ArticleIssue #164 | Creative Is Math Wearing a Costume
This week, I want to tie together a lot of what I’ve written about individually – hit rate, creative velocity, diversity, audience research – into one thread: creative is math wearing a costume.
The Artifact Was Never the Value
Most marketers, at some point, confuse the ad with the work. The ad is the deliverable. It’s what goes into the feed, what clients approve, what agencies bill against, what designers and editors spend their weeks producing. It’s tangible, visible, and entirely the wrong place to look.
It’s a costume.
The ad is a costume draped over the things that actually move performance: the offer, the underlying audience insight, the trust the brand has earned, the quality of the data, the sharpness of the positioning. Those are the body underneath the costume. The creative is just how it shows up in the feed. Change the costume and you change how it looks. Change the body and you change what it is.
Two brands can hire the same creative shop, pull the same template, produce ads that look nearly identical, and end up with radically different performance. The costume is the same. The body wearing it isn’t.
The Math Underneath
I’ve written about most of the individual pieces of this math over the past few years. I’ve just never put them in one place. So, the shortlist:
Creative performance follows a power law. Across millions of ads, a small minority of creatives (1-10%, depending on the account) account for 80-95% of spend and results. That’s the shape of creative performance everywhere. The exact ranges may vary, but it holds up in every vertical. Most of your creative budget funds ads that never scale. That’s not a failure of execution – it’s the shape of the distribution. If you’re uncomfortable with that reality, you’ll spend your career optimizing against the math rather than with it. The entire game is finding hits. Not making ads. Finding hits.
“Hits” has an actual definition. Most creative operations don’t use one, which is why most conversations about creative performance are unfalsifiable. A hit is not an ad that looks good in a review, or an ad the CEO happens to love, or an ad that posts a strong CTR in its first 48 hours. The definition we use is precise, and I’d argue it should be yours too: a hit is an ad that (1) produces at least 25 conversions in a 30-day window, (2) spends at least 25x your target cost per conversion (or AOV divided by tROAS, if you’re optimizing to return) & (3) produces those conversions at or below your target CPA. The specifics flex by vertical, but you need a definition this precise, because the third criterion is what separates a hit from an ad that’s simply consuming budget. Without a definition, hit rate is unmeasurable. Without hit rate, every conversation about creative volume is a waste of time.
Hit rate is the number that actually matters. Hit rate is the percentage of creatives you launch that meet the definition above. At a 20% hit rate, you need to produce 5 creatives for every 1 winner. At 40%, you need half as many. The agencies telling you to “make more creative” without knowing your hit rate are selling you inputs while ignoring the yield on those inputs. That’s like evaluating a sales team by call volume and ignoring close rate. Most accounts I audit run between 10% and 20%. Very few brands I’ve seen track this at all.
The reason hit rate dominates is that it’s operating leverage on the creative P&L. Cost per hit ( your real input price for buying scale) is defined as your cost per creative divided by your hit rate. If a creative costs $1,000 to produce and your hit rate is 10%, your cost per hit is $10,000. Raise the hit rate to 25% and your cost per hit drops 60% to $4,000.
That unlocks radically different economics and radically different competitive positions. At a 25% hit rate, a brand has two levers its competitor at 10% simply doesn’t have. It can hold hit volume flat and redeploy the 60% production savings into working media. Or it can hold production budget flat and generate 2.5x more hits at the same spend. Either path compounds. The advantage isn’t a one-time efficiency gain — it’s a permanent structural gap in how much capital each brand puts to work against the same goal, quarter after quarter.
On the flip side, when a CFO asks why the profitability of marketing investments isn’t there, “Our hit rate is too low” is almost always part of the answer. It’s also almost never the answer anyone gives, because it’s almost never the number being measured.
Volume is an output, not an input. This is where most marketers (and most agencies) get the math backward. Both tend to treat creative volume as a dial you turn: more budget, more creative; less budget, less creative. But the number of creatives you actually need is the output of four independent constraints:
- How fast your creatives fatigue, relative to the length of your audience’s consideration cycle.
- How many distinct personas and angles you need to cover before you’ve actually tested anything, vs. confirmed one execution of one message.
- Whether each creative has sufficient budget to generate the optimization signal required to exit learning.
