Optimizing for Clicks Un-Brands Your Advertising

Every time you pick the creative that wins on click-through, you are voting against the assets that make the ad yours. The colors, characters, logos, sound and shapes that let a viewer know who is talking are exactly what gets stripped when creative is tuned for immediate response. Only about 16 percent of advertising is both recalled and correctly attributed to the right brand. An ad that is remembered but not attributed to you does not build nothing, it builds the category leader, because misattribution flows to the biggest brand in the set. The right scoreboard is not CTR, it is the strength of your distinctive assets and your rate of correct attribution, which together tell you whether you have a branding problem or a message problem.

Why optimizing for clicks mechanically un-brands

Response optimization does not just fail to build brand, it erodes it, through a chain the industry has named. Adam Morgan's Four Horsemen of the Dullocalypse are Performance, Optimization, Averaging and Procurement. Optimization is the direct culprit, the more placements you run across the simpler your assets must be to survive them, which is why so many logos flattened to a monotone sans-serif. The output is dull, and dull is expensive. System1 and Peter Field put the penalty at 2 to 2.6 times more media spend for a dull ad to match an interesting one, an estimated 189 billion dollars a year in wasted US spend, vendor figures read directionally. And this dulling floods the worst place for it, Karen Nelson-Field and VCCP found in 2025 that 85 percent of digital ads get under 2.5 seconds of active attention, enough to build memory in roughly 1.5 seconds only if distinctive assets are deployed well, so optimization strips your identity in exactly the low-attention channels where it has to register instantly.

Why the real cost is misattribution, not weak recall

The damage is not that people forget your ad, they remember it and credit someone else. The most cited figure, only about 16 percent of ads recalled and correctly attributed, traces to Byron Sharp's How Brands Grow citing an Ehrenberg-Bass study of 143 television ads, so it is best cited via Sharp. The Ipsos number is primary, about half who can recall an ad, without the brand name, cannot attribute it to the advertiser. Kantar's mobile Brand A case is sharper, its ad failed to break through for 59 percent of consumers, and among those who recognized it, it was credited to rivals as often as to the brand. The reason this is a growth problem, not a vanity metric, is where the misattribution goes, disproportionately to the category leader, so a challenger running under-branded performance creative is literally buying awareness for the incumbent it is trying to beat. The counterintuitive fix is that more branding is not the answer, a Kantar copy-test found no correlation between branding scores and how often the brand was shown. The brand has to be woven into the memorable moment, and the Ipsos test is simple, can the story be retold without the brand, because if it can, the branding is the problem.

Why asset strength is the scoreboard, not CTR

If CTR is the wrong metric, the right one is the strength of your distinctive assets on two axes. Jenni Romaniuk's Distinctive Asset Grid scores every asset on Fame, the share of category buyers who link it to your brand, and Uniqueness, how far they link it only to you. A working threshold around 50 percent on each axis marks an asset strong enough to be a solo brand cue. The newest peer-reviewed benchmark, from Phua, Sharp and colleagues in the International Journal of Advertising, analyzed 1,162 assets and found a mean Fame of 26 percent and mean Uniqueness of 54 percent, the average asset distinctive but not yet widely known. The decisive detail is which types are strong, shape-based assets like logos and packaging strongest at 40 percent Fame and 71 percent Uniqueness, color weakest at 12 and 39. Color is the one marketers most casually alter to chase a click and the least ownable, while the shape, logo and character assets that direct-response creative crops or drops are the strongest you have. Mark Ritson's caution belongs here, distinctiveness is necessary but not sufficient, you still need relative differentiation to command a price premium. It lets your advertising compound, differentiation defends margin once you are recognized.

Why you must separate a distinctiveness problem from a message problem

Once you measure asset strength and attribution, the two problems need opposite fixes. High recall but low correct attribution is a distinctiveness problem, fixed by deploying stronger assets earlier and more consistently, not a new message. High attribution but low persuasion is a message problem, where no amount of extra branding helps. Ipsos's Power of You meta-analysis of 2,015 US video cases confirms assets are the mechanism, high-performing ads deployed brand assets about 34 percent more often. The same data exposes the underuse, logos and color appear in most ads but every other asset type in fewer than half, and an Ipsos and JKR study of more than 26,000 consumers found fewer than one in five brand assets are truly distinctive. Most brands leave their strongest, most ownable cues on the table while optimizing the weakest.

