A device breakdown splits your advertising performance metrics by the type of device a person used, showing how mobile, desktop, and tablet each contribute to impressions, clicks, spend, and conversions. It re-slices metrics you already have instead of adding new ones; a device breakdown reports where delivery landed, and conversion-event columns report what happened after the click.
If you read a campaign only on blended totals, you can cancel spend that is working on one device while it fails on another. Pull the device split and you can shift budget, fix a slow mobile landing page, or tighten targeting before a weak device quietly drains the account.
A device breakdown is a segmentation view in an ad platform's reporting that groups the same performance data by device category, usually mobile, desktop, and tablet. In ChatGPT Ads Manager it lives in the Segment menu, which controls device and country breakdowns and sits apart from the metric columns and conversion-event settings.
Every row in a device breakdown carries the metrics you already track, impressions, clicks, spend, CTR, average CPC, average CPM, and conversions, recalculated for each device group. Nothing new is measured; the totals are partitioned so each device shows its own share. For the view to be meaningful you need enough volume in each group, since a device with a handful of clicks produces rates you cannot trust.
A device breakdown answers a delivery question: where your audience engaged. Conversion-event columns answer a different question about post-click actions, and country or platform segments slice the same data along other axes. AirOps ties device-level ad performance to the same prompt and visibility data behind your organic AI search work, so paid and organic read from one system.
Resources: See how to set targeting, creative specs, and conversion tracking for ChatGPT ads
A device breakdown is produced by the reporting system once you open the segment, so your job is to turn it on and read it correctly. Here is the sequence it runs through.
Collect delivery. The platform logs every impression, click, and conversion with the device category attached to each event.
Open the segment. In Ads Manager you add the device segment from the Segment menu, which re-runs the current report split by device.
Partition the metrics. Each existing metric is recalculated per device group, so mobile, desktop, and tablet each get their own impressions, spend, and conversion figures.
Compare the rows. You read CTR, CPC, and conversion rate across devices to find where efficiency and volume concentrate.
Act on the gap. You shift budget, rework a device-specific landing page, or adjust targeting based on what the split shows.
The output tells you how each device contributed to delivery and cost over your chosen date range. It does not tell you who the person was or follow one user across devices, since the data stays aggregated and non-identifying.
The device split decides whether you can trust a spend decision. ChatGPT ads already give you limited context, so the few breakdowns you do get carry real weight when you judge whether to keep paying for a campaign. A device that looks fine in the blended view can be the reason your cost per conversion refuses to drop.
It protects budget decisions: blended totals hide a device that burns spend at a high CPC with no conversions, and without the split you keep funding it.
It guides creative and landing pages: a mobile-heavy audience needs fast, short pages, and the breakdown shows when mobile clicks are not converting so you fix the page instead of the ad.
It sharpens cross-channel reads: when you compare ChatGPT ads to search and social, device mix explains part of the cost difference, and ignoring it makes the channel look worse or better than it is.
SEO managers use a device breakdown to confirm whether mobile or desktop drives the ChatGPT ad clicks that later convert on their landing pages.
Content strategists use a device breakdown to decide whether a paid landing page needs a shorter, mobile-first layout before the next campaign flight.
Growth marketers use a device breakdown to reallocate budget toward the device where cost per conversion is lowest across their active ad groups.
The device breakdown comes from the Segment menu, which re-slices existing metrics by device, and it is separate from the Edit columns controls that add conversion events, so mixing them up leads you to look for a device split in the wrong place.
A device group needs enough impressions and clicks before its CTR, CPC, and conversion rate mean anything, because a small sample swings those rates wildly and tempts you to act on noise.
Device data in ChatGPT ads is reported in aggregate and does not follow a single person across devices, so a user who clicks on mobile and buys on desktop will not appear as one connected journey in the breakdown.
Reveals which device drives conversions in ChatGPT Ads Manager instead of a single blended number.
Protects spend by exposing a device that clicks but never converts.
Guides landing page fixes when mobile engagement lags desktop.
Sharpens cross-channel comparisons against Google Ads and Meta device data.
Speeds budget shifts toward the lowest cost-per-conversion device.
Flags platform or operating system patterns worth testing in future ad groups.
Add the device segment before you judge a campaign, so your first read already shows mobile, desktop, and tablet separately.
Set a minimum click threshold per device before trusting its rates, because thin groups produce unstable CTR and conversion numbers.
Pair the breakdown with your own analytics, since ChatGPT ads report aggregated data and your site sees the actual device behavior after the click.
Review the split over a stable date range, because changing attribution windows mid-period shifts conversion counts and distorts the comparison.
Act on the gap you find, moving budget or fixing a device-specific page, so the breakdown drives a change and not just a chart.
Avoid treating a low-volume device group as a verdict. A device with a dozen clicks can show a wild CTR or a zero conversion rate that vanishes once volume builds, and pausing it early throws away data you needed to make the call.
