Most cold email business cases are built backwards: someone picks a revenue target, divides by deal size, and derives a sending volume from it. That produces a plan, not a projection, and the plan is usually impossible.
Built forwards, the model has five inputs and it is unforgiving in a useful way.
The five inputs, with honest ranges
| Input | Realistic range | Note |
|---|---|---|
| Emails sent per month | Your mailbox count × 25–30 × 21 days | Capacity, not ambition. Size it here |
| Reply rate | 1–5%, occasionally 8% on a tight niche | The input people inflate most |
| Positive share of replies | 25–35% | The rest are no, not now, and wrong person |
| Positive reply to meeting | 30–50% | Depends almost entirely on response speed |
| Meeting to customer | 10–25% | The second input people inflate |
Worked through at the low end: 4,000 emails at a 3% reply rate is 120 replies, 36 of them positive, 14 meetings, and two customers. At the high end of every input the same volume produces closer to ten. That spread is the honest answer, and it is why a single-number projection is misleading.
The two inputs people get wrong
Reply rate. Public case studies quote 8 to 15% because those are the ones worth publishing. A well-targeted campaign into a defined niche with a real reason to write lands at 3 to 5%. Model at 2%, plan at 3%, and treat anything above 5% as evidence you have found something rather than as a baseline. Benchmarks by segment.
Close rate from cold meetings. Teams use their inbound close rate, which is typically two to three times too high. A cold meeting is with someone who was not looking. Use 10 to 15% until you have your own number.
The costs to include
- Infrastructure: mailboxes at $0.80 to $4 each, domains at roughly $6 to $14 a year. Small, and the part people focus on.
- Data: a credit per verified email, ten per mobile. At 4,000 emails a month across a three-step sequence that is roughly 1,300 new leads.
- Platform.
- Human time: the largest cost by a wide margin, and the one left out of every calculator. 120 replies a month is a meaningful share of someone’s week, and unanswered interest is the most expensive waste in outbound.
Cost per meeting, not cost per email, is the number to manage. At the figures above (14 meetings against maybe $200 of tooling and a third of a person), the tooling is rounding and the time is the business case.
What the model is for
Not forecasting. It is for two decisions.
Whether the arithmetic can work at all. If your deal size is $500 and the model says two customers a month, the channel cannot support the effort and no amount of copy fixes that.
Where the constraint is. Run the model, then ask which single input would change the outcome most. Under a 2% reply rate the answer is always targeting, never volume, and buying more mailboxes to fix a low reply rate doubles a bad result at double the cost.
Next: benchmarks to model against, what a lead actually costs, or infrastructure cost at 10, 100 and 500 mailboxes.
Frequently asked questions
What is a realistic cold email reply rate?
1 to 5%, occasionally 8% on a tightly defined niche with a real reason to write. Model at 2%, plan at 3%, and treat anything above 5% as a discovery rather than a baseline.
What close rate should I use for cold meetings?
10 to 25%, and 10 to 15% until you have your own number. Using your inbound close rate is the most common modelling error, it is typically two to three times too high.
How many emails can I actually send?
Mailbox count times 25 to 30 sends a day times about 21 working days. Capacity is set by mailboxes, not ambition.
What costs do ROI calculators leave out?
Human time, which is the largest cost. A hundred and twenty replies a month is a meaningful share of someone’s week, and unanswered interest is the most expensive waste in outbound.
What should I do if the projection is poor?
Ask which single input would change the outcome most. Under a 2% reply rate it is always targeting rather than volume, and adding mailboxes doubles a bad result at double the cost.