What a display campaign is and where ads run
Display campaigns run on the Google Display Network (GDNGoogle's network of display ad placements.): the publisher sites and mobile apps that sell inventory through Google, together with Google-owned surfaces such as Gmail and YouTube where Google documents that display ads can appear. The advertiser does not choose a page; the system matches each ad slot to the targeting on the campaign and bids for it in real time. Most of the ads are responsive display ads, which Google assembles from uploaded images, logos and text to fit whatever slot is won, although uploaded image ads in fixed sizes are still supported.
Three features of the auction matter from the first day. Pricing is impression-led: even when a campaign is bid on clicks or conversions, the underlying cost of inventory is best understood through cost per thousand impressions (CPMCost per 1,000 impressions.), which is why the CPM calculator is the right first tool for a display budget. Delivery is broad: a single campaign can serve across a vast number of sites and apps in a day, so inventory quality is something the advertiser manages rather than assumes. And attention is low: people are on the page for the content, not the ad, so creative has to earn a glance before it can earn a click. Buying the same open-web inventory through a demand-side platform, with direct deals and more granular controls, is the subject of the programmatic section of this site; the mechanics of the Google Ads version are covered here.
When display is the right tool (and when it is not)
Display works when the job is to be seen by a defined group of people, repeatedly, at a low price per impression. The clearest cases are RemarketingAdvertising to previous visitors or engagers. to people who have already visited the site, staying present with a known audience between purchase cycles, and introducing a brand to a new audience that is defined by interests or by the content they read. It also suits launches where the aim is to build recognition before a search campaign can capture the demand.
Display is the wrong tool when the brief is to harvest demand that already exists. Someone who has just searched for a product is better reached by a search ad; the display impression they see later is a reminder at best. It is also a poor fit when the measurement plan only credits clicks, because display clicks are scarce and the people who do click are not representative of the people the ads influence. Advertisers who judge display by cost per click alone usually conclude it does not work, which is a measurement problem more than a media one. The media-planning view is that display earns its place in a mix by what it does for the other channels, and should be planned on ReachUnique people who saw an ad. and FrequencyAverage times a person saw an ad. rather than on clicks.
The core guides
The core guides divide the subject cleanly.
- Display targeting options covers who sees the ads: audience segments, your own data, content targeting, optimised targeting, how to layer them for reach or for precision, and how to read the audience and placement reports.
- Responsive display ads covers what they see: the asset requirements, how Google assembles a responsive ad for each slot, creative rules that hold up at every size, and how to test assets.
Two neighbouring sections share the same auction and much of the same audience catalogue. The YouTube ads hub covers video campaigns, which use the same audience segments but have their own formats and counting rules. The bidding section covers the strategies that decide what an impression is worth, which matters doubly for display because the inventory is so varied.
Further guides
Further guides put targeting and creative to work on particular jobs:
- Launching a display campaign covers a launch in order: one job, a matching bid strategy, targeting, optimised targeting, exclusions, frequency, ads, measurement and a first review.
- Excluding existing customers from prospecting covers keeping prospecting on new customers: defining an existing customer, building exclusion lists, settings by campaign type and measuring it.
Placement hygiene: exclusions and brand safety
Because delivery is broad, a display campaign left on default settings will spend part of its budget on placements the advertiser would never have chosen: made-for-advertising sites, low-attention mobile game apps, parked domains and content next to which the brand should not appear. Google provides the controls, and in practice they need to be set rather than assumed.
- Content exclusions and inventory type restrict the kinds of content the campaign can serve against; Google documents inventory tiers that trade reach for sensitivity.
- Placement exclusionsLists of websites, apps, video channels or app categories where an advertiser's ads must not appear. block named sites, apps, YouTube channels and app categories, and can be applied at account level so that every campaign inherits them.
- Topic exclusions remove whole subject areas.
- The "where ads showed" report lists every placement that received impressions, and is the source of the exclusion list. Reviewing it in the first days of a new campaign, and on a schedule afterwards, is the single most valuable routine in display management.
The targeting guide covers these controls in detail, including the difference between excluding a placement and excluding a category.
Measuring display: view-through, reach, incrementality
Display is seen far more than it is clicked, so the measures have to include what happens after an impression. Google Ads reports view-through conversions: conversions that follow an impression without a click, within the window the advertiser sets. They are evidence of influence, not proof of it, because some of those people would have converted anyway. Reach and frequency reports show how many distinct people saw the ads and how often; the reach and frequency calculator helps plan the budget needed to reach an audience at a sensible frequency before the campaign runs.
The honest measure is incrementality: did the group that saw the ads convert more than an equivalent group that did not? Google offers conversion lift studies for eligible accounts, and a geographic holdout can be run without them. The measurement section covers conversion tracking, attribution windows and the gap between platform-reported and analytics-reported numbers, all of which bite harder on display than on search.
A furniture retailer reads its display campaign two ways
A furniture retailer runs a remarketing campaign to recent visitors and a prospecting campaign to in-market segments, spending about £6,000 a month across both. Read on clicks, prospecting looks hopeless: a 0.3% click-through rate and a cost per click-based conversion of £140, against £35 on search. Read on view-based measures, the picture changes: the prospecting campaign reached 410,000 distinct people at an average frequency of 4, and the view-through conversions attributed to it, if even a third were truly incremental, would bring its effective cost per acquisition to about £48. The retailer runs a four-week geographic holdout in two comparable regions and measures a 9% lift in total conversions where display ran. That figure, not the click-based one, sets the budget for the following quarter.
Common display problems
| Symptom | Likely cause | Where to fix it |
|---|---|---|
| Budget spent mostly on mobile apps and games | No app or placement exclusions; optimised targeting chasing cheap impressions | Targeting guide, exclusions section; diagnosis: display ads on low-quality placements |
| Ads shown far beyond the chosen audience | Optimised targeting or targeting expansion enabled by default | Targeting guide, optimised targeting section |
| Ads look wrong in some slots: cropped images, clipped headlines | Assets built for one size; missing square or landscape variants | Responsive display ads guide |
| High impressions, almost no engagement | Creative that does not read at a glance; audience too broad | Responsive display ads guide, creative by objective |
| Conversions reported by Google Ads far above analytics | View-through conversions counted in one and not the other | Measurement section |
| A campaign barely spends its budget | Stacked targeting, small lists, frequency caps, exclusions, strict targets or ads not serving | Display or video campaign not spending |
| Same person sees the ad many times a day | No frequency cap; small remarketing list with a large budget | Targeting guide, remarketing section |
| Cost per impression rising with no change to the campaign | Bidding strategy constrained by targets; inventory competition | Bidding section |
| Display or partner clicks that never turn into leads or sales | Low-intent placements, location reach or a landing page that does not convert | Clicks but no conversions diagnosis |
| A new campaign spends on the cheapest impressions from its first day | Default settings left in place: expansion on, no exclusions, no frequency cap | Launching a display campaign |
| Prospecting results look strong but bring few new customers | Existing customers not excluded, and bidding finds whoever converts most easily | Excluding existing customers |
The Google Display Network expertise page describes how the author approaches display planning and placement control in practice.
