How does display advertising work, and when should you use it?

Display advertising includes visual ads distributed across websites, apps and other digital environments. Banners, rich media formats and videos can be bought directly from a publisher or through automated systems. The format alone does not define the purpose: it can build familiarity, bring someone back to an offer or generate an immediate response.

The useful question is not whether banners work in absolute terms. It is which audience should receive which message, in what context and with what observable change.

When should you use display advertising?

Display can reach broad audiences before they actively search for a solution, so it can support launches, awareness and consideration. It can also remind people who have already interacted with a product, provided the pressure is proportionate and respects their preferences.

It is less suitable when the objective, audience and proposition are vague. Buying many low-cost impressions cannot compensate for irrelevant creative. Before investing, define the channel’s role in the customer journey and the alternative against which you will compare it.

If demand already exists and is expressed through search, search ads may capture it more effectively. If a new category needs to be explained, display can distribute a visual message, but it will require repetition and continuity.

How do you choose audiences and placements?

You can select editorial contexts, topics, geographical areas, audience characteristics or available behavioural signals. Apparent precision does not guarantee accuracy. Very narrow segments can cost more, limit learning and rely on weak inferences.

Build a list of suitable environments and another of exclusions. Check where ads actually appeared, not only where they were supposed to appear. Consider brand safety, editorial quality, viewability and non-human traffic.

Context can contribute to meaning. The same message alongside a relevant article and on a page crowded with ads creates different experiences.

How do you design the creative?

An ad has little time. It must make its subject, promise and action recognisable without demanding too much effort. Avoid tiny text and too many messages. Prepare variants for different formats instead of letting the platform crop them at random.

The landing page must continue what the ad promises. If the click leads to a generic home page, the visitor must reconstruct the journey. Speed, readability and visual continuity affect the outcome as much as the banner.

Test one meaningful difference at a time: proposition, image, audience or context. Changing everything together makes it difficult to understand why one variant worked.

How many times should an ad be shown?

One exposure may not be remembered, but excessive frequency wastes budget and causes irritation. Examine the frequency distribution, not only the average: some people may receive many impressions while others see only one.

Set limits appropriate to the duration, objective and stage. Refresh materials during a long campaign. Exclude people who have already completed the action when there is no useful follow-up message.

How do you measure results?

Impressions and reach describe distribution. Viewability estimates whether the ad had an opportunity to be seen. Clicks and conversions show responses, but attribution may give credit to exposures that did not change the decision.

For awareness objectives, assess qualified reach, frequency, recall or changes in searches. For direct response, check cost, quality, margin and new customers. Where possible, compare exposed and unexposed groups or comparable periods.

Display advertising is useful when treated as an experiment in communication and distribution. It becomes noise when success means only the number of times a file loaded on a screen.

How often should ads be refreshed?

There is no universal expiry date. Monitor frequency, declining response and elapsed time, but interpret these alongside the audience reached. Creative may seem tired because it is shown too often to the same people or because the proposition was irrelevant from the beginning.

Prepare variants that change the argument or evidence, not only the colour. One might clarify the problem, another show the outcome and a third answer an objection. Keep enough elements consistent to recognise the brand and understand which assumption you are comparing.

Archive versions, dates, audiences and destinations. Without a record, the team repeats rejected ideas or attributes the outcome to the wrong detail. Before replacing an ad, also check whether the page, availability or terms are the true cause of the decline.