Cold Email Infrastructure: The Capacity Math

The math almost always gets done. It just usually gets done in month two, when the campaign has been running for six weeks and the numbers are under plan. Someone finally sits down, works backwards, and finds that the campaign could never have hit the target. Not because of the copy. Not because of the list. Because the sending capacity never supported the goal in the first place.

Between a meeting target and the infrastructure that would carry it sit four multiplications. They take five minutes and need four numbers, three of which you already have. Run them before you approve the spend and you get either a plan or a scoping conversation. Run them afterwards and you get an explanation.

Why the math rarely gets done

The reason is mundane. The target is expressed in meetings, the infrastructure in mailboxes. Two units nobody converts into each other, because they live in different conversations — the target sits in a board meeting, the mailboxes sit on a vendor invoice.

The mistake also hides well. The first weeks are mailbox warm-up, then campaign ramp. Both produce exactly the kind of weak early numbers you patiently sit out. By the time somebody realises it was never a ramp problem, three months are gone — and against a three-week warm-up window, three months is not a rounding error.

The four numbers the calculation starts with

Three of them you normally have. The fourth is where it hangs.

1. The positive-reply target per week. If you plan in meetings, add a step: not every positive reply becomes a meeting.
2. The expected reply rate. This is the number that multiplies everything else, and the one most often estimated optimistically.
3. The channel split between LinkedIn and email, as the sequence design defines it.
4. The follow-up factor. A three-step sequence with one follow-up each doubles touchpoints against first-touch only.

For the reply rate, three bands are workable until you have your own data:

These are planning figures, not measurements. They stand in for your own data only until you have some, and they are deliberately not optimistic. Planning launch against the best band means planning the failure in.

The chain, worked backwards

Three formulas, in this order:

1. Weekly touchpoints = reply target ÷ reply rate
2. Senders or mailboxes = weekly touchpoints including follow-ups ÷ 100
3. Domains = round up (mailboxes ÷ 3)

Add a 20 percent buffer on mailboxes and domains — for spares when a sender leaves the company, and for reply bandwidth when a sequence runs better than planned. The buffer is not caution, it is experience: without it the campaign stalls on the first outage.

The worked example that ends in a target you cannot staff

A target of ten positive replies per week sounds modest. At average fit, an even channel split and a three-step sequence with one follow-up each, the math looks like this:

Forty LinkedIn sender profiles, ideally senior ones, is more than most mid-market companies can field. That is not an arithmetic error — that is the result. And it is exactly the number that belongs on the table before anyone signs.

Three levers remain when the math lands here. They are not equivalent:

Why a narrower audience rescues the math

The cut is the only lever that lowers the capacity requirement instead of paying for it. That is not theory — it is the difference between a campaign that pays for itself and one that needs more senders.

At node.energy the starting position was exactly the critical one: a very broad target market of 12,000 companies with no industry-specific restriction. Instead of scaling capacity up, the approach ran on signal-based outbound rather than mass outreach, explicitly so as not to burn the market. The published result: 5–6x more demos generated, increasing from an average of 5–6 per month to 30–35 per month, in under three months. Not through more mailboxes, but through a better hit rate per touchpoint.

b2match shows the same mechanism: the biggest deal in nine years, from outbound, in less than four months.

The three weeks nobody makes up

New domains and mailboxes carry a three-week warm-up window. During it, not a single campaign email goes out. This is not a precaution you trim when you are in a hurry — a mailbox that sends volume too early lands in spam, and repairing that is harder than starting over.

The decisive part: the bottleneck is the calendar, not the work. The setup itself — buying domains, setting authentication, provisioning mailboxes, configuring signatures — takes hours. The warm-up takes weeks, and no amount of extra effort or budget shortens it.

Two planning rules follow. Both are free, and both get broken regularly:

• Domain selection belongs on day one of the engagement, not two weeks after kickoff. It is an asynchronous decision — nobody needs a meeting for it.
• Sender names are locked once provisioned. Change your mind afterwards and the warm-up window restarts. So that decision needs an owner and a sign-off before anything is ordered.

The capacity everyone forgets: who handles the replies

Reply handling is one of the most common failure points in outbound, and almost never for the reason people assume. It is rarely the messaging. It is that nobody on the receiving side feels clearly responsible.

Two models work, and the choice follows the structure:

• One dedicated owner for all inbound replies across all campaigns. Right when there is one team and one region.
• A routing rule in the CRM that assigns automatically by segment, region or account owner. Right as soon as several teams or regions are involved.

What matters is not which one you pick, but that one model is chosen, written down and given a name. Every positive reply needs a named owner — without one, the most expensive stage of the funnel evaporates.

What works well in practice: a dedicated channel where every reply lands — positive, negative and neutral — and where the team logs what happened using reactions and short comments. It makes it immediately visible when a reply has been left sitting.

And one detail worth more than it looks: the neutral replies are the underrated part. Not a clear yes, not a hard no — that is where the largest unused potential sits, and where nothing gets done most often. An hour spent going through the neutral replies together after the first weeks finds more than any subject-line optimisation.

The pre-check in one sentence

There is one test that stands in for the whole calculation when time is short: if you cannot state the required sender count and the rough market size in a single sentence, do not sign and do not launch.

This is not a rhetorical test. Someone who can name the number has done the math. Someone who cannot has a target and a hope. The difference between the two shows up reliably in month two.

Once the math holds, the work moves to the list itself — the checks before the send are the step after this one. And if the open question is market size rather than capacity, that is a different calculation: see pipeline generation.