Revenue Per Employee: Why Headcount Stopped Being the Growth Lever

Key takeaways

‍

• Revenue per employee is total revenue for a period divided by the average number of employees over that period, and it only compares fairly within one sector.

• Hiring makes a poor growth lever for services firms, because payroll grows immediately while revenue grows only if the extra hours reach customers.

• The number moves when more of the week goes to revenue-generating work, through cutting non-revenue work, standardizing what performs, automating what’s stable and pricing on outcomes.

• Outcome-based contracts need evidence of what was delivered, so the revenue-generating share has to be measurable before you sign one with a client.

‍

Hiring is a linear answer to a growth problem that isn’t linear. Add ten people and payroll rises by ten salaries the day the offers are signed. Revenue rises only if the extra hours reach a customer, and some of them won’t.

‍

Adding people is the only growth lever that raises cost with certainty and output with luck.

‍

Revenue per employee shows the gap. It measures what an organization’s time produces, and it moves when more of each week goes to work a customer pays for, whether or not the team grows.

‍

What is revenue per employee?

‍

Revenue per employee is a company’s total revenue for a period divided by the average number of people it employed over the same period. It shows how much revenue the business generates for the size of its workforce, and companies usually report it for a fiscal year.

‍

Revenue per employee = Total revenue ÷ Average number of employees

‍

Use the average headcount across the period, or full-time equivalents where part-time work is significant, so that hiring or attrition partway through the year doesn’t distort the result.

‍

The figure varies enormously between sectors, because it mostly reflects how much human time a business model needs to earn a dollar. A software company with largely automated delivery will post a far higher number than a services firm built on people’s time, and the gap says more about the two models than about how either company is run.

‍

A useful revenue per employee benchmark comes from your own sector, ideally your sub-sector, from a source that names its method and the year it measured. A figure with no source on a conference slide isn’t a benchmark, however often it circulates.

‍

Comparing a BPO with a software company tells you little about either. For listed peers in your own sector, both inputs appear in annual reports, so a like-for-like comparison is straightforward to build.

‍

Headcount is a legacy metric

‍

Headcount was a reasonable stand-in for capacity when capacity meant hours. If each task took a fixed amount of time and a person did one thing at once, more people meant more capacity in a fairly direct way.

‍

Even the official measure of labor productivity counts hours over heads. The Bureau of Labor Statistics defines it as real output per hour worked. For a services business today, though, the useful variable is narrower again: the share of those hours that reaches a customer.

‍

Two teams of the same size can earn very different revenue because more of one team’s week goes to work a customer pays for. The other team loses its week to rework, switching between tools and coordination that never reaches a client. On an org chart, both are fifty people, and nothing on the chart shows the difference.

‍

Client pricing pressure makes that blind spot expensive. A client asking for a lower rate wants the same work from less of its budget, which only works if more of your week goes to billable work. Hiring to absorb the pressure adds cost to a margin that’s already shrinking.

‍

The AI question works the same way. When a client asks what AI saves them, they want to know whether the hours they pay for are going further, and a larger team answers in the wrong direction.

‍

The number can also mislead in both directions. A growing team can hide a falling revenue-generating share behind rising totals, and a team that shrinks through attrition can look more efficient while the work runs exactly as before. Reading revenue per employee alongside the share of time that reaches customers tells those cases apart.

‍

Utilization benchmarks by role and industry show how widely that share varies even within one role, before you compare whole organizations.

‍

Cost per employee, and what it leaves out

‍

Cost per employee is the easier number, and most finance teams know it well. It’s the fully loaded cost of employing someone: wages and salaries plus benefits, payroll taxes, equipment and a share of overhead.

‍

Benefits alone are a large part of that load. In the Bureau of Labor Statistics’ June 2026 employer cost data, benefits made up 30% of private industry employers’ compensation costs, which averaged $46.89 per hour worked. Equipment, software and overhead come on top of that.

