How to Calculate an Employee's Actual Cost and Stitch It to Revenue: A Departmental Guide

Key Takeaways

• Fully-loaded cost is the only honest denominator. Salary alone understates the real cost of an employee by a quarter to a half, and comparing revenue to salary flatters everyone.
• Velocity, productivity and efficiency do not produce the cost — they explain it. They convert a lump cost into cost per unit of output and reveal where cost is being wasted.
• Separate revenue generated from revenue influenced. Generated is directly attributable and rare outside sales; influenced is contributed-to and must be credited fractionally, never in full.
• Never sum influenced revenue at full value across people. If everyone claims the whole deal they touched, attributed revenue exceeds real revenue several times over — the same double-counting that breaks marketing attribution.
• The method is departmental by necessity. Sales generates, marketing influences, engineering and product enable, and support retains — four different relationships to revenue that one formula cannot capture.
• For most roles the honest output is a cost-to-value relationship with a confidence level, not a precise ROI. Spurious precision on individual revenue attribution is the field's main failure and its main abuse.
• Use this to find and fix system waste and misallocation, not to rank and threaten individuals — attribution built for judgement gets gamed; attribution built for understanding gets used.

1. The Short Answer: Cost, Drivers, and Revenue

Connecting an employee to revenue is a three-part model. First, calculate their fully-loaded actual cost — not salary, but salary plus benefits, taxes, tools, space, management overhead and ramp time, typically 1.25 to 1.4 times base salary and higher for heavily-supported roles. Second, use velocity, productivity and efficiency to explain that cost — to express it as cost per unit of output and reveal where it is being wasted — rather than to fabricate an output number. Third, attribute outcomes as either revenue generated (directly closed by the person) or revenue influenced (contributed to, credited fractionally), and never confuse the two.

The reason this must be done department by department is that the four major functions relate to revenue in fundamentally different ways. A salesperson generates revenue directly. A marketer influences it at scale but rarely closes it. An engineer or product manager enables it by building the thing that is sold. A support or success person retains it. Applying a single attribution method across all four — or worse, expecting all four to show 'revenue generated' — produces numbers that are confidently wrong and that punish the functions whose contribution is real but indirect.

This guide builds the model and then works it per department. It is the financial capstone to the workforce-metrics cluster: it uses [employee velocity](/guides/employee-velocity-guide), [employee productivity](/guides/employee-productivity-guide) and [employee efficiency](/guides/employee-efficiency-guide) as inputs, and it is the individual-level companion to team costing such as the [sales team costing guide](/guides/calculate-sales-team-costing-guide).

• AEO Quick Answer: fully-loaded cost as the denominator, velocity/productivity/efficiency to explain it, revenue generated vs influenced to connect to outcomes — done per department.
• The four functions relate to revenue differently: generate, influence, enable, retain.
• One attribution method across all of them produces confident nonsense.

2. Why Salary Is the Wrong Number

Almost every casual calculation of 'what an employee costs' uses salary, and almost every one is therefore wrong by a wide margin. Salary is the visible cost; the actual cost is substantially higher, and using the wrong denominator flatters every downstream ratio.

The gap between salary and actual cost comes from a stack of real, recurring expenses. Employer taxes and statutory contributions. Benefits — health, retirement, insurance. Tools and software licences, which for some roles run to thousands per year. Workspace, whether office or a remote stipend. Equipment. Training and development. Recruitment and onboarding, amortised over the tenure. And management overhead — the fraction of a manager's cost devoted to this person.

Two costs are routinely forgotten and both are large. Ramp time: a new hire produces little or nothing for weeks or months while being paid fully, and that unproductive period is a real cost that should be amortised across their productive tenure. And the productivity cost of management and coordination — the meetings, the reviews, the context-switching — which grows with organisation size and is rarely attributed.

The practical multiplier is commonly 1.25 to 1.4 times base salary for a typical role, and meaningfully higher for roles with expensive tooling, heavy management support, or long ramp. The exact figure matters less than the principle: comparing revenue to salary rather than to fully-loaded cost overstates every employee's return by the size of the gap, which is precisely the amount the naive calculation ignores.

