Utilization is an input, not a result
A firm with sixty billable people usually has a dashboard, and the dashboard usually leads with utilization. It is a capacity measure: how much of the time you paid for got pointed at a client. It says nothing about what that time earned, which is a chain of four more numbers, each multiplying the last and each leaking quietly. A firm can hold a good utilization number for two years while its effective rate falls fifteen percent, and nobody names the problem, because nothing on the page is pointed at it.
The mechanics below are the same for a design studio, an engineering consultancy, an accounting practice, an IT services shop or a specialist advisory practice, and the same whether the hour bills at $95 or $650.
You are probably here because
- Utilization looks fine and the bank balance does not
- Two people quoted two different utilization numbers for the same month
- Your biggest client is somehow your least profitable one
- Every fixed-fee project runs over and nobody can say by how much on average
All four share a root: the firm measures capacity and prices from cost, and never computes what an hour returns.
The number everyone quotes and nobody defines
Utilization is billable hours divided by available hours. Every word in that sentence is contested inside most firms, and the disagreement is rarely written down.
Billable hours. Hours coded to a client project, obviously. But does the two-hour scoping call before signature count? The rescue day you chose not to invoice? Most firms answer once, informally, in a hallway, and the answer drifts as people change.
Available hours. Three common denominators sit about ten points apart. The calendar gives 2,080. Subtract holidays and typical paid time off and you land near 1,850 to 1,900. Scheduled hours resolve part-time staff and mid-year starts correctly, at the cost of a number that moves weekly.
Pick the time-off-adjusted denominator, write it in one sentence, publish it, and do not change it mid-year. A firm that changes its denominator loses its own history, which is the only benchmark that ever meant anything.
Realization is where the money leaves without a decision
Realization is the fraction of work you did that turns into cash at your standard rate. Two independent leaks live here and most firms track one.
Rate realization is the discount agreed at contract time: rack rate $250, client negotiated $215, realization 86%. This leak is visible. A person decided it, in a meeting, on a date, and wrote it in a document.
Hours realization is work you did and did not bill. Someone worked 46 hours and the invoice went out at 40. This one is invisible. It gets decided by a project manager at eleven at night on the third of the month, under pressure to keep a client calm, and it appears in no contract anywhere.
Combined realization in the mid-eighties to low nineties is common and unalarming. Under 80% is a scoping problem wearing a billing problem's clothes. But the number is nearly useless without a reason attached, and attaching one is the highest-return reporting change most firms can make: every write-off gets a cause code. Five is enough — scope we gave away, estimate was wrong, rework on our error, client disputed it, deliberate relationship investment — and they map to five different departments. Without them, a nine percent write-off is weather.
The effective rate is the only rate that pays anyone
Effective rate is revenue divided by hours actually worked, unbilled hours included. It is the number that survives contact with reality, and almost nobody computes it.
Take a $250 rack rate, 90% rate realization, 93% hours realization: the effective rate is roughly $209. Multiply through the capacity side — 1,880 available hours at 70% utilization is 1,316 billable hours — and you get about $275,000 per delivery person per year. Against a fully loaded cost near $135,000 that is $140,000 of contribution before firm overhead, selling cost and unbilled management time.
The specific numbers are not the point. The chain multiplying is. A five-point utilization slip and a five-point realization slip are each small enough that nobody escalates them, and together they take about ten percent off the top line — usually the whole difference between a good year and a tense one.
Where the money usually leaks — our ranking by dollars recovered
Our ordering from firms we have worked inside, not a survey. The ranking is the useful part, not the values.
Why raising utilization can lower profit
Here is the trap, common enough to be a genre. Three people are on the bench. A client offers discounted work on a fixed fee that is already thin. Filling the bench raises utilization four points and the dashboard turns green, while effective rate falls, because you sold hours below what they cost to keep available. Read the two together: if utilization rises while effective rate falls, you sold hours you should have left alone. Sometimes that is right — holding a team together through a slow quarter is a real reason — but it should be a decision, not an accident.
