Winners don't run better ads. They run an assembly line.
If you run paid creative — DTC founder, e-commerce performance lead, agency buyer, or in-house growth team — this one is for you. Across eighteen months of platform spend and creative output, one trait separates the top 1% of advertisers from everyone else, and it isn't budget, tenure, or some clever efficiency hack. It's raw creative throughput. The advertisers who win ship creative at industrial scale, run dozens of campaigns at once, and get there fast. Volume is the moat — and your refresh cadence is the lever.
The Velocity Ladder.
Two stories the numbers look like they're telling — and aren't.
It isn't seniority.
The obvious read is that the winners are simply the accounts that have been around longest. They haven't. Climbing from the bottom half to the top 1%, output explodes 280× — yet average account age inches from 470 to 550 days. Roughly two and a half extra months. The winners didn't out-wait anyone. They out-produced them.
It isn't efficiency.
The second guess is that the big players found a cheap, clever way to mass-produce ads. The opposite is true. Divide 90-day spend by creatives shipped: the bottom half spends $104 per video, the top 1% spends $2,188 — roughly 21× more per creative, not less. There's no scrappy hack here. They industrialized, and they paid full freight to do it.
One number for the whole story.
The Creative Velocity Index (CVI) scores each advertiser tier by total creative output — videos, images, and interactives — shipped per advertiser per quarter, indexed so the median advertiser = 100.
It collapses the entire operating gap into a single, comparable figure. The typical advertiser scores 100. The top 1% scores nearly 7,700 — not because they're better marketers, but because they're running a fundamentally different machine.
Read it as a throughput benchmark: where does an advertiser sit on the velocity curve, and how far is the climb to the next rung?
Now find your seat on the ladder.
If you run paid creative on this platform — agency, in-house DTC team, founder doing your own buying — you already know your monthly creative output. Plug it in. The tool maps you to the tier you'd land in by the CVI ladder above, and tells you what it would take to climb a rung. The point isn't to flatter or scold. It's to make the gap concrete.
Roughly, what does your team ship in a month?
How it works. Your inputs are converted to a 90-day total (monthly × 3) and divided by the median-advertiser's quarterly output (222.3 assets), then × 100. Tier cutoffs are the midpoints between each tier's CVI score from the table above. Benchmarks reflect the gaming · agency-managed cohort — the highest-n, most reliable segment of the source data. Nothing is saved or transmitted. Refresh to clear.
And the format question already has an answer.
Before celebrating any winner, kill the obvious red herring. Vertical video isn't winning. It already won. Across every advertiser tier, the share of video uploads that are vertical sits at ~100% — winners and losers alike. The differentiator isn't format. It's throughput within format.
(long-form)
(interactive)
(portrait)
(short-form)
(static)
(long-form)
The growth is leaving gaming for the checkout cart.
Volume tells you how advertisers win. The eighteen-month trend tells you where the winning is moving. Gaming is still the cash base — roughly $985M / month in managed spend — but consumer and e-commerce spend is compounding more than 1.5× faster, and a wave of new self-serve advertisers is arriving to claim it. If you're a DTC operator wondering whether this platform is for you yet: the data says you and 2,166 other consumer self-serve advertisers just signed up in the last eighteen months.
Consumer self-serve grew 6.4× — but spend per advertiser fell.
The same window that saw self-serve consumer counts run from 400 to 2,567 also saw the average spend per self-serve advertiser fall from $49.5K to $13.4K per month. A natural read is that the cohort is weaker. We disagree — and want to flag the tension before anyone else does.
Lower per-advertiser spend during a sustained 6.4× count expansion is what it looks like when the front door opens. If you're a DTC operator just kicking the tires — running a few hundred dollars a day, A/B-ing your first vertical hook — you're exactly who that $13.4K average represents. That's not the platform softening. That's the next thousand customers walking in.
So what does this mean for you?
Count your weekly creative output — then double it.
Strategy isn't the gap. Cadence is. If you ship 30 video assets a month, your competitors outranking you on impressions aren't smarter — they're shipping 5–20× that. Brief faster, batch wider, kill weaker hooks sooner. The lever is throughput.
The bottleneck the industry will pay to solve is volume, not novelty.
Generative creative tools, briefing platforms, freelancer-roster software, agency creative ops — all sit on the load-bearing constraint. The DTC market is expanding 6.4×; the production stack expanding underneath it is undersupplied. If you ship anything that lifts a brand from 90 to 900 assets per quarter, you're selling water in the desert.
The thesis writes itself: creative throughput infra × the e-commerce wave.
