Why Creative Is Becoming the Biggest Performance Lever in Digital Media

Creative is no longer just what the platform delivers

It is tempting to think of creative as the final layer of a campaign – the message or image carried by the media. In automated advertising, that is no longer the whole picture.

Meta describes Andromeda as the retrieval stage of its advertising recommendation system. Its role is to narrow a pool of eligible ads to the candidates most likely to be relevant before more detailed ranking takes place. Meta’s latest technical explanation describes ad content being combined with real-time context as part of that matching process.

Google gives advertisers similar advice from a different angle. Its guidance for AI-powered campaigns asks for a range of on-brand headlines, descriptions, images and video formats so its systems have more relevant combinations to test and deliver.

These are platform sources, so their performance claims should be treated as such. But the direction is clear: creative is not simply shown after the system finds an audience. The assets, messages and formats supplied by the advertiser give that system options from which to learn and match.

In other words, creative has become an input to delivery.

That changes the job. A campaign with one generic proposition presented in ten slightly different layouts may contain ten files, but it still gives the platform one idea. A campaign built around several genuine customer problems, reasons to believe, emotional angles and formats creates a much richer set of possibilities.

The distinction is not between having less creative and more creative. It is between variation and diversity.

More assets do not automatically create more learning

We have previously written about the growing importance of creative diversification in paid media. The next challenge is to be more precise about what diversification means.

Changing a crop, background colour or headline may be useful for executional optimisation. It does not necessarily test a different reason why someone might care.

Meaningful creative diversity starts further upstream:

• Which customer problem are we addressing?
• Which benefit leads the message?
• What evidence makes the claim believable?
• Is the customer responding to reassurance, aspiration, urgency or practical utility?
• Does the idea suit the way the platform is actually consumed?

• Is the work recognisably ours?

Each concept should represent a hypothesis about the audience, not simply another item in the asset library.

That makes creative testing more commercially useful. Instead of learning that one colour or crop outperformed another, the business can learn that a particular problem, product benefit or proof point is more persuasive. Those findings can then influence landing pages, audience strategy, product positioning and future briefs.

Performance data becomes an input into creative strategy, rather than a weekly instruction to make more versions of the last winning ad.

Attention is essential, but it is not the outcome

The first job of any advert is to earn enough attention to communicate something.

Viewability only establishes that an advert had the opportunity to be seen. The IAB and Media Rating Council now position attention as a complementary signal that can add context beyond delivery metrics, while explicitly warning against treating it as a standalone measure of effectiveness.

That qualification matters – an advert can attract attention because it is visually striking, confusing or completely out of character for the brand. It can generate a click without creating trust. It can be memorable while leaving people unable to identify who it was for.

Creative effectiveness therefore needs several layers of evidence.

Predictive attention and creative-scoring tools can help identify whether branding, messages and calls to action are likely to be noticed before spend is committed. We have explored that role previously. Live testing can then show how different concepts perform in market. Brand, conversion and commercial data establish whether that response was valuable.

No single measure should be allowed to redesign the work by itself.

Academic evidence supports that broader view of creativity. A meta-analysis covering 93 datasets found positive effects from advertising creativity, but also found that creativity works better when it combines originality and appropriateness rather than originality alone. The work needs to be different enough to be noticed and relevant enough to mean something.

Attention opens the door. Relevance, recognition and persuasion determine what happens next.

Distinctiveness and optimisation have to work together

Automated platforms are designed to find and exploit patterns. That is useful, but it can encourage short-term convergence.

A particular execution performs well, so investment moves towards it. New versions imitate it. The message becomes narrower. What began as optimisation gradually turns into repetition.

The immediate numbers may still look efficient, while the brand becomes less distinctive and the creative system learns less.

Kantar and WARC matched around 450 ads from their respective creative quality and ROI databases and found that ads combining strong short-term and long-term qualities generated more than four times as much profit as weaker work. Their analysis also found a stronger relationship between profit and measures connected with future brand demand than might be expected from a purely short-term view.

This does not make every performance advert a brand campaign. It does mean conversion-focused creative still needs recognisable brand assets, a coherent point of view and a reason to be remembered.

Variety should broaden the ways a brand can connect without making every execution feel as though it came from a different business.

The new bottleneck is not production

AI has made it possible to generate and adapt assets far more quickly. Templates, feeds and dynamic creative systems can handle products, prices, locations and formats at a scale that manual production could not.

That solves a production problem. It does not automatically solve the creative one.

If a weak idea can now be turned into fifty assets instead of five, the result is simply more efficient mediocrity. Worse, it creates more material for media and marketing teams to approve, traffic, analyse and explain.

The real constraint is increasingly the operating model around the work.

Many organisations still brief creative after the channel plan has been agreed. Creative, media and data teams work to different timelines. Performance gets reported by campaign or ad ID, while the underlying proposition or concept is lost. Insights sit in dashboards but rarely make it back into the next brief.

Under those conditions, producing more assets adds activity without adding much intelligence.

A stronger model connects five things:

1. Audience and commercial insight – what behaviour or barrier are we trying to change?
2. Creative hypotheses – which genuinely different ideas are worth testing?
3. Platform and placement design – how should each idea behave in its environment?
4. Measurement – what will attention, response and commercial data each tell us?
5. Learning – how will the result change the next decision?

The advantage does not come from having the fastest production process. It comes from reducing the distance between insight, idea, delivery and learning.

Creative is becoming a performance system

Creative is not the only determinant of digital performance. The offer, product, landing experience, budget, measurement and quality of the conversion signal still matter. No serious marketer should pretend otherwise.

But creative is becoming one of the most valuable variables marketers can actively shape.

It influences whether people notice the brand and understand the proposition. It gives automated platforms different options to match with people and placements. It determines the quality of the hypotheses being tested. And, when measurement is connected properly, it produces learning that reaches beyond the media campaign.

That requires creative, media, data and technology to operate as one system – an approach we are continuing to develop at Space & Time.

The platforms may decide which advert has the opportunity to appear.

Creative gives the system something worth choosing – and gives the audience a reason to care.