Unit economics and survivors

#Retainer pricing, by niche and tier

Pricing varies enormously by scope and niche, and the sources here are almost entirely agency-marketing blogs rather than a single audited industry survey — treat the ranges below as [Directional], triangulated across several independent listings that broadly agree with each other, not a precise benchmark.

The Databox rows come from Databox's own published survey of agencies using its reporting software — a real, named, disclosed-operator-adjacent source (a software vendor surveying its own customer base), which is a meaningfully better tier than an anonymous blog post, though it's still a self-selected sample of agencies that already use one particular reporting tool, not a random sample of the whole industry. [Directional]

A pattern worth naming directly: SMMA-specific pricing sources cluster lower than general "marketing agency" pricing sources. The SMMA-specific figures (the $1,000–$2,000/month "sweet spot for new agencies" commonly cited in SMMA-course marketing content) are consistently below the general local-business retainer figures above. Some of that gap is real — a narrower, social-only scope should cost less than full-service SEO-plus-PPC-plus-content. But some of it plausibly reflects that SMMA-specific content is written by course sellers whose own students are new, under-differentiated operators charging entry-level rates, not a reflection of what the service is actually worth once specialised. Treat the low end of the SMMA-specific pricing content with that in mind.

#Worked example: what a solo operator's margin actually looks like

Take a solo operator running 6 clients at $2,000/month each — the middle of the entry-level SMMA range above. Gross revenue: 6 × $2,000 = $12,000/month. Subtract a realistic tool stack (see Launch, capital, and operations for the full breakdown; call it $400/month for a CRM/automation platform plus ad-account tools) and a part-time contractor for content production, say 40 hours/month at $25/hour = $1,000/month. Total costs: $400 + $1,000 = $1,400/month. Gross margin: $12,000 − $1,400 = $10,600, or 88% — before the operator's own time is counted as a cost. This is the arithmetic behind the commonly-cited "50–70% margin for solo operators" figure: it's realistic if the operator's own labour isn't priced in, because a true solo operator is the labour. The moment that operator hires a second person to free up their own time, the real margin compresses toward the lower end of that range or below it — which is exactly the point at which many single-operator SMMAs plateau. [Directional] on the underlying margin range — sourced from SMMA-industry content with the same course-seller caveat as above — but the arithmetic itself is load-bearing and checkable against any real client count and rate.

#What it costs the agency to acquire a client

This is the number SMMA-course marketing rarely states honestly, because it's the number that determines whether the business model in the previous section actually works.

• Cold email, done properly in 2026 — dedicated sending domains, inbox warming, a genuine SDR or the operator's own time, and the tooling to manage deliverability — realistically runs close to $5,000/month in tooling and labour once done at a volume that produces results, and multiple sources flag that agencies under roughly $30,000/month in revenue struggle to run this channel profitably at all. [Directional] — sourced from cold-email-tooling vendors, who have some incentive to make the "old way" (raw cold email) look expensive relative to whatever they're selling instead; treat the specific dollar figure with that caution, but the directional point — cold email deliverability infrastructure got materially more expensive and technical through 2025–2026 as spam filtering tightened — is corroborated independently by the compliance material in Compliance and the guru economy.
• Cold-email conversion, once delivered: a "good" email-to-meeting conversion rate is above roughly 0.4% by one named 2025 benchmark report; average reply rate across a large B2B email dataset runs around 3.4%. [Directional]
• Cold calling: industry-average conversion in 2025–2026 runs 2–3%, with a large study across 200,000+ calls landing at 2.3%; a commonly-cited 2026 B2B-agency benchmark is 25–40 dials per booked meeting, compressing to 15–20 dials per meeting with verified direct-dial data and a parallel dialer. [Directional]

#Worked example: what a "6–10 week to first client" timeline actually requires

Take the 25–40-dials-per-meeting figure above and a realistic solo-operator pace of 40 dials/day (roughly what one person can sustain alongside actual client delivery work, not a full-time SDR's volume). At the low end (25 dials/meeting), that's a booked meeting every ~0.6 working days; at the high end (40 dials/meeting), one meeting per working day. A realistic close rate on a booked discovery call for an unproven solo operator — not a named benchmark, a conservative planning assumption — might run 1-in-4 to 1-in-6. Running the high-end case (40 dials/meeting, 1-in-6 close rate) forward: one meeting per day, one close roughly every 6 working days, so a first signed client is plausible inside 2–3 weeks of consistent daily dialling once outreach actually starts — and realistically, a new operator needs 1–2 weeks before that to build a niche-specific pitch, a portfolio or case study to show, and a target list. Grounded estimate: 3–5 weeks from a genuine standing start to a first signed client, assuming daily outreach discipline and no non-standard friction. This is meaningfully faster than the multi-month timelines the licensing-heavy models researched elsewhere on this platform require, because SMMA has no professional-licensing wall at all — see Compliance and the guru economy for the one legal floor that does apply. Treat any course promising a first client in days, not weeks, as marketing rather than a plan to budget time against.

