Display Advertising: What It Is Genuinely Good

Updated September 2026 · Written and maintained by the Progression Agency strategy team

Display advertising buys visual placements on sites and apps, shown to audiences selected by interest, behavior or context rather than to people who searched. That means low click rates by design, harder measurement, and real inventory-quality risk — and it also means reach that search cannot provide. This covers what display is actually good at, what to measure, and the maintenance that stops it drifting toward the cheapest impressions available.

What is display advertising?

Buying visual ad placements on websites and apps, shown to audiences selected by interest, behavior, context or a list you provide — rather than to people who searched for something.

That distinction is the whole thing. Search advertising reaches demand that already exists; display creates or reminds. Judging one on the other’s metrics is where most disappointment starts.

How is display different from search advertising?

Search captures intent that is already there. Display interrupts, which means lower click rates, different measurement, and a different job in the funnel.

A display click rate that would be catastrophic in search is normal in display, and comparing the two directly makes display look broken when it is working as designed.

The bottom row is the one advertisers new to display underestimate. Inventory quality varies enormously, and unmanaged campaigns can spend a meaningful share of budget in places nobody would choose.

What is display advertising actually good at?

Reaching people who do not know you exist, staying present through a long consideration period, and bringing back people who already visited.

Retargeting is where most businesses see the clearest return, because it addresses people who have already shown interest and simply did not act yet.

Retargeting site visitors

The highest-intent display audience available, and where most programs should start.

Reaching audiences before they search

Useful where the purchase involves a long consideration period and few people search early.

Staying present during a long decision

B2B and considered purchases where a search campaign alone reaches people only at the end.

Reaching customer lists

Uploading a list to reach existing customers or lookalike audiences.

Geographic and contextual reach

Placing ads where the context is relevant rather than where the person is known.

Supporting other channels

Display frequently makes search and direct traffic perform better without taking credit for it in the reporting.

What is display advertising bad at?

Producing immediate direct-response conversions cheaply, and being measured with the same attribution logic as search.

A business that needs inquiries this week is usually better served by search. Display is a different instrument and rewards a different time horizon.

Where do display ads actually appear?

Websites, apps, and increasingly connected television and digital out-of-home — bought through networks and exchanges rather than negotiated site by site.

The buying mechanism matters, because it determines how much control you have over where the ads run and how much of the spend reaches the publisher.

The control column is the reason inventory quality varies so much: the cheapest routes are cheapest partly because you know least about where the impression was served.

What should a display campaign be measured on?

Reach and frequency against the intended audience, assisted conversions, incrementality where you can test it — and view-through with appropriate skepticism.

Last-click attribution systematically undercounts display, and view-through attribution systematically overcounts it. Neither number alone is the answer.

Reach and frequency

Whether you actually reached the intended people, and how often. Frequency is the most commonly ignored and most commonly wasted variable.

Assisted conversions

Where display appeared in a path that converted, which is closer to the truth than last click.

View-through conversions

Treat as directional. A view-through counted after somebody scrolled past an ad they never saw is not a result.

Incrementality testing

Holding out a region or audience and comparing. The most trustworthy answer available and the least frequently run.

Brand and direct search volume

A working awareness campaign shows up as more people searching your name, which is a real and checkable signal.

Cost per thousand impressions

A buying efficiency measure, not a performance one. Cheap impressions that reach nobody relevant are not cheap.

What is frequency capping and why does it matter?

Limiting how often one person sees the same ad. Without it, a small audience absorbs the whole budget and the campaign becomes an irritation.

Excessive frequency is one of the most common and most fixable waste sources in display, and it actively damages brand perception rather than merely wasting money.

How do you avoid wasting budget on poor inventory?

Exclusion lists, placement reports read regularly, category and content exclusions, and viewability thresholds.

This is ongoing maintenance rather than a setup task. Inventory changes constantly, and a campaign left alone drifts toward the cheapest available impressions.

Read the placement report

The single most useful maintenance habit. It shows exactly where the money went.

Build and maintain exclusion lists

Apps and sites that generate clicks and no outcomes recur across campaigns.

Exclude irrelevant content categories

Both for brand safety and because relevance affects performance.

Set viewability requirements

An impression that never entered the viewport is not an impression in any useful sense.

Watch mobile app inventory specifically

Accidental clicks in apps are a well-known source of meaningless traffic.

Check the geography actually served

Broad location settings routinely deliver outside the intended area.

Does creative matter more in display than in search?

Substantially. In search the intent does the work; in display the creative has to earn attention that nobody offered.

This is why display programs with one static ad set underperform: the creative is the variable with the most leverage and the one most often left unchanged.

What makes display creative work?

One clear idea, legibility at small sizes, an obvious brand, and a specific reason to act — in that order.

Display ads are seen peripherally and briefly. Anything requiring reading fails, and anything that does not identify the advertiser is spending money on somebody else’s awareness.

The landing page row is not really a creative issue, but it destroys more display performance than any creative decision: an ad about a specific thing landing on a homepage asks the visitor to start their search over.

What does display advertising cost?

It is bought on cost per thousand impressions far more cheaply than search clicks — but cheap impressions and cheap outcomes are different things.

The relevant number is cost per outcome, and for display that is usually higher than search while reaching people search cannot.

How much budget does display need to work?

Enough to reach a meaningful share of the intended audience often enough to register — which means a small budget spread thinly achieves nothing.

Concentrating a modest budget on a narrow audience, usually retargeting, beats spreading it across a broad one, which is the most common small-budget mistake.

Should a small business run display advertising?

Retargeting, usually yes. Broad prospecting display on a small budget, usually not.

The reason is arithmetic: broad reach requires enough impressions to matter, and a small budget spread across a large audience reaches nobody often enough to have an effect.

How does display relate to programmatic?

Programmatic is how most display is bought — automated auctions across many publishers — rather than a different kind of advertising.

The distinction people are usually reaching for is between self-serve platforms and managed programmatic buying, which differ in control, minimum spend and cost structure.

What about privacy changes and third-party cookies?

Audience targeting that relied on cross-site tracking has become less reliable, which has pushed attention back toward contextual placement and first-party data.

The practical consequence for advertisers is that your own customer and visitor data has become more valuable relative to bought audience segments.

What should you do before starting?

Verify conversion tracking, decide what job display is doing, agree how it will be measured, and prepare more than one creative.

Agreeing the measurement before launch is what prevents the common argument three months in about whether it worked.

What audience targeting options actually exist?

Broadly: your own data, audiences the platform infers, contextual placement, and geography — and they differ in reliability as much as in reach.

Your own data is the most reliable and the smallest. Inferred audiences are the largest and the least verifiable, which is a trade worth making consciously.

The bottom row is not always wrong — for genuinely mass-market products it can be efficient — but for anything with a defined buyer it is where budgets disappear quietly.

Media impressions

The unit almost all advertising is priced in, and the one most loosely defined.

What is a media impression: one instance of an advertisement being delivered, which is not the same as being seen. The distinction between a served impression and a viewable impression — where a defined portion of the ad was actually on screen for a minimum duration — is the single most important thing to establish before comparing two prices, because comparing a viewable rate against a served one makes the better inventory look expensive.

In public relations the word is used differently again, where impressions frequently means the total audience of every outlet that carried a mention. That number is a reach estimate rather than a measurement, and quoting it alongside advertising impressions in the same report gives the two very different things a false equivalence.

Frequently asked questions