Outbound and GTM-as-a-service

A software company built a directory of the agencies that resell it, and the directory became the most legible supply-side map in this atlas. Clay lists roughly 178 solution partners across four public tiers — Elite Studio, Studio, Advanced Artisan, Artisan — with named firms at the top including Growth Engine X, RevPartners, Frontal AI and Go Nimbly (Clay; GTM Engineering; Reachly).

The platform underneath is real money: $100M Series C at $3.1B in August 2025, a $55M DST-led tender at $5B in January 2026, and $150M ARR as of May 2026, up from $108M at the end of 2025 (Sacra; BusinessWire; TechCrunch).

The surprising part is what the directory does not say. Tier names are published; the two numbers that decide whether an agency layer is a channel or a franchise — what a partner earns on a referral, and what margin a partner keeps reselling Clay credits to its own clients — are not published anywhere. The public artefact is a leaderboard. The economics are private.

#Whose budget

The startup buyer, at its most ordinary. Every seed-to-Series-B B2B company doing $100k+/month in revenue is a candidate (ColdIQ) — which means a large count of small, mortal accounts rather than a short list of enormous ones. See How much money is actually in the buyer pool: roughly 8,000–11,000 companies globally raised $10M+ over eighteen months, at a benchmark ACV around $19K, with a 14-month median runway behind them.

Published prices cluster tightly. Clay-agency retainers run $3.5k–$10k/month mid-market, $1k–$3k at the budget end and $15k–$50k+/month enterprise, on 3–6 month minimums (Reachly). The largest practitioner survey — State of GTM Engineering 2026, n=228, including 67 agency operators and 30 freelancers — puts the median retainer at $5k–$8k/month across a full range of $1k–$33k, split by scope: Clay builds and enrichment $2k–$5k, managed outbound $5k–$10k, full-stack GTM $10k–$20k, advisory $3k–$8k/month or $200–$400/hour (GTME Pulse).

The older appointment-setting layer prices the same job more crudely. Belkins publishes "from $5,000/month" against 100 guaranteed appointments a year and 1,500 leads a month; SalesRoads charges $9,950 per four weeks for one SDR and $16,750 for two (RevenueFlow).

#Can you get the supply

Yes, and that is the problem.

Clay expertise was genuinely scarce in 2024–25 — scarce enough that Clay had to manufacture it, through certifications, bootcamps and the partner directory itself. By 2026 the manufacturing worked. 178 listed partners plus a documented backlash over a Clay pricing pivot (Revnu) is what the end of a scarcity looks like from the outside.

Underneath the GTM-engineering framing, cold-calling supply is not scarce at all, and never was. It is the most portable labour in this atlas: a laptop, a dialler and a list.

The one real constraint is that solo operators are simultaneously the supply and the middleman. Freelancers start at $2k–$4k/month per client, reach $4k–$6k within six months and $6k–$10k after a year (GTME Pulse) — at which point they are charging what a small agency charges and keeping all of it. Arbitrage only appears when an agency bills $8k–$15k and staffs the work with junior or offshore delivery. supply: 2 — assembling it is a hiring problem, not a sourcing one, and anyone can run the same ads tomorrow.

speed: 5, the highest in the atlas alongside Design, video and content production and Paid creators, clipping and UGC ad ops, and for the same reason the supply score is low. A solo operator bills $2k–$4k/month per client from the first engagement (GTME Pulse), the tooling is a Clay seat and a dialler, and the buyer is a founder with an empty pipeline signing a three-month retainer out of a discretionary line — no procurement, no pilot, no licence. Weeks from a standing start to a paid invoice. That is the whole warning in one number: the entry cost that lets you in tomorrow is the entry cost that let the other 178 listed partners in yesterday.

#What the spread looks like

Wide, on inference rather than disclosure.

Multi-operator agencies charge 40–60% above solo rates for identical scope, averaging $8k–$15k/month (GTME Pulse). Against roughly 0.3–0.5 FTE of an operator costing $3k–$6k/month fully loaded, that implies a gross margin of about 55–75% [UNVERIFIED — inferred from the two published rate cards, not from any operator's accounts].

The best public datapoint is ColdIQ: $7M ARR on about 20 people, roughly 70 active clients, $5k–$10k/month packages, onboarding 30+ new clients a quarter (ColdIQ). That $7M is billings, not net revenue — an agency is principal on the whole retainer, so unlike a marketplace its headline and its take are the same figure, and the margin question is entirely on the cost side. This is the distinction GMV is not revenue exists to police, and it is why agency ARR looks flattering next to a marketplace's.

spread: 4. Real, but 55–75% of a $6k retainer is a smaller absolute prize than 30% of a lab contract. See What a rake can actually be.

#Can you hold it

Barely. This is the weakest column on the card.

Three-to-six-month minimums exist because churn without them would be brutal. The client can hire the operator — the operator's own rate card shows what they earn direct, and $6k–$10k/month freelance against a $12k agency retainer is an argument the client will eventually make. That is Getting cut out with the numbers published on both sides.

Worse, the deliverable is auditable. Meetings booked either happened or did not, so there is no outcome ambiguity to hide a bad quarter behind, and Belkins' "100 guaranteed appointments" makes the guarantee explicit. And the customer base carries the venture mortality rate: on Carta's Class of 2018 cohort, 62% of seed-funded startups shut down within seven years and only 15% reach Series B (SaaStr on Carta). hold: 2.

#What AI does to it

Ambivalent, and Clay is the proof of both readings.

Clay grew ~500% at peak by automating the enrichment and research a junior SDR used to do by hand (Sacra). That automation created the GTM-engineering role and the agencies that staff it — a rare case of a model-adjacent tool manufacturing a services category rather than deleting one.

But the same automation is a treadmill. Each capability Clay ships converts a billable agency task into a checkbox, and the agency has to climb to the next unautomated rung to keep its rate. Meanwhile the deliverable itself — cold email at volume — degrades as everyone gets the same tooling; reply rates are a shared resource being consumed. ai: 2, and see What better models do to each layer.

#What would kill it

Two other endings. The salary anchor drops — a startup that can run signal-based outbound with one operator and a model does not need a $12k retainer, and offshore headcount is already at 30% of allocation at a stated 40–50% cost arbitrage (SaaStr on ICONIQ). The middle contracts — seed funding in North America was down 27% year-on-year in Q2 2026 with early-stage deal count at a five-quarter low (Crunchbase), and this vertical sells almost entirely into that middle.

#Who is already there

The structural read: this is an agency vertical wearing marketplace vocabulary, and the choice between the two shapes is the whole question — see Marketplace, staffing firm, BPO or agency. Nobody here has demonstrated the network effect that would justify a software multiple (What the public market pays for labour).

#Where the record is thin

Also missing: any operator's actual delivery cost (the 55–75% band is arithmetic, not accounts), any churn or retention figure for the category, and any measure of what share of a Clay agency's clients came through the directory rather than through the agency's own cold-start work. ColdIQ is the single disclosed P&L datapoint in a vertical with 178 listed firms.