#Churn: the real number, and why "good churn" numbers disagree with each other

• Retainer-based agencies: roughly 18% annual client churn, with an average client lifespan cited around 56 months — those two figures are at least internally consistent (1 ÷ 0.18 ≈ 5.5 years ≈ 66 months, in the right neighbourhood of the 56-month figure once accounting for churn compounding within a year rather than at a single point).
• Project-based work: churns closer to 42% annually, average lifespan around 24 months — a very different, and much weaker, retention profile than retainer work.
• By agency size: $1M–$5M-revenue agencies (11–25 employees) run roughly 24% annual churn; $5M–$10M-revenue agencies roughly 19%; large 51+-employee full-service agencies achieve the best retention, around 12–15%, generally attributed to dedicated account teams and broader service integration that makes a client harder to fully replace elsewhere.
• Roughly 43% of B2B agency churn happens in the first 90 days — the onboarding window, before a new client has seen a full reporting cycle or real results.

Where these sources disagree with each other, stated plainly: different blogs in this space cite "good" churn as anywhere from 3–5% to 8–10% annually, without a shared methodology or named study behind either figure — a sign that these specific numbers are being repeated between sources rather than independently measured, even though the broader pattern (retainer-based retention around 75–85% annually, worse for project work, better for larger full-service shops) is consistent enough across independent sources to treat as [Directional]. Treat the specific "good churn" percentage as folklore-tier and the broader pattern as reasonably solid.

#What actually differentiates a survivor

Pulling the threads from this lesson and the root-mechanism lesson together, three factors show up consistently across independent sources as separating agencies that last from ones that don't — none of them is a secret, and none of them is what most SMMA marketing content leads with:

1. Niche specialisation over generalist "social media management." Directly follows from the execution/judgment split in the previous lesson — a specialist sells judgment a platform's automation can't replicate; a generalist competes with a free platform feature.
2. Retainer structure over project-based work, because the churn data above shows retainer relationships are structurally stickier — a client on a monthly retainer has to actively decide to leave, where a project-based client simply doesn't rebook.
3. A real onboarding process in the first 90 days, because that's where the 43%-of-churn concentration above says the relationship is actually won or lost — a client who sees a full reporting cycle and a clear expectation-setting process before month three is measurably less likely to be part of that 43%.

#The failure-rate claim this course will not repeat

A "92% of agencies fail" or "90% of marketing agencies fail" figure circulates in SMMA and agency-coaching content. It has no traceable source: one widely-shared LinkedIn post pushing the 90%+ figure explicitly concedes "this 90% is just opinionated." Other cited figures in this space disagree with each other and with that one — "75% of small agencies fail within five years" in one source, "34% of advertising agencies close within five years" in another — with no shared methodology visible behind any of them. [Speculative — discard]

Separately, a specific claim about a class-action-style "SMMA lawsuit" over course sellers' unfulfilled income promises turned up in low-quality aggregator search content with no named plaintiff, filing, or court docket this research could locate. [Speculative — could not verify, likely unreliable content] — named here so it isn't repeated as fact, not because this course found evidence it's false, but because it could not find evidence it's true either, and a specific-sounding unverifiable claim is worse than an honestly vague one.

The grounded comparison point is general US business survival data, not an agency-specific number: new business establishments survive their first year at roughly 76.8%, falling to about 51.2% by year five, per the Bureau of Labor Statistics' Business Employment Dynamics programme. [Established] — a real, named, primary federal data series, though it covers all new employer establishments across every industry, not agencies specifically, so it's a floor for comparison rather than a direct agency-failure figure. An SMMA, being low-capital and often solo-operated, plausibly has a different survival profile than the all-industry average in either direction — this research found no agency-specific survival study rigorous enough to state a number, and says so rather than inventing one.