Compute fully-loaded cost first, for every role, because it is the denominator of everything that follows. A revenue-to-cost analysis built on salary is measuring against a number that is a quarter to a half too small.

• Salary is the visible cost; actual cost is 1.25-1.4x salary, higher for supported roles.
• The stack: taxes, benefits, tools, space, equipment, training, recruitment, management overhead.
• Routinely forgotten and large: ramp time and coordination/management cost.
• Comparing revenue to salary overstates return by exactly the size of the gap.

3. How Velocity, Productivity and Efficiency Fit In

The three workforce metrics are frequently misunderstood as ways to measure an employee's output directly. In this model they play a different and more honest role: they explain the cost, converting a lump sum into cost per unit of output and revealing where the cost is being wasted. They are the bridge between the cost and the value, not a substitute for measuring the value.

Velocity converts cost into cost-per-throughput. If you know a person's fully-loaded cost and their velocity — completed work per period — you can express cost per unit of finished work. This is meaningful in repeatable work and treacherous in knowledge work, exactly as covered in the velocity guide, so it is used where units are comparable and avoided where they are not.

Productivity is the cost-to-output ratio itself, at its most direct: output per unit of input, where the input is the fully-loaded cost. This is the cleanest of the three for economic purposes, because it is already framed as value per cost, and it is why economic productivity — output per dollar of loaded cost — is the workhorse metric of this whole model.

Efficiency explains where the cost is being wasted. A person's fully-loaded cost buys a certain amount of resource; efficiency reveals how much of that resource produced useful output versus rework, waiting and system-imposed waste. A high-cost employee at low efficiency is expensive not because they are overpaid but because a large fraction of what you pay for is being consumed by waste — usually system waste they do not control, as the efficiency guide details.

The crucial discipline: these metrics explain and adjust the cost side of the model; they do not manufacture the revenue side. It is tempting to multiply someone's velocity by a revenue-per-unit figure and call it their revenue contribution, but that fabricates attribution the data does not support for anything but the most directly-generating roles. Velocity, productivity and efficiency tell you what the cost buys and where it leaks; revenue attribution is a separate, more careful exercise handled in the next sections.

• The three metrics explain the cost, not the revenue — they are the bridge, not a substitute.
• Velocity: cost per unit of throughput (repeatable work only).
• Productivity: output per loaded cost — the workhorse economic ratio.
• Efficiency: where the loaded cost is wasted, usually to system waste the person does not control.
• Do not multiply velocity by revenue-per-unit to fake attribution — that is the classic error.

4. Revenue Generated Versus Revenue Influenced

The revenue side of the model rests on one distinction that, ignored, produces the single most common and most damaging error in workforce attribution.

Revenue generated is revenue a person directly and attributably closed. A salesperson who owns a deal and signs it generated that revenue. It is directly attributable, it is rare outside of closing roles, and it is the only kind of revenue that can honestly be assigned to one person in full.

Revenue influenced is revenue a person contributed to but did not solely close. The marketer whose campaign sourced the lead, the solution engineer who won the technical evaluation, the product manager who built the feature that clinched the deal, the support agent whose service earned the renewal — all influenced revenue, none generated it alone, and each must be credited fractionally, never in full.

The catastrophic error is crediting influenced revenue at full value across multiple people. If the marketer claims the whole deal, and the SDR claims the whole deal, and the AE claims the whole deal, and the solution engineer claims the whole deal, the sum of attributed revenue is several times the actual revenue. This is the exact double-counting that breaks marketing attribution, transplanted to people: everyone credited in full, the total a fiction. Any model that lets attributed revenue exceed real revenue has failed at arithmetic before it failed at fairness.

The honest treatment of influenced revenue is fractional credit that sums to one across all contributors — the same logic as multi-touch attribution in marketing, and subject to the same limits. Credit can be split by rule (equal, or weighted toward certain stages), by a model, or by judgement, but however it is split, the fractions for a given deal must sum to 100%, not to 400%. This is the [attribution](/glossary/attributions) problem applied to people, and our [attribution modelling](/solutions/attribution-modeling) work is about making it defensible.

State which kind of revenue you are attributing every time. 'This person is associated with $2M of revenue' is meaningless until you say whether they generated it (closed it) or influenced it (touched it, at what fractional credit). The two differ by an order of magnitude in what they claim.

