete weeks, using the enquiry conversion action. Show spend, clicks, conversions and the top search terms per device.” Review the evidence returned before interpreting the ratios.
If the review leads to a change, Adlane prepares it as a proposal that shows the current setting and the proposed action for your approval. Nothing is applied by the conversation itself, and any new campaign starts paused. Use the proposal to record the hypothesis, the evidence and the date you will re-check the same devices with the same definitions.