‍

A cost figure on its own can’t show what the spending returned. Two teams with the same fully loaded cost per employee can earn very different revenue, so the cheaper team may still be the less valuable one.

‍

The denominator (average number of employees) is where cost per employee gets slippery. Contractors, part-time staff and outsourced seats may or may not be counted, and including them in cost while leaving them out of headcount overstates what each person costs. Choose one definition, use it for both cost and revenue per employee, and keep it stable from year to year.

‍

The figure can also move for reasons unrelated to performance, such as a benefits renewal or a shift toward more senior roles, so read every change in it with its cause attached.

‍

Profit per employee joins the two sides. It’s profit, usually operating or net, divided by the same average headcount used for revenue per employee, so it shows whether the revenue a workforce generates covers everything it takes to run the business. Read over several years alongside revenue per employee, it shows whether growth comes from a healthier business or simply a larger one.

‍

You need all three. Cost per employee tells you what a team takes to exist, and revenue and profit per employee tell you what that spending buys.

‍

What actually moves the number

‍

Four moves raise revenue per employee without a single new hire. Each one increases the share of the week that reaches customers, which is where the gap between similar teams comes from.

‍

• Cut the non-revenue share of the week. Rework, switching between tools, coordination and copying data between systems all use hours without producing anything a client pays for. Cutting that share raises the revenue-generating portion of every week. It’s also the move most firms skip, either as an oversight or because it's seen as too difficult to measure. Look first at work done twice, where a task is redone after review or re-entered in a second system.

• Standardize on the path that performs. When several ways of doing the same work coexist, some reach the customer faster. Making the best path the default lifts the average without asking anyone to work harder, and the metrics that matter when evaluating outsourced work usually show which path is winning. Write it down, train the team on it, and check after a quarter whether the gap between paths has closed.

• Automate what is stable. A stable, well-defined, repetitive step is a good automation candidate, because automating it frees the time it used to take for work that still needs judgment. The frame is force multiplication: the same team producing more revenue-generating work. A step qualifies when it runs the same way most of the time and its inputs arrive in a predictable form. Automating an unstable step repeats its inconsistency faster, so measure the step before and after to show where the freed time went.

• Price on outcomes rather than hours. Hourly pricing rewards work that takes longer, whatever anyone intends. Outcome-based pricing rewards the revenue-generating share directly, once you can measure that share and defend it to a client. Start with one client and one service line where the baseline is already stable, and widen it once the model has held through a full cycle.

‍

Most firms have room on more than one of these at once, and the order matters. Cutting non-revenue work and standardizing come first, because they create the stable process that automation and outcome pricing both rely on. All four leave the team the same size and change what its week produces.

‍

Measuring it credibly

‍

Credible measurement stays at the team and process level. The question is how much of a team’s collective time reaches the customer, and the answer belongs to the team as a whole.

‍

The number rests on time sorted into categories: revenue-generating work, internal coordination, rework and everything else that takes hours without reaching a client. A defensible baseline covers a full business cycle, since one week that happens to look clean or chaotic misrepresents the norm.

‍

Agree on the categories with the team leads who run the work, so the split matches how the work is organized. Then keep the definitions fixed, because changing what counts partway through a year breaks the comparison. Revenue sits on the other side of the ratio, so make sure the time data and the revenue figures cover the same dates.

‍

Three traps distort the result:

‍

• Seasonality. A quiet month and a peak month show very different splits, and comparing them produces a trend that isn’t real.

• Mix shift. When the work itself changes, with complex cases replacing simple ones, the number moves even though nothing about how the team operates has changed.

• Counting hours as output. The most damaging of the three. Busy hours and revenue-generating hours are often different things, and treating them as one is how a struggling process looks fine on paper.

‍

Used this way, the measure points at the process, where the room to improve sits, and never at a person. The baseline should be one a team lead can defend in a meeting, with every figure traceable to its source.

‍

Report the result to the board as a share of team time next to revenue for the same period, so both sides of the ratio are visible together. When a team’s share falls, ask what changed in the work before drawing any conclusion about the team.

