For seventy years, advertising treated polish as a proxy for quality. A brand that could afford a flawless television spot was a brand with money, and a brand with money was probably a brand you could trust. The polish itself carried information. That logic held because polish was expensive. It took crews, studios, retouchers, and time, so only serious players could produce it.
Generative tools broke the proxy. Flawless is now the default output of a text box. Perfect lighting, perfect grading, perfect grammar, perfect symmetry, available to anyone in seconds. And when a signal becomes free, it stops being a signal. The audience has already internalized this faster than most marketers have: a suspiciously perfect image now reads as cheap, not premium. The polish that once said "we invested" now says "we prompted."
This inverts the entire aesthetic economy of our industry, and the inversion follows a rule I want to name precisely, because I watch brands misapply it every month. Imperfection is not the new value. Cost is the value it points to. The scratch on the hand thrown mug matters because it proves a hand. A scratch added in post proves nothing, and audiences are getting frighteningly good at telling the difference.
Key Findings
- Signals work only when they are expensive to fake. Polish lost its cost around 2024, so it lost its meaning, and the market is repricing toward signals machines cannot cheaply imitate.
- The valuable imperfections are byproducts of real process: hesitation in a voice, a specific opinion that risks something, texture from a real place and day. Cosmetic roughness applied to synthetic work is already being detected and punished.
- Craft signals compound for people and small teams faster than for large brands, because provable individual authorship is the cheapest honest signal available.
- Big brands can still buy craft signals, but only by actually buying craft: real artisans, real time, real constraints, shown transparently.
- By 2030, most premium categories will run a two track creative economy: synthetic volume for coverage, certified craft for meaning, with almost nothing surviving in between.
The Signal Cost Law
Here is the principle underneath all of this, and I will state it as a law because I believe it is falsifiable: the persuasive value of any stylistic signal converges on the cost of faking it. Call it the Signal Cost Law. Peacock tails work because a sick peacock cannot grow one. Handmade joinery works because a factory line cannot hesitate. And a flawless brand film no longer works because a teenager can generate one before breakfast.
Run every creative decision through that filter and the confusion of the current moment resolves. Should your ad look rougher? Wrong question. The right question is: what does this creative choice prove that a competitor with a prompt cannot prove? A shaky camera proves nothing; anyone can shake a camera, and a model can simulate the shake. A founder answering hostile questions on camera, unedited, proves nerve. A campaign shot in one take in your actual factory proves the factory. Proof is the product. Roughness without proof is just bad production values wearing a costume.
Why Fake Imperfection Fails Faster Than Fake Polish
There is a tempting shortcut here, and half the industry is taking it: keep the synthetic pipeline, then art direct in some grain, some stutter, some "authentic" mess. I expect this to become the most punished creative strategy of the decade, for two reasons.
First, detection is asymmetric. Audiences forgive polish that turns out to be synthetic, because polish never claimed to be anything but surface. But imperfection is a claim about origin. When manufactured roughness gets exposed, and it does get exposed, the brand has not merely bored the audience, it has lied to them about the one thing imperfection exists to communicate. Second, the tell is coherence. Real process leaves correlated fingerprints: the lighting matches the location matches the accent matches the wear on the equipment. Simulated mess is statistically tidy underneath. People cannot articulate why it feels off, but they feel it, the same way everyone learned to smell a stock photo without ever taking a course in stock photography.
The Texture Budget
Practically, craft cannot be everywhere. No brand can hand make everything, and no brand should. What I recommend to clients is a deliberate allocation I call the Texture Budget: choose the small number of touchpoints where proof of human craft carries the persuasion load, fund them extravagantly, and let everything else be efficiently synthetic and honestly disclosed.
The logic is portfolio logic. One genuinely crafted flagship, a documentary, a product film with a named maker, a yearly report written recognizably by your founder, radiates trust across a thousand synthetic touchpoints that merely need to be competent. Spread the same money evenly and you get uniform mediocrity that signals nothing. Concentration is the strategy. This is also why the death of cheap distribution matters here: when reach is bought rather than given, as I argued in my essay on the death of organic reach, you want your paid impressions carrying your highest proof creative, not your average.
The Small Player Advantage
The strategic surprise of the craft economy is who it favors. For decades, production quality advantaged incumbents. Craft advantage flips that, because the cheapest honest signal in the market is a real person with a name, a face, a history, and something to lose. A solo consultant publishing sharp, risky, evidently personal analysis carries more proof per dollar than a Fortune 500 content studio, because her signal cannot be delegated and everyone knows it.
I have watched this play out with my own material. The pieces that move pipeline are never the smoothest ones. They are the ones with a specific opinion attached to my name that would embarrass me if I were wrong. That is costly signaling in its purest form, and it is why I tell operators that a personal brand is not a marketing channel but a proof asset, a point I develop in my guide to building a personal brand. Big brands can rent this through creators and founders, but they cannot manufacture it internally, and the renting shows.
What Craft Does to Category Structure
Zoom out to 2030 and I expect most premium categories to be visibly two track. Track one is synthetic volume: product feeds, variants, performance creative, personalization, produced at near zero marginal cost and disclosed as such. Track two is certified craft: scarce, named, slow, expensive, and doing all the meaning making. The middle, professionally polished content with no proof of anything, gets crushed, because it costs real money while signaling nothing the free tier does not.
This will restructure agency economics too. The billable hour survives on track two only. Agencies that positioned themselves as efficient producers of polish are competing with a marginal cost of zero, and that fight has one ending. Agencies that can organize real craft, find the makers, design the constraints, document the process credibly, are selling something that gets scarcer and pricier every quarter. I am rebuilding parts of my own firm around exactly this split.
What I Would Do About It
Audit your creative through the Signal Cost Law. For each flagship asset, write one sentence: this proves X, and faking X would cost a competitor Y. If you cannot fill in the sentence, the asset is polish, and polish is free now.
Set your Texture Budget explicitly. Pick two or three touchpoints per year where you will spend disproportionately on provable human craft, with named makers and documented process. Let everything else be efficient and honestly synthetic. The disclosure is not a confession, it is what makes the crafted tier believable.
Never simulate imperfection. If the mess is not real, cut it. The downside of exposed fake craft is categorically worse than the downside of admitted synthesis.
And if you are small: lean into the one advantage money cannot buy back. Publish under your name, take positions with real risk attached, show your process while it is still ugly. Every flaw that comes from an actual hand is, for the first time in the history of this industry, worth more than the airbrush.