Sales headcount is not sales capacity

[Updated on 29.9.2026] Rewritten. The model and its parameters are unchanged.

Most sales budgets we see are built the same way: take the number of sellers, multiply by a quota, and spread the result over twelve months. That figure is the capacity of a team where nobody leaves, nobody is new and everyone sells equally well. No such team exists, so the budget misses, and the explanation is usually looked for in the market or in the sellers rather than in the model.

This post describes the model we use to budget new customer acquisition in B2B SaaS: first the one condition without which no model matters, then the four places where headcount and capacity diverge. It is a generalization; every company adds parameters of its own. The parameters here come from the outbound teams we have built and from the actuals of our portfolio companies.

Who leads them

If the owner expects a new account executive to come in and start selling, no model matters; the plan has failed before the first call is made. A strong sales team is led by example, with discipline and a constant push, and that takes the full time of a sales-focused founder or, in a larger company, a sales manager. Without one there is no point in hiring account executives. They will fail, and the only results are lost money, lost time and a reputation among salespeople as a company where sellers do not succeed.

The opposite case is a company that gives new account executives a platform to join: sales training that consists of the materials, the most common arguments and counterarguments, lead lists ready to call, and a culture of doing. That is what puts a new hire in a position to succeed. Combine it with a founder selling next to the account executives, and the numbers below apply. Without it, there is no reason to look at the numbers yet.

Four leaks between headcount and capacity

Tenure. In practice the average tenure of a seller is around 24 months, so each month roughly one in 24 leaves: headhunted, changing industry, on parental leave. With three sellers this looks like an odd thing to model, but over a larger team and a longer horizon it is exactly what happens, so the model churns the salesforce by 1/24 every month.

Trial period. About a third of recruits leave during or right after the trial period, on their own initiative or the company's. Most budgets assume zero.

Ramp-up. A seller hired in January is not a full seller in February. Learning the company and the product, and building the confidence to close, takes months; six is a reasonable default for an account executive. A budget that counts every hire at full output from the first month is wrong for half the year.

Role weights. An account executive, a key account manager and a sales development representative do not carry the same capacity. Normalize one role to 1 and weight the others from your own numbers, for example AE = 1 and KAM = 1.7. SDRs are counted separately, because they produce meetings rather than signed revenue.

The sum of the weighted, churned and ramped heads is the salesforce; the same calculation for SDRs gives the SDRforce. The rest of the model runs on these two numbers. Headcount does not appear in it again.

Hire four to have two

Put the four leaks together and a practical rule follows. Four account executives start in January for a new market. A third of recruits leave during the trial period, so one of the four goes, sometimes two. The rest take six months to reach full output, and at a 24-month average tenure ordinary turnover takes about one more before the year is out. At the end of the first year the team is two heads, and over those twelve months the four hires produced the output of about two full sellers. By the end of the second year, without further hiring, one is left. If the plan needs two sellers at full output a year from now, hire four.

From capacity to net signed ARR

Capacity gives the number of calls and meetings the team can handle in a month. Conversion turns them into deals: separate hot and cold leads, and take the conversion rate for each from your own history rather than from a benchmark. Two deductions are usually forgotten before the result is called signed ARR. Signature and first invoice are rarely the same month, so revenue lags the deal. And a share of new customers leaves during their own trial period, so part of what was signed never becomes recurring revenue. What remains is net signed ARR.

Costs

Direct costs are always in the budget: salaries, salary-related expenses, commissions. Indirect costs are usually missing: recruitment fees, software seats, and the cost of letting people go. With a churning team these are recurring costs, since every seller who leaves is replaced through a new recruitment and a new ramp-up.

Comparing actuals with the model

Track five to fifteen parameters every month and put the actuals next to the model's estimates. Then leave the estimates alone. They are long-run averages, and the monthly actuals revert to them; a bad quarter is not a reason to lower the tenure assumption, and a good one is not a reason to raise conversion. Revise a parameter only when there is enough new data to show that the average has moved. Three rules make this work: base every assumption on your own historical numbers where they exist, build the model bottom-up from sellers and meetings rather than top-down from a revenue target, and apply the same prudence as in accounting.

The parameters we follow in every sales organization: number of meetings and deals; pipeline ARR and signed ARR; salesforce and sales headcount, side by side; ARR per salesforce unit; gross sales efficiency; potential market size; and how many times a year the team goes through the whole addressable market. On the last one, above two and a half rounds a year the same prospects are being contacted every few months, and the response rates show it.

When the first year in a new market misses the plan, we look first at who leads the team, then at these four leaks, and only then at the market. In most cases the plan was built on headcount, and either the leadership or the capacity was never there.