‍

What outcome-based contracts require

‍

Outcome-based contracts require evidence of what was delivered before they’re safe to sign, a higher bar than an hours-billed contract has to clear.

‍

An hours-billed contract needs little more than a timesheet. An outcome-priced one needs a defensible link between the work performed and the result the client pays for, so the revenue-generating share of the week has to be measurable before negotiation starts. Finding out partway through the first quarter, when the numbers stop matching expectations, is already too late.

‍

You need two things in place first: a stable baseline for how the work runs today, and a clear definition of the outcome being priced. The measure should also hold steady from month to month unless performance changes.

‍

Without that, you’re agreeing to a figure nobody can explain when the client challenges it, and that challenge tends to arrive when the margin is already thin. Agree on what happens when volume swings outside the expected range as well, since outcome pricing can penalize a team for a spike it didn’t cause.

‍

Build in a formal review after the first quarter, so both sides can check the measure against what was delivered before the terms roll forward. Clients who ask for outcome-based terms are asking for proof, and a firm without it negotiates from a weaker position than its work deserves.

‍

‍What remote BPO teams get wrong about SLAs is often the same mistake in another form: committing to a standard before confirming the work can reliably support it.

‍

Where the data comes from

‍

Each of these moves depends on a measurement that most companies don’t yet have anywhere finance or operations can see it.

‍

Insightful is a work data platform. Its Workforce Analytics product measures where time goes at team and process level, including how much of it lands on revenue-generating work. That turns output per person into a number you can manage rather than one you calculate after the fact.

‍

You decide what counts as revenue-generating, by labeling the applications and sites that make up that work and by setting up billable projects, so the split reflects your business. Time is reported by team across applications, websites and projects, which lets you read the split against revenue for the same period. If your definition changes, relabeled categories can be applied to past data, so the trend stays comparable.

‍

The scope is narrow on purpose. It doesn’t set prices, negotiate contracts or decide what to automate. It supplies the baseline those decisions need, measured at individual, team, and process level so the output describes how the work actually happens in full context.

‍

Make the number a managed one

‍

The growth lever is the share of each week that reaches the customer, and that share can be measured. Once you can see it, revenue per employee stops being a figure you report once a year and becomes one you manage week to week.

‍

That holds whether the pressure comes from a client’s rate card or from questions about what AI should be doing for your margin. Start with one team over a full cycle, and compare the share of its week that reaches customers with the revenue it earns.


The first result is usually a clearer view of where the non-revenue share sits, so you can make informed decisions that impact your margins right away.

‍

See where your own team’s time is going with a free trial.

‍

Frequently asked questions

‍

What is revenue per employee?

Revenue per employee is total revenue for a period divided by the average number of employees over the same period, usually a fiscal year. A higher figure means more revenue for the size of the workforce. Compare it only within a sector, because business models need very different amounts of human time per dollar.

‍

What is a good revenue per employee figure?

There’s no single good figure, because revenue per employee depends heavily on sector and business model. Benchmarks travel badly between industries, since each needs a different amount of human time to earn a dollar. Trust a benchmark only if it comes from your own sector and names its source and year.

‍

How do you increase revenue per employee?

Increase the share of the week that goes to revenue-generating work instead of adding people. That means cutting rework and coordination overhead, standardizing on the process path that performs best, automating stable repetitive steps to free time for higher-value work, and pricing on outcomes instead of hours.

‍

What is cost per employee?

Cost per employee is the fully loaded cost of employing one person: wages and salaries plus benefits, payroll taxes, equipment and a share of overhead. It’s well understood and easy to calculate, but it measures only what a team costs to exist and says nothing about the revenue that spending returns.

‍

Why is headcount a poor growth metric?

Headcount measures capacity as if it were still just hours, when what drives revenue is how much of each hour reaches the customer. Adding people raises payroll immediately, but revenue rises only if the extra hours go to work clients pay for, so headcount alone can’t explain growth.

‍

‍