Why AI is both the accelerant and the cure

AI is making the crisis worse by default and can reverse it by design. Because large models predict the most probable output, they regress toward the mean, driving homogenized content and model collapse, and the misfires are here, from Coca-Cola's AI holiday ads to McDonald's pulled Netherlands Christmas ad, omitting brand soul and defaulting to category-generic imagery. Pointed the other way, multimodal AI turns volume into brand-building. Feed a campaign's frames into a vision model to build a brand-presence timeline, what share carry a distinctive asset and how early the brand registers, since CreativeX finds only 28 percent follow best practice. Then encode your asset rules as a spec the model must follow, the character early and in most frames, the sonic logo on every cut-down, the shape never cropped, enforced by a validator that blocks off-brand variants. That is how scale becomes brand-building, variants that carry the same codes rather than subtly un-branded ones. The one thing AI cannot do is the Fame and Uniqueness test, which needs real light category buyers, not the model, since those scores require consumer data AI cannot fake. The limits are real, AI vision estimates presence, not memory, distinctiveness is built through years of repetition no model can shortcut, and optimizing for presence alone can produce mechanically consistent but dull work, the neutrality the cost of dull warns about.

The distinctiveness scoreboard (copy this)

Run this on your creative review before you crown another CTR winner.

• Add two metrics beside CTR on every review, an asset-presence score from a vision audit and a correct-attribution rate from testing. If a CTR winner scores low on attribution, log it as a loss.
• Run a Fame and Uniqueness baseline. Test your 5 to 10 candidate assets among light category buyers and plot them on the grid.
• Protect and prune by quadrant. Defend the assets scoring high on both, invest in the high-uniqueness ones, and stop casually altering any asset above the 50 percent thresholds, especially color.
• Write brand codes into your AI pipeline. Encode the non-negotiable rules, asset timing, frame coverage, sonic logo on every cut-down, no cropping the shape, enforced by a validator on every generated variant.
• Re-audit on a cadence. Vision-audit every new asset before launch, refresh Fame and Uniqueness twice a year, and treat a rising misattribution rate or falling asset-strength score as an early warning that optimization is un-branding you.

The full workflow, auditing asset presence across a campaign, designing the Fame and Uniqueness test, diagnosing a distinctiveness problem from a message problem, and enforcing brand codes on AI-generated variants, is packaged as a reusable Claude skill. Get the free skill.

What to do Monday

Change the scoreboard before the creative. Put an asset-presence score and a correct-attribution rate next to CTR on every review, and reclassify any click-winner that cannot be attributed to you as the loss it is. Run a Fame and Uniqueness baseline among light category buyers, protect the strong assets, and stop altering them to chase placements. Then encode your brand codes as rules your AI pipeline must follow, so scale carries your identity instead of eroding it. Recognition compounds and clicks do not, and if your advertising is not attributed to you, you are paying to build the brand you are trying to beat.

Sources: Byron Sharp, How Brands Grow, on recall and correct attribution, citing an Ehrenberg-Bass study of 143 TV ads; Ipsos, John Hallward on ad attribution, and the Power of You meta-analysis of 2,015 US video cases; Kantar, the Brand A break-through and misattribution case, and the Link copy-test branding finding; Jenni Romaniuk, Building Distinctive Brand Assets, the Distinctive Asset Grid; Phua, Sharp and colleagues, the distinctive-asset benchmark, International Journal of Advertising, 2026; System1, eatbigfish and Peter Field, the Cost of Dull, and Adam Morgan's Four Horsemen; Karen Nelson-Field and VCCP Media, 2025; Ipsos and JKR, the distinctiveness study of more than 26,000 consumers, and CreativeX best-practice data.