- And whether the account as a whole has enough budget to absorb the total count without fragmenting signal across too many creatives.
Whichever of those four constraints is tightest is the one governing your required output. Hit rate then converts that constraint into required production volume. The brands that skip straight to “make more ads” are solving for the wrong variable – and usually accelerating the wrong problem.
There’s a structural wrinkle here worth mentioning: production-billing agencies have a built-in incentive to recommend more volume regardless of diagnosis. It’s how they grow the account. You should be skeptical of any creative prescription that happens to align with the service provider’s revenue model, because the advice and the revenue model are almost always coupled, and the coupling is rarely acknowledged. The honest version of the conversation is: “Before we recommend producing more ads, let’s first determine whether production is your binding constraint.” Most brands/agencies never have that conversation.
Why This Matters Right Now
For the past decade, producing ads at volume was hard. Editors, designers, studios, UGC pipelines, review processes – production costs were real. That friction was the moat that enabled agencies to charge for production.
That friction is gone. Production is trending toward zero marginal cost for anyone willing to use AI tools. The value of being able to produce a costume has collapsed. The value of being able to design the body wearing it has not.
The default response to cheap production is to produce more – and that response, applied at industrial scale by teams who don’t know their hit rate, is going to destroy accounts. Hit rate doesn’t hold steady as volume increases. It degrades, because the body under the costume doesn’t improve simply because costume changes take less time and costume production gets cheaper. Temu didn’t improve fashion.
In fact, the opposite happens: the degradation compounds. Lower hit rate means higher cost per hit. Higher cost per hit means more production budget required to generate the same number of hits. More production pressure means less time per ad concept, which means more AI slop, less thought per brief/concept and even lower hit rate. The brands that step onto this treadmill don’t realize they’re on it – they think they’re executing. They’re actually paying more each month for the same number of winners and convincing themselves that the problem is a production shortfall rather than a diagnostic one.
And the underlying ad platform environment doesn’t fix itself either. A few weeks ago, I audited an account spending $40K a month where CPMr had climbed to $119 vs. a target of $32. Meta had long since stopped finding new people. It was hammering the same audiences at rising frequency, running up the cost of reaching people who’d already decided whether or not to buy. Every new creative launched into that account was more meat into the grinder. The agency had told the brand they needed more creative. In the abstract, that advice is almost always correct. In that account, it was exactly wrong, because creative volume wasn’t the binding constraint. Fresh messaging angles were the only lever left to unlock audience response that Meta’s algorithm had stopped finding on its own.
This is the pattern about to play out at scale. AI-driven production is already flooding Meta + TikTok. Hit rates ( already low in most accounts) are falling further. The combination of more advertisers flooding the channels + more dollars migrating from traditional channels means CPMs are (still, somehow) climbing. Brand teams are subscribing to more AI tools, cranking out more assets and wondering why nothing is scaling. The honest answer is more difficult. It requires sitting down and fixing the diagnostic work nobody built the muscles for.
Meanwhile, the brands using the same tools to sharpen hypotheses, tighten offer iteration, audit audience research, and repair reach before incremental volume are pulling further ahead.
The signal-to-noise ratio in the auction is about to get considerably worse, which means the marginal value of being disciplined is about to get considerably higher. That’s a compounding structural advantage. Andromeda already rewards the inputs more than the finish. Cheap production just turns up the volume on that dynamic.
When the costume becomes free, the body wearing it matters more, not less.
The Takeaway
Creative has always been math wearing a costume. The costume is what everyone debates because the costume is what everyone can see. The math is invisible, which is why nobody argues about it- and exactly why all the leverage lives there.
If you want a single test for whether your creative operation is actually working, try this. Walk into your next creative review and ask two questions. What is our hit rate over the last 90 days? And what is the sharpest insight behind the ads we’re shipping this week? If the room can answer both crisply, you’re operating on the math. If the room only has an answer for what the ads look like, you’re operating on the costume. Most rooms, in my experience, can only answer the second.
The brands that win from here won’t be the ones with the best-looking ads. They’ll be the ones who figured out that creative was never really about the creative. It was always about the math. The only thing that’s changed is that the math is no longer hideable.
Cheers,
Sam