• Generated: directly closed, attributable in full, rare outside closing roles.
• Influenced: contributed to, credited fractionally, never in full.
• The catastrophic error: full-value influenced credit across many people, so attributed revenue exceeds real revenue.
• Honest influenced credit is fractional and sums to 100% per deal, not 400%.

5. The Cost-to-Value Model

With cost and revenue defined honestly, the model that connects them is a ratio and a confidence level, not a single precise number.

The core ratio is value over fully-loaded cost. For a directly-generating role, value is revenue generated; for an influencing role, value is fractionally-credited influenced revenue; for enabling and retaining roles, value is a defensible proxy for their contribution, which the departmental sections address. The ratio expresses how many dollars of value each dollar of fully-loaded cost is associated with.

Attach a confidence level to every such ratio, because the confidence varies enormously by role. A closing salesperson's generated-revenue ratio is high-confidence — the attribution is clean. A brand marketer's influenced-revenue ratio is low-confidence — the attribution is genuinely fuzzy. Reporting both as if equally precise is the spurious-precision trap that discredits the whole exercise. A number with a stated confidence is useful; the same number presented as exact is misleading.

Read the ratio as a relationship, not a verdict. A ratio below one does not mean an employee is not worth their cost — it may mean they are in an enabling role whose value the ratio cannot fully capture, or early in ramp, or bearing system waste that suppresses their measurable output. The ratio is a prompt to understand, not a trigger to act, and treating a low individual ratio as grounds for a decision is exactly the abuse that makes people game the inputs.

Use trends and cohorts over point estimates. An individual's ratio in a single period carries heavy noise from deal timing, task mix and factors outside their control. The ratio's trend over time, or the aggregate ratio for a team or role, is far more reliable than any individual point figure — the same reason [revenue per employee](/glossary/ltv) is most honest at the company and team level.

The output of the model, for most roles, is therefore a cost-to-value relationship with a confidence band and a trend — not a precise individual ROI. Building it to produce a precise number where the data does not support one is how the model becomes both wrong and dangerous.

• Core ratio: value ÷ fully-loaded cost, with value defined by the role's relationship to revenue.
• Attach a confidence level — high for closers, low for brand marketers — never present both as exact.
• A low ratio is a prompt to understand, not a trigger to act.
• Prefer trends and team aggregates to noisy individual point estimates.

6. Department by Department: Sales

Sales is the one department where revenue generated is directly and cleanly attributable, which makes it the simplest case and the source of the mistaken assumption that all departments can be measured this way.

Cost: fully-loaded, and unusually high variable components. A salesperson's actual cost includes base plus commission and bonus (on-target earnings), the sales tech stack (CRM, dialler, sales engagement, data), and often significant management and enablement overhead. Ramp is long and expensive in sales — new reps frequently take months to reach full productivity — so amortised ramp is a substantial cost that is easy to omit.

Value: revenue generated, credited to the closing rep, but with a critical caveat about influenced revenue within sales itself. In a team-selling motion, an SDR sourced the meeting, a solution engineer won the technical evaluation, and a sales manager coached the deal — so even in sales, the closed revenue is partly influenced by others, and crediting the AE with 100% of a team-sold deal overstates the AE and erases the SDR and SE. Where roles are specialised, split the credit; where one rep does everything, they generated it.

The right ratio: generated (or fractionally-credited) revenue over fully-loaded cost, which is the cleanest cost-to-value ratio in the company. This is where the model is most trustworthy, and it connects directly to the metrics in the [presales, sales and post-sales KPI guide](/guides/kpis-presales-sales-postsales-guide) and the team-level economics in the [sales team costing guide](/guides/calculate-sales-team-costing-guide).

The trap specific to sales: over-crediting the closer and under-crediting the enablers, which distorts hiring and comp decisions toward closers and away from the SDRs, SEs and managers whose influenced contribution is real but hidden. Even in the department where generation is clean, the influenced/generated distinction matters.

• Cost: high variable comp, expensive tech stack, long expensive ramp — amortise it.
• Value: generated revenue, but split credit in team-selling motions.
• The cleanest cost-to-value ratio in the company.
• Trap: over-crediting the closer, erasing the SDR, SE and manager who influenced the deal.