Two compounding curves in the same dataset point the same direction. Performance is throughput-constrained (the 280× gap). Demand is migrating to consumer (4.1× vs gaming's 2.5×, 6.4× more self-serve advertisers). The investable bet is whatever collapses the unit cost of high-volume creative at the moment the next thousand DTC advertisers arrive.
Read the investor memo ↗The bottleneck is creative throughput. The demand is moving to e-commerce. Whoever connects the two owns the next leg.
Two things in this data are hard to argue with. One: what separates winning advertisers isn't budget, tenure, or efficiency — it's creative throughput. Volume is the binding constraint on performance. Two: demand is migrating — consumer / e-commerce spend is compounding far faster than gaming, and self-serve advertisers are arriving more than 6× over. Put them together and the highest-leverage move a platform can make writes itself: collapse the cost of industrial-scale creative.
Do that, and you don't just widen the lead for the top 1% — you hand the surging bottom 99% the one lever proven to move the line, at the exact moment e-commerce demand arrives to use it. That's a flywheel: more advertisers → more creative throughput → more performance → more advertisers.
Two more findings sitting in the same data.
Issue 01 is one story, sharply told. But this dataset has more in it than the headline, and editorial honesty means showing the work. Here are two second-order findings we caught while building Issue 01 — both pulled from the same three CSVs, both defensible. Neither made the headline because Issue 01 needed to be one moat, not three. They're here so you can see what we left on the cutting-room floor — and because at least one of them is the cover of Issue 02.
The interactives gap is wider than the video gap.
The 280× video gap is the headline of Issue 01. It isn't the widest gap in the data. The widest gap is in interactives — playables and HTML5 mini-experiences. Top 1% gaming-managed advertisers ship 1,747 interactives per quarter. Bottom-50% advertisers ship 4.5.
That's 388× — about 40% wider than the video gap. The format the average advertiser quietly ignores is the format the elite double down on.
Peak vertical? Maybe.
Vertical video uploads compounded for 22 straight months. Then April 2026 hit 831,404 uploads — an all-time peak. May 2026 dropped to 751,183. A 9.6% month-over-month decline in the format that's defined the entire window.
One data point is not a trend. But after eighteen months of monotonic growth, the first down month deserves a note — not a celebration of "vertical video collapsed," but a flag for advertisers planning their second-half 2026 creative ops. If April was the peak, the calculus on hook fatigue and format diversification changes.
How this was built — and what was thrown out
- Cohort. Every Finding-01 number uses gaming, agency-managed accounts — 270 to 428 advertisers per tier, the segments with enough advertisers to be reliable. Thinner segments were not used for headline claims.
- Tiers are defined by 90-day spend rank. So a spend gap between tiers exists by construction. We never treat it as a finding. Every claim above rests on a variable free to move independently of spend: creative output, campaign count, or account age.
- Excluded. The top-1% consumer / managed cell (n = 5) and its 25,186-interactives-per-advertiser figure. Five advertisers isn't a pattern; it's an outlier wearing a costume. Likewise excluded: the top-25% gaming / self-serve cell (n = 2).
- Set aside as a red herring. Vertical-video share runs ~100% in every tier — winners and losers alike. It is table stakes, not a differentiator, so it earns no place in the story of why anyone wins.
- Mean vs. median. A handful of mega-spenders inflate the top-1% gaming mean ($33.5M) versus its median ($19.6M). Where the divergence matters, both numbers are read side-by-side.
- Finding 02 uses monthly managed-account spend (Dec 2024 – May 2026), indexed to Dec 2024 = 100, plus self-serve advertiser counts from the same period. The honest tension — 6.4× advertiser growth alongside falling spend-per-account ($49.5K → $13.4K) — is surfaced on-page rather than buried; we read it as expansion runway.
- Associative, not causal. Stated plainly in the close above. The arrow could run either way. The defensible claim is the operating pattern, not the mechanism.
- Data. Illustrative, anonymized, aggregated platform data covering Dec 2024 – May 2026 (format detail extends back to Jun 2024). Figures rounded for display. No external data used.
- What didn't make Issue 01, but is in the data. Two more findings sit in the same CSVs — the interactives gap (388× — wider than the video gap) and a possible peak in vertical-video uploads (April 2026, −9.6% in May). Both are surfaced in the Annex above rather than buried in this list, because editorial honesty means showing what we left out, not just what we picked.
- Build. Data parsing, recomputation, and SVG/CSS chart authoring done with Claude Code as a coding partner — charts are computed from the raw CSVs and hand-rendered, not AI-image-generated. Writing and editorial choices are mine.
