What is creative intelligence?
In advertising, creative intelligence is the practice of breaking ads into their elements, such as the hook, format, message, visuals and offer, and linking those elements to performance, so a team can learn which creative choices work for which audience and objective.
The term is also used in psychology for the ability to produce original ideas. This guide uses the advertising sense: evidence about creative, gathered from the ads a brand has already run.
- 01TagHook, format, message, visual, offer
- 02CompareSimilar ads, same channel and objective
- 03TestTurn a pattern into one defined change
What creative intelligence looks at
Every ad is a bundle of choices. Creative intelligence names them consistently so the same choice can be found across hundreds of ads.
| Element | Examples | Compare it against |
|---|---|---|
| Hook | The first seconds of a video or the first thing a static ad shows: a face, a product, a question, a result | Other hooks in the same format and placement |
| Format | Video length, static image, carousel, creator-style or studio | Ads with the same objective, since formats suit different jobs |
| Message | Benefit, feature, price, social proof, urgency | Ads for the same product and audience |
| Visuals | People or product first, setting, color, framing | Ads with a similar message, so the visual is the main difference |
| Offer and call to action | Discount, free trial, learn more, shop now | Ads at the same stage of the customer journey |
A pattern in past ads is a reason to run a test, not proof of what to make next.
Creative intelligence vs creative analytics vs creative testing
The three terms are related steps in one workflow:
- Creative analytics reports how each ad performed: spend, impressions, clicks and results by ad.
- Creative intelligence explains why, by connecting those results to the elements inside each ad.
- Creative testing checks an explanation by changing one element and comparing against the original under matched conditions.
How creative intelligence works
The steps are simple. The discipline is in keeping comparisons fair.
- Collect every ad with its delivery and results, and keep the creative file itself.
- Tag each ad's elements with one shared vocabulary, so the same hook or message is named the same way everywhere.
- Compare like with like: the same channel, objective and format, over a similar period.
- Look for elements that appear repeatedly in stronger or weaker ads, and note how many ads support each pattern.
- Turn the strongest pattern into a test with one defined change, and record the result back into the library.
Creative metrics to define once
Platforms count these slightly differently, so write your definitions down and keep them fixed. Common versions:
- Hook rate = 3-second video plays / impressions. How often the opening stopped the scroll.
- Hold rate = plays to a longer mark, such as 15 seconds or completion / 3-second video plays. How often the ad kept attention after the hook.
- Click rate = link clicks / impressions. How often attention became a visit.
- Cost per result = spend / results. What each outcome cost, on the platform's own attribution.
Where creative intelligence stops
Creative is one of several reasons an ad performs. Audience, budget, bidding, timing and the offer all move results too, so a creative pattern is evidence to test rather than a rule.
Prism, Mosaic's creative intelligence capability, works within that limit: it reviews ads against similar ads, explains which elements sit behind the results and proposes specific tests. It does not predict bookings, revenue or return on spend.
Frequently asked questions
Is creative intelligence the same as creative testing?
No. Creative intelligence finds patterns across the ads you have already run. Creative testing checks one of those patterns by changing a single element under matched conditions.
What data does creative intelligence need?
The ads themselves, their delivery and results by ad, and consistent tags for each element. Results without the creative file cannot say which choice made the difference.
Can artificial intelligence (AI) tag ad creative?
Yes. AI can describe what is in an image or video, such as people, products, setting and on-screen text, which makes tagging consistent at scale. A person should still review the vocabulary and spot-check the tags.
Can creative intelligence predict which ad will win?
It can point to promising patterns, but past performance depends on audience, budget and timing as well as creative. Treat its findings as hypotheses and confirm them with a test.
