Sales Capacity Planning: The Headcount a Target Implies, and the Gap It Hides

A board sets next year's revenue number in October. By November it has been divided by a quota and turned into a hiring plan, and by the following September the team is short and nobody can say which assumption broke. The plan was arithmetic all along, and the arithmetic was mostly assumptions wearing the clothes of measurements.

Sales capacity planning is the model that produces that hiring number, and the useful version of it is less about the formula than about which of its inputs anybody has actually measured.

What the model computes

Salesforce's sales capacity planning guide defines it as the process of identifying and forecasting the number and type of sales representatives needed to achieve a given revenue goal, and notes that it is also called sales force planning or sales capacity forecasting. The basic formula it publishes is short enough to hold in your head: the number of reps, multiplied by individual quota multiplied by average quota attainment, gives sales capacity.

That page is honest about what the short version leaves out. It says the basic formula does not account for ramp times, churn, or new hire plans, and recommends mapping the reps you start each quarter with, plus planned hires, minus average churn. For the ramp adjustment it recommends counting fully ramped reps at full weight and new hires at half, and its published worked example runs a team entering a quarter with twenty fully ramped reps, expecting two hires and typically losing two reps a quarter, to a capacity of nineteen complete reps. The inputs in that example are Salesforce's own illustration rather than a measurement of any real team, and the shape is what matters: the number of bodies and the number of selling units are different quantities, and the plan is built on the second.

The same page lists what a real model has to take in, and the list is the reason capacity planning belongs to whoever holds the data rather than to whoever holds the target: team size and roles, inbound lead volume, the pipeline required to hit the goal, average sales cycle length, average deal size, average quota attainment across teams and individuals, average ramp times, the number of reps ramping at any moment, and average turnover.

Two directions, and the gap between them is the plan

A capacity model is worth building twice, in opposite directions, because each direction answers a question the other cannot.

Run it top down and the target tells you the headcount it implies: this number, divided by a quota that a person at this attainment level can carry, requires this many selling units. Run it bottom up and the current team plus the hiring plan tells you what can actually be produced: these people, at these ramp states, at their measured attainment, produce this much.

Neither number is the plan. The difference between them is, and the value of doing both is that the difference is a quantity somebody has to decide about rather than a surprise that arrives in month nine.

1. Step 1Top down, from the target

   Divide the number by an achievable quota to get the selling units the target implies. This is an argument about the target, not about the team.

2. Step 2Bottom up, from the roster

   Count fully ramped sellers at full weight, ramping sellers at part weight, subtract expected attrition, and multiply by measured attainment.

3. Step 3Read the gap

   The difference is the part of the number that nothing currently produces. It is the whole output of the exercise.

4. Step 4Decide what closes it

   Hiring, quota per head, the segment mix, or a motion that does not consume a seller's attention. Each has a different lead time.

5. Step 5Name what stays uncovered

   Where the gap is not closed, say so before the period starts rather than discovering it in the third quarter.

Three adjustments decide whether the model is honest

Ramp is a discount on new hires, not a delay before they count. A seller hired in month two of a quarter is not a seller for that quarter, and the convention of counting them at part weight exists because counting them at full weight is how a plan promises capacity that has not arrived. The corresponding number is ramp time, and it is worth measuring for your own segments rather than borrowed, because a longer cycle and a more technical product both push it out.

Attrition is not a risk, it is an input. A model that treats turnover as an unfortunate exception plans a full roster and staffs a smaller one every period. Subtracting an expected loss looks pessimistic on the page and is simply what the last several quarters did.

Attainment is the input that decides everything and the one most often set to a hundred percent. A capacity model built on the assumption that every seller hits quota is a model of a team that does not exist, and it will overstate capacity by exactly the amount by which the assumption is wrong.

That third one is where the model quietly lies, and it is worth being specific about the mechanism with arithmetic. Suppose a plan carries ten fully ramped sellers on a quota of six hundred thousand each, and the model assumes full attainment. It reports six million of capacity. If the team's measured attainment is closer to seven tenths, the same roster produces four million two hundred thousand, and the plan is short by one million eight hundred thousand before a single thing has gone wrong. Every figure in that paragraph is invented for the illustration and describes no real company. The point that survives the invented numbers is that the attainment assumption moves the output more than the headcount does, and it is the input least likely to have been measured.

