For years, the craft of performance advertising centred on targeting. Teams built detailed audiences by interest, behaviour and demographics, layered lookalikes on top, and spent much of their time refining who would see an ad.
That approach has been weakened from two directions. Privacy changes, such as Apple's App Tracking Transparency, which since 2021 has required apps to ask permission before tracking people across other companies' apps, and browsers blocking third-party cookies, have reduced the signals available for detailed targeting. At the same time, the major platforms have moved towards broad, automated targeting, where their systems decide who sees an ad based on who responds to it.
The result is a shift in where the advantage lies. When the platform chooses the audience, the creative becomes one of the strongest signals it has about who the ad is for.
How creative now does the targeting
In automated campaigns, the platform shows an ad to a wide range of people and learns from who engages with it. The content of the ad, its images, words, format and tone, shapes who responds. That response then guides who the system shows it to next.
In practice, an ad that speaks clearly to first-time parents will tend to find first-time parents, and an ad that speaks to small business owners will tend to find them, even if both campaigns are targeted broadly. The creative filters the audience.
This has a significant consequence. A single ad trying to appeal to everyone gives the platform little to work with. A set of distinct ads, each aimed at a clear group or need, gives it much more.
What good creative looks like now
A few principles hold across most platforms.
| Be clear quickly. | On fast-scrolling feeds, the first moments of an ad decide whether anyone sees the rest. Show the product, the problem or the promise early. |
|---|---|
| Make distinct variations. | Rather than small tweaks to one idea, build genuinely different concepts aimed at different motivations, such as price, quality, convenience or identity. |
| Design for the platform. | Vertical video, sound habits and native styles differ between platforms. Ads that look like they belong tend to perform better than ads that look like they were moved from elsewhere. |
| Feature real use. | Showing the product in the context where it is used, by people the audience recognises, tends to help the platform connect the ad to the right people. |
Building a creative system
Because creative now does so much work, it needs to be managed as a system rather than produced as a series of one-off campaigns.
Plan volume and variety.
Automated campaigns benefit from a steady supply of new creative. Plan production to deliver regular variations, not a single large launch.
Test with a clear question.
Each test should answer something specific: which message, which format, which opening. Change one major thing at a time where you can, so the results are useful.
Learn from the results.
Record which concepts work for which audiences and markets. Over time, this becomes one of the most valuable things a marketing team owns.
Use AI where it helps.
Generative AI can speed up variations, resizing and adaptation into new formats and languages. It is most useful when it scales good ideas rather than replacing the thinking behind them. Human review remains essential for brand consistency and accuracy.
Treat creative as a system
With less data available for targeting and more automation in how ads are delivered, the quality and variety of creative has become a central lever of performance. The teams that win will be the ones that treat creative as a strategic system: clear ideas aimed at distinct audiences, produced at a steady pace, tested carefully and improved over time.