7. Department by Department: Marketing

Marketing is the department most often mis-measured, because it influences revenue at scale but rarely generates it directly, and forcing a generated-revenue frame onto it either erases its contribution or fabricates one.

Cost: fully-loaded, including the large non-headcount spend marketing controls — media budget, tools, agencies, content production. A distinction matters here: the marketer's own cost is separate from the budget they deploy, and conflating the two produces nonsense ratios. Measure the person's cost against their contribution, and the campaign budget's return separately.

Value: influenced revenue, credited fractionally, and this is where the double-counting danger is highest. A single deal may be influenced by a brand campaign, a content piece, a paid ad and an email, each run by different marketers — and crediting each in full multiplies the deal several times over. The honest approach is fractional multi-touch credit summing to one, held at low confidence because marketing attribution is genuinely uncertain, especially for demand-creation work whose effect is lagged and cross-channel.

The specific danger: last-touch thinking applied to people. Just as last-click attribution over-credits capture channels, last-touch people-attribution over-credits whoever ran the final email and erases the brand and content work that created the demand — punishing exactly the demand-creation marketing whose value is real but indirect. This is the people-level version of the trap in the [performance, social and retention marketing KPI guide](/guides/kpis-performance-social-retention-marketing-guide).

The honest output for marketing: a low-confidence influenced-revenue contribution measured through incrementality where possible, plus leading indicators of demand created (pipeline sourced, branded search, assisted deals) rather than a false claim of generated revenue. A marketing function that reports 'we generated $10M' is almost always double-counting; one that reports 'we influenced this pipeline, sourced this much of it, at this confidence' is telling the truth.

• Cost: separate the marketer's own cost from the budget they deploy — do not conflate them.
• Value: fractional multi-touch influenced credit, held at low confidence.
• Danger: last-touch people-attribution erases demand-creation marketing, exactly like last-click.
• Honest output: influenced contribution via incrementality plus demand leading indicators, not fake generated revenue.

8. Department by Department: Product and Engineering

Product and engineering are enabling functions: they build the thing that is sold, so they influence essentially all revenue and generate none of it directly, which makes naive per-person revenue attribution both impossible and misleading.

Cost: fully-loaded, and often high — senior engineering salaries, expensive tooling and infrastructure, long ramp on complex systems. The cost side is straightforward; it is the value side that resists the standard model.

Value: the trap is trying to attribute revenue to an individual engineer, which cannot be done honestly — no single line of code generated a deal, and features are built by teams over time and monetised through a whole go-to-market motion the engineer did not run. Attempting per-engineer revenue attribution produces numbers so fabricated they discredit the exercise.

The honest approach is to measure at the level where the value is real: the team, the product area, or the feature. Team output over team cost, product-line revenue over product-line cost, or the revenue impact of a shipped capability over the cost of building it — these are defensible where per-person attribution is not. This is a direct application of measuring one level up, exactly as the productivity guide argues for knowledge work.

Where individual contribution must be understood, use it as productivity and efficiency (from the workforce-metrics cluster) plus qualitative peer and outcome assessment, explicitly not as an individual revenue number. An engineer's value is real and often enormous; the honest way to express it is through the team's and product's economics and through structured judgement, not through a fabricated personal revenue figure that will mislead every decision built on it.

The reframing: for enabling functions, the question 'how much revenue did this person generate' is the wrong question. The right questions are 'what is this team's or product's cost-to-value ratio' and 'what did this capability enable', which are answerable and useful where the individual version is neither.

• Enabling function: builds what is sold, influences all revenue, generates none directly.
• Cost: high and straightforward; the value side resists per-person attribution.
• Measure at the team, product-area or feature level — where the value is actually real.
• For individuals, use productivity, efficiency and structured judgement, never a fabricated revenue number.

9. Department by Department: Customer Success and Support

Customer success, account management and support are retaining functions: they influence revenue by keeping and expanding it, so their contribution is best measured through retention and expansion rather than through new-revenue generation.

Cost: fully-loaded, typically lower base than sales but with meaningful tooling and, for success roles, a book of business that frames their value. The cost side is standard.