Modelled at full attainmentThe assumption most plans start from

• Ten sellers, quota of six hundred thousand each
• Reported capacity of six million
• Requires every seller to reach quota
• Reads as a plan and behaves as a hope

Modelled at measured attainmentThe same team, one honest input

• Same ten sellers, same quota
• Capacity of four million two hundred thousand at seven tenths attainment
• The shortfall is visible in October rather than in July
• Produces a smaller number and a decision that can be made

Our quota attainment entry carries the position that matters here: a broad team miss is usually evidence about the plan rather than about the sellers. A capacity model is where that evidence should have been used, because attainment is an input to it and the plan that produced the miss was built on the optimistic version of the same number.

Capacity and coverage are two different constraints

A capacity model says how many selling units the number requires. It says nothing about whether the market can be reached by those units, and treating one as the other is the most common way a plan is wrong while every number in it is correct.

The territory carve is where the second constraint lives, and the live position there is worth restating: a plan can show every account owned and zero accounts abandoned while most of the market sees nothing all year, because assignment gets recorded as coverage and the two are different facts. A coverage model assigns customers to segments, segments to roles and roles to people. A capacity model assumes those people can produce. Both can be internally consistent and jointly wrong.

The practical version is that a capacity plan should be read against a workload number as well as a revenue number. Take an invented pair of figures purely to make the shape visible: a model that says fourteen sellers, against a carve that gives each of them six hundred accounts. Neither number describes a real company. The plan has answered how many people the target needs and left unanswered whether the target is reachable by fourteen people working the way this team works, and only the first of those two questions has an owner.

The gap has two answers and only one of them is hiring

Every capacity model on the shelf resolves its gap the same way, by adding heads, because that is the variable the model exposes. It is also the slowest and most expensive lever, and it compounds: a hire is a fixed cost that arrives immediately, produces on a ramp, and carries the same attrition probability as everyone else.

Our note on why hiring more reps is usually the expensive answer argues the case in full, and the fully loaded cost of a sales development seat is the worksheet that prices the option properly before it is chosen. The other answers to the same gap are less visible in the model and often faster: raising conversion inside the funnel the team already has, changing the segment mix so the same effort meets larger contracts, or moving part of the market to a motion that does not consume a seller's personal attention.

Pressure-testing a capacity plan

• ✓Yes: The attainment figure is measured from your own last four periods, not borrowed
• ✓Yes: Ramping sellers are discounted, and the ramp length comes from your own segments
• ✓Yes: Expected attrition is subtracted as an input rather than noted as a risk
• ✓Yes: The model was run top down and bottom up, and the gap is stated as a number
• ✓Yes: The workload per seller is stated beside the revenue per seller
• ✓Yes: Every answer to the gap was priced, not only the hiring one
• ~Depends: Quota per head has been checked against what the current team actually reaches
• ✕No: A plan whose only variable is headcount, presented as arithmetic

Where this sits against outbound

A capacity plan describes selling capacity, and selling capacity is downstream of conversation supply. A model that produces a headcount has assumed each of those sellers will have enough qualified conversations to work, and that assumption sits outside the model entirely. Where it fails, the visible symptom is an attainment miss, which then feeds back into the next capacity plan as evidence about the team.

That is the loop worth breaking, and it is a supply question rather than a headcount question. Counting pipeline at the top against counting it at acceptance is where the supply number gets defined honestly, and pipeline coverage is the ratio that turns a revenue target into a pipeline requirement before any of it is divided by heads. Where the crediting rule is what determines which of several roles gets recognised for that pipeline, the crediting policy is the neighbouring decision, and it is worth settling before two roles are pointed at the same accounts.

RevenueFlow is paid on attended meetings that meet criteria agreed in writing before launch, which prices the supply side in the same unit a capacity model is built from. See what a first campaign produces before a gap is closed with headcount that takes two quarters to produce anything.

The short version

Capacity planning turns a revenue target into a number of selling units. The formula is reps multiplied by quota multiplied by attainment, and the adjustments that make it honest are ramp, attrition and a measured attainment figure rather than an assumed one.

Build the model in both directions and treat the gap as the output. Read the result against a workload number as well as a revenue number, because a plan can be right about how many people the target needs and wrong about whether those people can reach the market. Then price every answer to the gap rather than the only one the model makes visible, since hiring is the slowest lever in a plan whose problem is usually the supply of conversations rather than the supply of sellers.

Definitions, the formula and the ramp convention above are quoted from the Salesforce page linked in the text, fetched and verified 21 August 2026. Its worked example uses its own illustrative inputs. Verify current guidance against the source before relying on it.