AirOps: Connects your ChatGPT ad performance, including device-level results, to the prompt and visibility data behind your organic AI search, so paid and organic sit in one view.
ChatGPT Ads Manager: Provides the native device segment where you split impressions, clicks, spend, and conversions by device category.
Google Analytics 4: Captures the on-site device behavior after the click, letting you check whether ChatGPT ad clicks on a given device actually convert.
Open the segment. This week, go into ChatGPT Ads Manager, open an active campaign, and add the device segment from the Segment menu to see your current numbers split by device.
Set a baseline. Export the split over your last full reporting period so you have a stable starting point before you change anything.
Check volume. Mark any device group with too few clicks to trust, and set it aside until it collects enough data to read.
Connect your analytics. Confirm that your site analytics tags device category on ChatGPT ad traffic, so you can verify conversions the platform reports in aggregate. Google Analytics 4 or your own tool can show the device mix your ad actually drove.
Act and recheck. Make one change based on the split, shifting budget or fixing a device-specific page, then pull the breakdown again after a full period to confirm the move worked.
A device breakdown groups ad metrics by the device category a person used, usually mobile, desktop, and tablet.
You produce it by adding the device segment in your ad platform, which re-slices existing numbers without measuring anything new.
Its main constraint is volume, since a device with few clicks gives rates too unstable to act on.
Its main risk is reading aggregated data as if it tracked individuals, which it never does.
The leverage sits in reallocation, because the split shows exactly where to move budget or fix a page.
A device breakdown and a conversion-event breakdown answer two separate questions, and they live in two separate controls. The device breakdown comes from the Segment menu and splits your existing metrics by the device a person used, telling you where delivery and spend landed. A conversion-event breakdown comes from the Edit columns settings and splits your single conversions total into the individual events behind it, telling you which actions people took. Marketers often open the Segment menu expecting to separate purchases from sign-ups and leave confused, because that split is not there. The Segment menu is the where axis of a campaign, and the conversion-event columns are the what axis of the same data. You will usually want both: the device view to judge delivery efficiency, and the event view to judge outcome quality. Using one when you needed the other is a common reason a reporting read goes sideways.
Check the device breakdown on a cadence that matches your spend and your conversion lag, and skip the fixed daily habit. For most campaigns a weekly review is enough, because conversions in ChatGPT ads can take a day or two to appear and daily readings mostly show noise. If you are spending heavily or running a short test, look every few days, but wait until each device group has collected enough clicks to produce stable rates. Avoid rechecking right after you change an attribution window, since the conversion totals shift for reasons that have nothing to do with device performance. The practical rule is to review when you are about to make a decision: before a budget change, before pausing a device, or before reporting results upward. Reading it more often than you will act on it only tempts you to react to swings that settle on their own.
The device breakdown varies between ad groups mainly because the audiences and contexts behind them differ, and device is downstream of that. Different context hints reach people in different situations, and someone asking a quick question on a phone behaves differently from someone researching on a desktop. Ad group volume also matters: a group with few clicks will show a lopsided device split that evens out as it collects more data. Creative and landing pages play a part too, since a page that loads slowly on mobile will depress mobile conversions and tilt the split toward desktop. Time of day and day of week shift the mix as well, because device usage patterns change across a week. None of this means one ad group is broken. It means the device split reflects real differences in who each group reached, which is exactly why reading it per group is more useful than reading one blended number.
Not directly, and it helps to be honest about that. You cannot pick which device a ChatGPT ad serves the way you might set a bid adjustment in other platforms, because the device breakdown reports where delivery happened and does not let you dictate it. What you can do is influence it indirectly. Your context hints shape which situations you reach, and those situations skew toward certain devices, so a hint about on-the-go tasks will tend to pull mobile. Your landing page decides whether clicks on a given device convert, and a faster mobile page lifts the mobile numbers you later read in the split. Campaign settings such as platform targeting also move the mix. So treat the breakdown as a signal you steer through creative, pages, and targeting, then read the result to confirm the steer worked. Trying to control device directly will only frustrate you; shaping the inputs is the realistic lever.
There is no universal benchmark, and anyone who hands you one for ChatGPT ads is guessing. OpenAI has said it does not yet publish performance benchmarks across advertisers or industries, so a healthy device split is defined against your own business instead of an external standard. A good breakdown is one where each device you care about has enough volume to read, where cost per conversion on your priority device sits at or below your target, and where no single device is quietly consuming spend without returning anything. Judge it by whether it changes a decision: if the split tells you to move budget, fix a page, or keep going, it is doing its job. Compare it to your own prior periods to see direction, since trend tells you more than a snapshot. The worst version is a breakdown you cannot act on because every group is too thin to trust, which points to a volume or targeting problem instead of a performance verdict.