Capability Gap

All verticals

Design, video and content production

The highest inferred take rate in the atlas attached to the weakest defence: abundant supply, a salary-anchored price, and a buyer alternative that gets cheaper every quarter.

avoidlow confidence8 minupdated 2026-08-29design · creative · content · subscriptions · devtools
Who is buying
Series A/B devtool and SaaS startups spending $50k–$300k a year on creative and content combined
Who is selling
Globally distributed designers, editors and engineer-writers; abundant and reachable through ordinary channels
The spread
70–88% [WEAK — supply-side figures are self-reported]; 55–65% on generous assumptions
Size of the pool
No category figure. Superside, the largest named vendor, reported ~$44.9M revenue in 2024
Read
A 70–88% spread on Glassdoor-grade evidence, earned by operations rather than by scarcity — and the entire pitch is a comparison to a designer's salary that generative tooling is dragging down.
Budget depth
2
Supply difficulty
1
Spread
4
Holdability
2
AI direction
1
Speed to first dollar
5

This is the most commoditised vertical in the atlas, and it is also the one with the widest apparent spread. Both statements are true at once, and the tension between them is the whole page.

Superside's published pricing gives an effective in-plan rate of about $83/hour at its Growth tier. Multiple Glassdoor reviews report the network's experienced designers being paid $10–$12/hour, with reviewers describing the pool as entry-level or offshore (Vidico's pricing breakdown; Glassdoor review; Glassdoor hourly pay). That is a 70–88% spread [WEAK — the supply side of this ratio is self-reported review data, not accounts].

A spread that wide with supply that abundant is not a moat. It is a price that has not finished falling.

Whose budget

Moderate, discretionary, and anchored to a salary the buyer already pays.

A Series A/B devtool company spends roughly $50k–$300k a year on creative and content combined. Superside's average contract value is about $90k/year — the good case, not the median (Vidico).

The tiers: Starter $5k–$8k/month for 40–80 hours, Growth $10k–$18k/month for 80–160 hours, Enterprise $20k–$40k+/month for 200+ hours. Layered on top are a mandatory $1,000/month service fee, $1k–$5k onboarding, overages at $100–$200+/hour, 10–25% rush fees, and "Project Boosters" with a $10,000 minimum — add-ons that lift total spend 15–25% above the base subscription (Vidico).

The add-ons are the product

A subscription priced at $83/hour with overages at $100–$200/hour is not really a subscription; it is a retainer with a meter. The 15–25% add-on lift is where the turnaround guarantee gets paid for, and it is the closest thing this vertical has to pricing power — the buyer is not paying for design, which is abundant, but for design by Thursday, which is not.

The pitch is explicitly "cheaper than one designer." That makes it a displacement sale, permanently benchmarked against a headcount cost — the same trap as Outbound and GTM-as-a-service and the mirror image of Adversarial evals and red-team crowds, where no prior headcount existed. budget: 2: real money, small per account, discretionary, and first to be cut.

Can you get the supply

Trivially. This is the only supply: 1 in the atlas.

Designers, video editors and technical writers are abundant, globally distributed and reachable through ordinary channels — job boards, Contra, Dribbble-adjacent listings, an ad. Nothing about assembling them requires a relationship, a licence, a rig or a reputation. Draft.dev's network of 300+ technical experts is a genuine asset for devtool content, but it is an asset built by recruiting writers, which anyone with a budget can do (Draft.dev).

Superside's own shape confirms it: roughly 850 staff against ~$44.9M of 2024 revenue is about $53k of revenue per head, which is only possible with a globally distributed, low-cost delivery workforce (Latka).

What is genuinely hard is quality control at volume and hitting a turnaround SLA — brief intake, QA, escalation, rework. That is an operational moat, and operational moats are copyable by anyone willing to build the same process. The competitor a vendor must beat is not another agency but the buyer's own offshore hire: offshore headcount allocation at funded companies rose from 24% to 30% year-on-year at a stated 40–50% cost arbitrage (SaaStr on ICONIQ).

The same abundance is why this is the atlas's fastest vertical to a first invoice. speed: 5. Nothing sits between a standing start and money: the supply answers a job board this week, a $5k–$8k/month Starter plan is a marketing lead's discretionary signature with no procurement, and the deliverable ships inside the first month. Weeks, not quarters. Read that beside supply: 1 and it is an indictment rather than a feature — a business you can start in a fortnight is a business your customer can start in a fortnight, and so can everyone else.

What the spread looks like

Wide on the headline, narrower on any honest reading, and the evidence quality is the worst in the atlas.

At $83/hour effective and $10–$25/hour supply cost, the gross margin is 70–88%. Assume instead a blended $30/hour supply cost plus project-management overhead and it lands at 55–65% — still a good business, still not a defensible one.

