The short version. A bid adjustment raises or lowers your bid for a specific device, location, time, audience, or demographic. In the manual-bidding era they were a primary lever. Under Smart Bidding, most of them are signals the algorithm reads and largely sets for itself, not dials you turn by hand. That shift changes what they are for: diagnosis, not tuning.
Stop hand-tuning adjustments Smart Bidding ignores. Start reading them as a diagnosis of where spend and conversions have come apart.
A bid adjustment is a modifier, expressed as a percentage, applied to your bid in a defined context: a device, a location, an hour of the week, an audience, or a demographic. Set a +20% mobile adjustment under manual bidding and every mobile bid rose 20%.
Under Smart Bidding, that logic mostly no longer applies. Target CPA and Target ROAS set their own device, time, and audience bids from live signals, and they generally ignore the manual adjustments you enter. The important exception is exclusions: a -100% device or location adjustment still removes that segment entirely. Everything between is now advisory at best.
So the question is not what percentage to set. It is what these breakdowns tell you about the account. Read correctly, they are one of the cheapest diagnostics in Google Ads. Read as dials, they are wasted effort.
The device breakdown surfaces the single most common misallocation I find. Across 28 accounts, mobile absorbs a median 72% of spend. In 24 of those 28 accounts, mobile converts worse than desktop, with a median conversion rate 28.6% lower.
This is not an argument to cut mobile. Mobile volume is real, and for many businesses the mobile visit precedes a desktop conversion. It is an argument to look. When 72% of budget flows to the device that converts 28.6% worse, that gap is either justified by your funnel or it is a leak. Most accounts have never checked which. The device breakdown answers it in one report.
If device is the adjustment people set backward, time and demographics are the ones they never open at all.
Real dayparting, a schedule that is not simply on all day every day, exists on only 3 of 28 accounts, covering 0.7% of campaign settings across the corpus. Income is barely on the map: only 13 of 33 accounts even populate an income breakdown, and exactly one account ever sets a nonzero income bid adjustment.
Under Smart Bidding, that is partly rational. You are not meant to hand-set an hourly bid curve the algorithm will override. But populating and reading these dimensions is not the same as tuning them. An account that never looks at its own performance by hour, or by income tier where relevant, is flying blind on two axes where real money hides. The adjustment being advisory does not make the data worthless. It makes the data a diagnostic you are throwing away.
Profit Forensics
I examine one week of your Google Ads account and find the money it is losing, or prove it is clean. Signed, either way. For accounts spending $50,000 or more per month.
| Dimension | Manual adjustment | What to do instead |
|---|---|---|
| Device | Mostly ignored (exclusions still apply) | Diagnose the mobile-desktop mismatch; fix funnel or tracking |
| Schedule (dayparting) | Overridden by Smart Bidding | Read performance by hour; exclude only clear waste windows |
| Location | Mostly ignored; exclusions apply | Exclude non-converting geographies outright |
| Demographic / income | Rarely used, largely advisory | Read the breakdown for mismatch; act through targeting, not modifiers |
Do bid adjustments still work with Smart Bidding?
Mostly not as dials. Target CPA and Target ROAS set their own device, time, and audience bids and generally ignore manual adjustments. The exception is exclusions: a -100% adjustment still removes a segment. Use the breakdowns for diagnosis and use exclusions for hard cuts.
Should I lower my mobile bid adjustment if mobile converts worse?
Usually not directly, because Smart Bidding will discount the manual modifier. First confirm the mismatch is a real leak and not a funnel where mobile assists a desktop conversion. If it is a leak, fix the mobile landing experience and conversion tracking, and exclude only if a segment clearly should not run.
Is dayparting worth setting up?
Setting a manual hourly bid curve rarely helps under Smart Bidding, which is why almost no one does it. Reading performance by hour is still worth it: it can reveal windows of pure waste you can exclude, and it tells you whether your budget and conversions align across the week.
Which bid adjustment matters most?
Device, as a diagnostic. It exposes the most common misallocation, budget flowing to a worse-converting device. After that, exclusions for clearly non-converting locations. Fine-tuned percentage adjustments on other dimensions rarely move the outcome under Smart Bidding.
The proprietary figures here come from a forensic analysis of Google Ads accounts representing $133M in managed spend, September 2024 to February 2025. Device figures cover 28 accounts (median mobile spend share and the mobile-desktop conversion-rate gap). Dayparting adoption is measured across 28 accounts and 15,168 campaign-setting rows; income-adjustment adoption across 33 accounts. Related reading: Target CPA bidding and portfolio bid strategies.