What a quarter of outbound costs when you pay per attended meeting

A quarter in which Recursive Scale produces no attended meetings invoices nothing. No setup fee, no retainer, no minimum, and no charge for the sending infrastructure, the data or the ramp. Meetings that a qualified prospect attends are quoted per ICP in the $200 to $800 range, so a quarter producing fifteen of them costs between $3,000 and $12,000. A retainer covering the same quarter is owed in full either way.

A quarter with no meetings invoices nothing

Outbound is bought in quarters because that is how long it takes to know anything. The table below sets the same quarter against four published alternatives, at three levels of output.

Cost over one quarter	0 attended meetings	15 attended meetings	45 attended meetings
Recursive Scale at $200	$0	$3,000	$9,000
Recursive Scale at $800	$0	$12,000	$36,000
SalesRoads, 1 SDR at $9,950 per four weeks	$29,850	$29,850	$29,850
EBQ full-time at $10,000 per month	$30,000	$30,000	$30,000
Sopro from $3,800 per month	$11,400	$11,400	$11,400

Two numbers in that table move and three do not. The EBQ price is conditioned on an annual commitment, so a quarter that produces nothing is followed by three more that are already bought. Martal Group publishes no price at all and requires a three or four month pilot campaign before the monthly subscription begins, which is the same structure expressed as time instead of money.

Recursive Scale carries the opposite exposure. Unpaid ramp is a Recursive Scale cost, so stretching it has no commercial upside, and a quarter that produces nothing has been funded entirely from this side of the table.

Coverage is what the price buys

Recursive Scale works the full reachable market for a client’s ICP rather than a sample of it, and the per-meeting price is what funds that volume. Volume here is not a vanity figure. Sending capacity is hard-limited by the mail platforms: Google Workspace caps a paid user at 2,000 messages and 3,000 external recipients per rolling 24 hours, and Microsoft caps a mailbox at 10,000 recipients per day and 30 messages per minute while telling bulk senders to go elsewhere. Reaching a market of any size means running many inboxes on many domains, which is infrastructure rather than effort.

Recursive Scale runs that infrastructure on its own domains and inboxes. The client’s sending reputation is never the thing being spent, and the client keeps the data and the reply history.

The reason coverage matters commercially is narrower than most outbound marketing suggests. A campaign that reaches a fraction of the market produces a result that cannot be distinguished from noise, which means the client learns nothing from a quarter of work and pays for it anyway.

A small campaign cannot tell you whether the message worked

Detecting whether one message outperforms another is a sample size problem with a published answer, and the answer is larger than most campaigns. The calculation below uses the standard two-proportion power formula at 95% confidence and 80% power, with the baseline reply rate measured by Sales.co at 2.09% of unique contacts emailed across 161 campaigns and 61,770 classified replies.

Improvement to detect	Prospects per variant	Prospects for a two-arm test
10% relative lift	77,132	154,264
25% relative lift	13,201	26,402
50% relative lift	3,656	7,312
Doubling the reply rate	1,090	2,180

Anyone can reproduce those figures: n per arm equals [1.96 times the square root of 2p(1-p), plus 0.84 times the square root of p1(1-p1) plus p2(1-p2)], squared, divided by the squared difference in the two rates.

Set that against published campaign volume. SalesRoads publishes 500 prospect records per SDR per month, which is 1,500 across a quarter. A test needs 2,180 prospects to detect a doubling of the reply rate, so one SDR quarter cannot resolve even that, and a doubling is the largest effect anyone credibly claims. Unify GTM, the most careful treatment located, declines to give a universal threshold and then uses “400 comparable contacts per variant” in its own worked example, roughly an order of magnitude below what its example would need.

Volume is therefore not an optional extra on top of the message work. Full-market coverage is the only condition under which a message test returns an answer, and the attended-meeting price is the mechanism that pays for it without billing the client for the attempt.

The denominator decides what a benchmark means

Reply rates are quoted across this market on at least four different denominators, and the same campaign can be described four ways. Belkins reports 0.45% across 7,530,489 emails sent in 2025 and states the point directly: “A 5% reply rate against openers and a 0.45% reply rate against total sends can describe the same campaign; they’re just measuring different things.” Saleshandy reports 3.7% of emails delivered across 53.1 million emails. Instantly reports 3.43% of emails sent but publishes no sample size. Sales.co reports 2.09% per unique contact.

Booked and attended are the same trap one stage later. SmartReach publishes the only complete funnel found on a single consistent denominator, from 40 million emails across 5,000 campaigns: roughly 10% reply, 2.2% positive, 1.2% meeting booked, 0.6% opportunity per prospect, which works out at about 83 prospects per booked meeting. Gong separately puts it at 344 cold emails per meeting across 28 million emails, and at a four to five step sequence the two measurements land in the same place.

Recursive Scale invoices on attendance, which is the one denominator a client can verify without trusting anyone’s reporting. A prospect either joined the call or did not.

What is agreed in writing before anything sends

Qualification is written before launch as a named list of titles, plus a substantially equivalent role reasonably relevant to the client’s offering, so a near-miss on a job title is settled by the agreement rather than by an argument at invoice time. The client approves the messaging, the sender names and the target markets before the first email goes out. The term runs open-ended with 14 days’ notice on either side.

Published price is rarer than it should be in this market. SalesHive and Martal Group publish none. Leadium publishes no rate for itself but does publish the arithmetic, noting that a $4,500 per month programme booking eight qualified meetings costs $562 per meeting and that the same programme at twelve meetings costs $375. Both figures are for booked meetings, not attended ones.

Who should not buy this

Recursive Scale is the wrong purchase for several kinds of company, and the cheapest time to find that out is now.

Companies with fewer than roughly 10,000 reachable buyers do not have a market that supports full coverage. The sample-size table above is the reason: a small market cannot produce a readable test, and working it hard exhausts it instead.

Companies with an annual contract value below roughly $10,000 will struggle to carry any price in the $200 to $800 band, because the meeting-to-close rate needed to justify it is higher than most such businesses run.

Companies that want a predictable monthly invoice should buy a retainer. Attendance billing makes cost track output, which means a strong quarter costs more than a weak one, and some finance functions would rather have the fixed number.

Companies that want to own and warm their own sending domains are also a poor fit, since outreach runs on Recursive Scale infrastructure by design.

Sources: SalesRoads, EBQ, Sopro, Martal Group, SalesHive, Leadium, Sales.co, Belkins, Saleshandy, Instantly, SmartReach, Gong, Unify GTM, Google Workspace sending limits, Microsoft Exchange Online limits. All pages checked 11 September 2026.