Technical content is a similar shape with worse data. Agencies charge devtool startups roughly $500–$2,000+ per technical article; writer pay is not published, and market rates for engineer-writers in the low hundreds per piece imply a 50–70% take [UNVERIFIED — the supply side here is an assumption, not a source].

Every take-rate number on this page rests on Glassdoor

Superside discloses revenue and headcount but not cost of delivery. The $10–$12/hour figure comes from employee reviews — self-selected, unaudited, and possibly describing one region or one seniority band rather than the network average. The atlas quotes it because it is the only supply-side evidence that exists, and flags it because it should not carry the weight the arithmetic puts on it.

Note also the GMV is not revenue point: Superside's ~$44.9M (up from $30.8M in 2023) is agency revenue, recognised gross, because Superside is principal on the whole engagement. A marketplace routing the same work would book only the fee. Comparing the two headline numbers directly is the standard error in this sector. spread: 4 — the level is high, the confidence is not.

Can you hold it

No. Both sides can leave, and the buyer's exit is trivially cheap.

There is no lock-in mechanism in a design subscription. No data accumulates, no workflow gets embedded, no compliance artefact expires. A brand-guidelines file transfers in a morning. The switching cost is one month's notice plus the annoyance of re-onboarding — which is exactly what the $1k–$5k onboarding fee is trying to convert into friction.

Supply leaks in the ordinary way: a designer who does good work for a client meets the client. Contra exists explicitly to make that transaction commission-free, which is Getting cut out as a business model rather than as a risk.

The one durable asset is the buyer relationship at enterprise scale — a $20k–$40k/month account with a dedicated team and an approved brand process is genuinely annoying to move. That is a small fraction of the logo count. hold: 2.

What AI does to it

This is the ai: 1 of the atlas: models are eating the work directly, at the exact layer where the margin sits.

The margin lives in high-volume, spec-driven production — ad variants, social cuts, landing-page iterations, formulaic technical explainers. That is precisely the output generative tooling produces most convincingly and most cheaply, and the buyer does not need to fire the vendor to benefit; they simply need fewer hours.

The pattern is documented elsewhere in this sweep and it is brutal. Legal document review — the closest structural analogue, a human-supply arbitrage on a task that became model-tractable — went from a $460,000 quote to $36,000 for the same 100GB, 250,000-document matter (Decover). Undifferentiated egocentric video went from a few dollars an hour to free (DreamVU). Nothing protects creative production from the same repricing except taste, and taste does not scale to 850 people. See What better models do to each layer.

Superside's own trajectory is the tell: $400M valuation in 2021 on $35.1M raised, revenue at ~$44.9M in 2024, and no comparable raise in the category since (Latka).

What would kill it

What would kill it

The price of the alternative keeps falling and the buyer notices. The entire category is sold as arbitrage against a designer's salary. When one in-house designer with current tooling produces what a $10k/month plan produced last year, the plan does not get renegotiated — it gets cancelled. This is not a hypothetical decline; it is the mechanism by which the 2021 marks became unrepeatable.

Two faster deaths. The Glassdoor number becomes a story — a business paying $10–$12/hour and billing $83 is one well-read post away from a supply revolt and a buyer-side embarrassment, and the receipts are already public. The mid-market contracts — this vertical sells into the segment that is already in a mild recession behind the headline funding numbers: seed down 27% year-on-year, early-stage deal count at a five-quarter low (Crunchbase).

Who is already there

CompanyModelMoneyEvidence quality
SupersideSubscription creative, tiered hours~$44.9M revenue 2024 (from $30.8M 2023); $400M valuation 2021; $35.1M raised; ~850 staffRevenue via Latka estimate; delivery cost undisclosed
Design PickleUnlimited-design subscription, category originatorNot disclosed in the recordNone found
AwesomicYC; subscription talent matchingNot disclosed in the recordNone found
ContraCommission-free freelance marketplaceNot disclosed in the recordTake rate is zero by design
Draft.devTechnical content, 300+ expert writersNot disclosedNetwork size self-published
HackmambaDevtool technical contentNot disclosedNone found

Contra deserves a note: a commission-free marketplace in a vertical with a 70–88% agency spread is an explicit bet that the spread is unearned. Whether it can fund itself without one is the Marketplace, staffing firm, BPO or agency question in its sharpest form.

Where the record is thin

Almost everything below the revenue line

No company in this vertical discloses cost of delivery, contributor pay, churn or customer concentration. Superside's revenue figure is a third-party estimate; its $400M valuation is five years old and has not been re-marked in public. Design Pickle, Awesomic, Contra, Draft.dev and Hackmamba have no financial disclosure at all in the record — six named companies and one estimated income statement between them.

Two specific holes worth naming. Nobody has published what a technical-content writer is actually paid per piece, which means the 50–70% take on that sub-vertical is a guess resting on a guess. And there is no measurement anywhere of how much creative volume has already shifted to in-house generative tooling — the single number that would settle whether this vertical is shrinking or merely maturing.