Value: retained and expanded revenue, credited appropriately. A CSM who prevents churn preserved revenue that would otherwise have been lost — measurable through gross revenue retention on their book. An account manager who drives expansion generated new revenue from the existing base, which is closer to generated than influenced and can be credited more directly. Support, further from the deal, influences retention diffusely and is best measured through its effect on retention and satisfaction rather than through revenue attribution to individuals.

The honest metric set: gross and net revenue retention on the book (from the [post-sales KPIs](/guides/kpis-presales-sales-postsales-guide)), expansion revenue for account management, and for support, retention and time-to-resolution effects rather than a revenue figure. Retained revenue is genuine value — a dollar kept is a dollar not re-acquired at full CAC — and it is frequently undervalued precisely because it does not show up as 'new' revenue.

The trap: measuring support and success on activity (tickets closed, calls made) instead of on retained value, which rewards motion over the outcome the customer bought and the revenue actually preserved. A support function optimised for ticket closure closes tickets; one measured on its retention effect retains customers, and the two are not the same, as the post-sales KPI guide details.

The value of getting this right: retaining functions are systematically under-credited because their contribution is prevented loss rather than visible gain, and a cost-to-value model that captures retained and expanded revenue corrects a bias that otherwise starves the functions protecting most of the company's lifetime value.

• Retaining function: keeps and expands revenue; measure via retention and expansion, not new revenue.
• CSM value: gross retention on the book; account management: expansion (closer to generated).
• Support: retention and resolution effects, not individual revenue attribution.
• Trap: measuring on activity (tickets closed) rather than retained value.

10. Stitching Cost to Revenue Without Fabricating

The act of connecting an employee's cost to revenue is where good intentions most often produce dishonest numbers, so a few rules keep the stitching defensible.

Rule one: attributed revenue must never exceed real revenue. Sum all the influenced credit across all people for a given deal and it must total the deal, not a multiple of it. If your model's attributed revenue across the company exceeds the company's actual revenue, the model is double-counting and every ratio built on it is inflated. This is the arithmetic guardrail that catches the most common failure.

Rule two: match the attribution method to the role's real relationship to revenue. Generated for closers, fractional influenced for marketers and enablers, retained for success. Forcing one method across all roles either fabricates generation where there is none or erases influence that is real.

Rule three: carry the confidence level through to the conclusion. A high-confidence sales ratio and a low-confidence brand-marketing ratio should not be compared as if equally solid, and any decision that treats them as equivalent is built on a false equivalence.

Rule four: prefer team and cohort aggregates to individual point estimates wherever the individual attribution is weak. The team's cost-to-value ratio is almost always more defensible than any individual member's, and for enabling functions it is the only honest level.

Rule five: use the model to understand, not to judge. The moment individual revenue attribution drives pay or firing decisions, people optimise the inputs — claiming influence, gaming the credit split, avoiding un-credited but valuable work — and the model's numbers detach from reality. Attribution built for understanding stays honest because no one benefits from distorting it; attribution built for judgement is corrupted by the incentive it creates. This is the same Goodhart dynamic that runs through every metric in the workforce cluster.

• Attributed revenue must never exceed real revenue — the arithmetic guardrail.
• Match the method to the role: generated, influenced, retained.
• Carry confidence levels through to the conclusion; do not compare high- and low-confidence ratios as equal.
• Prefer team and cohort aggregates where individual attribution is weak.
• Use the model to understand, not to judge — judgement corrupts the inputs.

11. Building the Model: A Practical Sequence

Implementing this is a few weeks of finance and operations work, and it front-loads the cost and definition work that makes the revenue side honest.

Week one: build fully-loaded cost for every role. Assemble the cost stack — salary, taxes, benefits, tools, space, equipment, training, recruitment amortised, management overhead, ramp amortised — and compute a defensible loaded cost and multiplier per role. This is the denominator of everything, and it is pure accounting, so get it right first.

Week two: classify each department's relationship to revenue — generate, influence, enable, retain — and choose the honest attribution method for each. This is where you decide, per function, whether you are measuring generated revenue, fractional influenced revenue, team-level enabling value, or retained revenue.

