On one account we audited, the conversions split like this: sixteen from mobile phones, two from computers.
Nobody in that business had ever seen that number. It is not hidden and it is not hard to find. It sits behind one dropdown that most people never open, and until you open it every figure in your account is an average of two audiences behaving completely differently.
Segment by Device Before You Change Anything
The dropdown is called Segment, and Device is one of the options. Apply it and every row splits into mobile, desktop and tablet.
Do this before you touch a bid, a keyword or an ad. It costs nothing and it takes ten seconds, and it routinely changes what the rest of the account looks like.
What you are checking is simple. Where is the spend going, where are the conversions coming from, and are those the same place. When they are not, you have found something that no aggregate number could have told you, because an average of a good performer and a bad one looks like a mediocre performer and nobody investigates mediocre.
The sixteen-to-two account is a useful shape to hold in your head. Two conversions from desktop is not a small version of sixteen. It is a different question. Either desktop demand genuinely is not there, in which case the spend going to it needs justifying, or something in the account is preventing those ads from competing. Those two explanations lead to opposite actions, and the aggregate number cannot distinguish between them.
The Split Is Rarely What People Assume
The assumption is usually that consumers are on phones and businesses are on desktops, and it is unreliable enough that you should not plan around it.
We audited a manufacturer selling technical components to engineers. The instinct there is that a professional buyer researches at a desk during office hours. What we actually said to them was the opposite: engineers keep odd hours, they search for a specific part when the need arises, and it does not much matter when your office is open. The buying moment is not synchronized to the seller's working day.
That cuts both ways. Some local consumer services turn out to have a meaningful desktop share, because people research a considered purchase in the evening on a laptop. Some B2B accounts are overwhelmingly mobile.
The point is not that one answer is right. It is that the answer is in your account, it takes ten seconds to retrieve, and guessing at it produces confident decisions built on a stereotype.
A Bid Adjustment Is a Standing Instruction
Here is the finding that makes this chapter worth a Friday afternoon.
On a sports chiropractic account we audited, almost every conversion was arriving as a call from a mobile ad. Desktop was producing essentially nothing, and not in a way that looked like weak demand. It looked like the ads were barely appearing there, and the likely cause was a device bid adjustment sitting in the account and suppressing them.
Nobody in that business had set that adjustment recently. Somebody had set it once, for a reason that may well have been sound at the time, and it had been quietly executing ever since.
That is the nature of an adjustment. It is not a decision you made, it is a decision that keeps happening. And because it is expressed as a small percentage on a settings screen rather than as a number in a report, it never appears in the place where anybody would go looking for a cause.
Most of Your Adjustments Are Inert. One Is Not.
This is the part that has changed and that most advice has not caught up with.
If you are running Smart Bidding, which covers Target CPA, Target ROAS, Maximize conversions and Maximize conversion value, most manual bid adjustments are not being applied. The strategy sets bids per auction using far more signal than a device percentage carries, so your carefully considered plus twenty percent on mobile is largely decorative.
With one significant exception. A minus one hundred percent adjustment still works, because it is an exclusion rather than a bid.
Put those two facts together and you get the trap. An account that migrated from manual bidding to Smart Bidding at some point carries a settings screen full of adjustments that no longer do anything, plus any exclusion-strength adjustment that absolutely still does. The harmless ones create the impression that the screen is inert. The one that matters hides among them.
There is a related trap in the opposite direction. Because Target CPA treats a device adjustment as a change to the target rather than to the bid, an adjustment that reads like a modest bid nudge can be quietly loosening what you are willing to pay for a conversion on that device. It is doing something, just not the thing the number appears to say.
So the audit action is not to tune your adjustments. It is to find whether any of them amounts to switching a device off, and to confirm you meant it.
There is one more reason to look now rather than later. Every chapter in this series that told you to open a report assumed the report describes one population. Segmented by device, several of those reports say something different, and the search terms people type on a phone are frequently not the terms they type at a desk.
Mobile Is Not a Device. It Is a Different Visit.
Even where the split is healthy, a mobile visit and a desktop visit are not the same event, and treating them as one is how accounts underperform on the device producing most of their volume.
The mobile visit is shorter, more likely to be interrupted, and far more likely to end in a phone call than a form. That last point matters more than any bid setting: if most of your mobile conversions arrive as calls, then your call asset and its tracking are not a nice extra. They are the main conversion path for the majority of your traffic.
Page speed lands differently too. We routinely find sites scoring near the bottom on mobile and near the top on desktop, which is not a cosmetic problem when mobile is where the customers are. It feeds landing page experience, and therefore quality score, on the device carrying most of the spend.
Ad Schedule Is the Same Idea, Measured in Time
Device splits your traffic by how people arrived. The ad schedule splits it by when, and the same discipline applies.
Segment by day and by hour, and look at where cost and conversions diverge. You are looking for the same shape: money going out at times that produce nothing, or good hours you are underfunded in.
The mistake to avoid is aligning your schedule to your office hours by default. Those are two different things. Your office hours describe when you can answer. Your customers' search hours describe when they decide, and for anything researched rather than urgently needed, those overlap less than owners expect.
Before restricting a schedule, ask one question: what happens to a lead that arrives at nine in the evening? If it goes to voicemail and gets called back in the morning and converts at a normal rate, those hours are working and switching them off would cost you real business. If it vanishes, the problem is your intake, not your schedule.
The Twenty Minute Version
- Segment by device. Campaigns view, Segment, Device. Read cost against conversions for each.
- Compare the conversion split to the spend split. If they disagree materially, that gap is the finding.
- Open your device bid adjustments and read them as instructions rather than settings. Who set this, when, and does the reason still hold?
- Find anything at or near minus one hundred percent. Under Smart Bidding that is the one adjustment still doing real work, and it may be switching off a device you want.
- Check whether mobile conversions are calls. If they are, verify the call asset is recording, because that is your main path.
- Test your own site on a phone, on mobile data, not office wifi. Time it.
- Segment by day and hour. Look for spend in hours that never convert.
- Do not match your ad schedule to your opening hours without checking what actually happens to an out-of-hours lead.
None of this needs budget or a rebuild. It is one dropdown and one settings screen, and between them they describe who your customers actually are and when they turn up.
We segment by device on every Google Ads audit before looking at anything else, because it decides how to read every other number in the account. Hand it to us and we will do the pass properly. Or open the segment dropdown tonight and look at one month. If the conversions are stacked on one device and the spend is not, you have your afternoon's work.
