SUNDAY SCARIES: Algorithmic Camouflage

Dentsu Creative's What's Trending called out a reality that needs more attention: Algorithmic Camouflage. Brands are adapting content to mimic whatever platforms seem to reward; not to express what they stand for, but to stay visible. DC is right about the behavior. But the mechanics underneath are nastier than even Dentsu spelled out.

1. The Conditions

Two of the biggest platforms in the world rewrote their operating systems within months of each other. Both should have cited Niccolo Machiavelli as their corporate strategist.

TikTok's U.S. algorithm is being retrained on American-only data following the January Oracle restructuring. Creators are reporting sudden view drops and inconsistent distribution during the transition. The completion rate threshold for viral distribution has reportedly jumped from 50% to 70%.

Forrester predicts trends under the retrained algorithm will “feel distinctly American” — which sounds worrying in itself — but the real consequence is that the cultural inputs shaping brand behavior on the platform are narrowing.

Meta's Andromeda engine completed global rollout last fall, replacing the ad delivery infrastructure that had powered Facebook advertising since inception. As of February, detailed targeting inputs are treated as suggestions. The platform has politely reminded brands that it's in charge now.

Two trillion-dollar black boxes. Brands' rational response to systems they cannot read is to copy whatever worked last week. Multiply that across thousands of brands and the monoculture is just arithmetic.

Takeaway: At scale, camouflage stops being a survival strategy. It becomes the uniform.

(Sources: Forrester via PostEverywhere; Social Media Examiner; 1ClickReport, "Meta Andromeda 2026 Update")

2. The Catch

Here's where it gets structurally ugly.

Meta's Andromeda uses computer vision to assign every ad creative an Entity ID — a semantic fingerprint based on visual DNA, not copy changes. If two creatives share visual similarities, Meta groups them into one Entity ID. A brand launching 100 ads may find only 10 are being served sufficiently. All that iterative labor — swap a headline, trim two seconds, change the CTA — it doesn't register. Same visual bones, same ad in Meta's eyes.

Then Meta released a Creative Similarity Score. A score above 60% triggers retrieval suppression. CPMs go up. The penalty compounds based on the degree of redundancy in your creative library.

To say it plainly: the platform that created the pressure to conform is now charging you for conforming.

The exit requires more creative, faster, in more formats (which means more spend, more production, more structural dependence on the platform). And that's not a broken business. It's as intentional as Apple changing their charging ports and cords with every new iPhone generation.

Takeaway: The opacity means their system is working. You pay for the camouflage and you pay for the antidote.

(Sources: Clutch.co; Social Media Examiner; AdMove, "How Meta Andromeda Works"; PPC Blog Pro)

3. The Cost

Razorfish, drawing on Publicis Media's synthesis of over 100 trend sources, found that consumers spend 13+ hours daily with technology and media, yet 75% of digital advertising isn't seen long enough to create memory.

The usual media explanation is fatigue. The real explanation is categorization. Your brain doesn't ignore an ad because it's seen too many ads. It ignores it because it thinks it's seen this one before (or something close enough to be filed under “already processed, nothing new here.”)

Novelty triggers encoding. Familiarity triggers dismissal. The camouflage works so well the brand becomes invisible to the brain before it even becomes visible to the eye.

Platform-side data tells the same story from the other direction. Frequency increases, but reach stalls and costs climb. The media team may diagnose it as market saturation. The actual issue is creative similarity: Meta won't send your ad somewhere new because it looks just like everything else in your library.

A brand that optimizes its way into sameness has voluntarily left the brand category and joined the commodity category. Price becomes the only argument, which is a race with no finish line and no winners.

Takeaway: The media strategy designed to make you visible is the exact thing making you invisible.

(Sources: Razorfish / Publicis Media, "5 Consumer Trends Rewriting the Brand Playbook"; Evolut Agency)

4. The Contradictions

“Fruit Love Island” — an AI-generated serialized dating show that's exactly as the title would lead you to believe it is — generated 200 million views on TikTok. The same Gen Z audience that torches brands for AI-generated ads voted enthusiastically for their favorite anthropomorphized “fruit couples”.

Ad Age's Gillian Follett said the asymmetry isn't about the technology — it's about the implied choice. When a big brand uses AI, Gen Z reads: they could have paid a real artist and chose not to. But the same moral blame game doesn't happen with a creator making something totally unhinged for fun. The audience is judging both the thing and the reasoning that produced it.

Meanwhile Bilt, a finance platform, is funding a sitcom-style TikTok series about a woman navigating New York. No product in the plot. It works because it comes from a genuine commitment — to a world, a character, a creative sensibility — not from a platform playbook.

Rachel Karten calls this the effort heuristic: perceived effort signals quality. On a platform optimized for volume and velocity, evidence of effort reads as authenticity. Because those savvy audiences know it's genuinely rare.

The brands that will win this are the ones with a point of view that's clear enough to where the algorithm's pressure doesn't change what they make. You can't fingerprint a perspective. You can only have one or not.

Takeaway: The exit from camouflage isn't a smarter creative strategy. It's authorship.

(Sources: Ad Age Insider / Gillian Follett; Rachel Karten, "Evidence of Effort," Link in Bio)

This Week's Red Thread: Cultural Camouflage

About 25ish years ago, my brother was just starting his first year at Vassar College as I was starting my last year at Middlebury College.

Being an opinionated younger brother, he couldn't stop telling me how much more “truly alternative” the place was — as opposed to the preppy sameness that plagued my school. (He wasn't wrong about Middlebury, but that's not the point here).

I visited him on a beautiful fall weekend, and in the span of 48 hours, I heard The Strokes' song “Last Nite” at least 100 times. Every single dorm room, every common area, every outdoor speaker, car stereo, campus bar and house party were proudly playing it over and over as though they'd all stumbled on the original vinyl recording in the bathroom of CBGBs. This “alternative” Northeastern college campus had all the diversity in musical taste as North Korea.

The fact that their official anthem was The Strokes was ironic in itself. These Nepo-baby rebels grew up on the mean streets of the Upper East Side of Manhattan and all met at a combo of the Dwight School, and Institut Le Rosey, the most expensive of all the Swiss boarding schools. But their blood sweat and tears were spent cosplaying as dive-bar Velvet Underground doppelgängers.

Everyone at Vassar thought they were being alternative by listening to The Strokes. The Strokes thought they were being alternative by pretending to grow up on the Bowery. Nobody was being alternative. Everyone was camouflaged as something the system around them rewarded.

That was one type of cultural social pressure. Now it's algorithms that manipulate culture. And the camouflage is everywhere.

My brother wasn't wrong that Vassar had taste. He was wrong that taste and conformity are opposites. Taste is what happens after you've absorbed the influences and decided which ones are yours. Conformity is what happens when you skip that step.

Fashion at scale is camouflage. Camouflage at scale is monoculture.

And the exit from both has always been the same: know what you actually stand for, make things from that place, and be consistent enough that no one can mistake you for anyone else.

The algorithm didn't invent this problem. It just runs it faster and sends you the bill.

Happy Easter and happy Passover to all who celebrate...

— NB