Week three: instrument the revenue side. For sales, clean deal attribution with credit splits for team selling. For marketing, multi-touch influenced credit summing to one, ideally validated by incrementality. For product and engineering, team and product-area economics. For success, retention and expansion on the book. Confirm the arithmetic guardrail: attributed revenue reconciles to real revenue.

Week four: build the cost-to-value ratios with confidence levels and trends, at the honest level for each function — individual where attribution is clean, team where it is not. Present them as relationships to understand, with confidence bands, not as precise verdicts.

Ongoing: review the ratios as trends, revalidate the cost model as compensation and tooling change, and watch for the model being pulled toward judgement use, which is when it starts to be gamed. A cost-to-value model is a lens for finding misallocation and system waste, and it stays useful only as long as it is used that way.

• Week 1: fully-loaded cost per role — pure accounting, the denominator of everything.
• Week 2: classify each department's revenue relationship and choose its attribution method.
• Week 3: instrument revenue per method; confirm attributed revenue reconciles to real revenue.
• Week 4: build cost-to-value ratios with confidence and trends, at the honest level per function.

12. A Worked Example (Illustrative Model)

The figures below are an illustrative model to demonstrate the method across departments. They are not real employee data.

A software company examines one deal worth $120,000 in annual value and the four people associated with it, each at their fully-loaded cost.

The account executive (fully-loaded cost $180,000/yr, base $110,000) closed the deal: generated revenue, but it was team-sold. The SDR (loaded $95,000) sourced it, the solution engineer (loaded $160,000) won the technical evaluation, and a content marketer (loaded $110,000) produced the material that first attracted the buyer. Crediting all four in full would attribute $480,000 to a $120,000 deal — the double-counting failure.

The honest split, summing to the actual $120,000: AE credited 45% ($54,000 generated-with-influence), SE 25% ($30,000 influenced), SDR 15% ($18,000 influenced), marketing 15% ($18,000 influenced, low confidence). The fractions sum to one; attributed revenue equals real revenue.

The cost-to-value read, across many such deals rather than this one: the AE's generated ratio is high-confidence and clean. The SE's and SDR's influenced ratios are moderate-confidence. The marketer's is low-confidence and better validated through incrementality across the whole pipeline than through single-deal credit. Meanwhile the engineers who built the product this deal bought appear nowhere in the deal-level attribution — and correctly so, because their value is captured at the product-and-team level, where it is real, not at the deal level, where it would be fabricated.

The lesson the example teaches: the naive version — everyone claims the deal, engineers get a made-up per-head revenue number — produces attributed revenue several times real revenue and a set of individual figures that would misdirect every hiring and comp decision built on them. The disciplined version reconciles to reality, credits each function through its real relationship to revenue, and honestly says where individual attribution stops and team-level economics must take over. That boundary — between what can be attributed to a person and what can only be attributed to a team — is the whole difference between a useful model and a dangerous one.

13. Putting It Together

Connecting an employee to revenue honestly is three disciplined moves: compute fully-loaded actual cost rather than salary; use velocity, productivity and efficiency to explain what that cost buys and where it leaks; and attribute revenue as generated or influenced, fractionally and reconciled to reality, through each department's real relationship to revenue.

The departmental structure is not optional detail — it is the core of the method. Sales generates, marketing influences, product and engineering enable, and success retains, and a model that respects those four different relationships tells the truth where a single formula fabricates. The most common and most damaging error, across all of them, is crediting influenced revenue in full across many people, so that attributed revenue balloons past real revenue and every ratio inflates.

For most roles the honest output is a cost-to-value relationship with a confidence level and a trend, not a precise individual ROI. Insisting on a precise per-person revenue number where the data cannot support one produces spurious precision that misdirects decisions and, once tied to judgement, corrupts its own inputs.

Use this model to find misallocation and system waste and to understand where value is really created — not to rank and threaten individuals. If you want the cost-to-value model built against your actual departments and revenue data, with the attribution reconciled honestly, that is where our [business operations](/solutions/business-ops) and [unit economics](/solutions/unit-economics) engagements do their work — sitting on top of the workforce metrics in the [velocity](/guides/employee-velocity-guide), [productivity](/guides/employee-productivity-guide) and [efficiency](/guides/employee-efficiency-guide) guides.