It also helps to stop treating the bench as pure waste. A firm at 90% has no capacity for the good project that appears on a Tuesday and no slack when someone gets sick. Honest targets differ by role: delivery staff 60–75%, staff-augmentation placements 75–85%, account-carrying managers 40–60%, owners and principals 20–40%.
Timesheet lag decides whether any of this is usable
This is the most ignored operational fact in a services business. If timesheets close on the fifth, then on the third you are flying on last month's data and every corrective action is five weeks late.
Worse, hours entered a week late are entered from memory, and memory rounds: toward whole numbers, toward the project the person enjoyed, away from the awkward hour spent fixing something that should not have broken. Late timesheets are not merely late. They are wrong in a direction.
Close weekly, within one business day — not because precision is a virtue, but because a manager can act on a Monday number and cannot act on one that arrives on the sixth. The enforcement is a manager who refuses to approve an incomplete week, never a reminder email. Every practice management system ships with reminder emails and they have never worked anywhere.
Build a rate from cost, then price from something else
Cost tells you the floor. It contains no information at all about the ceiling.
The floor is arithmetic. Take base salary hourly and multiply by roughly 1.25 to 1.45 for payroll taxes, benefits, insurance and equipment; your exact multiplier is in your own books. Then allocate overhead — rent, administration, recruiting, sales, and the unbilled time of everyone who manages — commonly 60% to 120% of direct labor. What comes out is the hour at which you break even, usually higher than people expect. Market rates tend to land two and a half to three and a half times that, which is where the old rule about thirds comes from. It is a sanity check, not a strategy.
The ceiling is set by four things absent from your books: what the client's alternative costs, what the outcome is worth, how urgent it is, and how many firms can actually do the work. The most common pricing mistake is not underpricing. It is uniform pricing — one card covering both a routine migration eleven firms could do and the piece of work where the client has three real options in the country.
| Pricing model | Fits when | What you take on | How you know it is going wrong |
|---|---|---|---|
| Time and materials | Scope is genuinely unknowable and the client stays engaged | Nothing. The client carries estimate risk | Their finance team asks monthly for a forecast you do not produce |
| Fixed fee per phase | Scope is definable one phase at a time | Estimate risk, in exchange for margin upside | You stop tracking hours against the fee, so the estimate never improves |
| Retainer or capacity block | Demand is ongoing but arrives unpredictably | An obligation to keep people available | A month where they use 40% of the block and start counting |
| Milestone or outcome | The deliverable is objectively testable | Acceptance risk, sometimes on things you do not control | Acceptance criteria containing the word satisfaction |
| Blended rate | Mixed team, buyer wants one number | Team-mix risk | The mix drifts senior and the blend was priced junior |
A fixed fee is a hypothesis, and it needs a scoreboard
A fixed fee is estimated hours times a target effective rate, plus contingency. Anything else is a guess with a decimal point on it.
The contingency is not padding. It is the price of the risk you just took off the client's books, and it belongs in a range: ten to fifteen percent for a shape you have delivered several times, twenty-five to forty for a first of its kind. If you cannot say which bucket a project is in, that is the finding.
Then keep tracking hours against the fee. Firms drop this constantly — it is fixed price, why count hours — and give up the only mechanism that improves the next estimate. After a dozen projects tracked by type you have a real estimate-to-actual multiplier for your own people, which is worth more than any published benchmark.
The four numbers that fit on one page
Utilization by role group, never firm-wide. Effective rate over a trailing ninety days. Write-offs grouped by cause code. Committed hours in the next eight weeks. Everything else on a services dashboard is derived from these four or is decoration. The fourth is the only leading indicator on the list: utilization tells you what happened, backlog tells you what is about to.
The client-level view most firms skip
Compute effective rate by client, not only by role. Nearly every firm we have looked inside has one or two clients running thirty percent below the average, very often the largest — the account that got a volume discount in year one and has been quietly growing scope since. Then add cost to serve, which lives outside the timesheet: the invoicing portal that eats two hours a month, the ninety-minute weekly call with four of your people on it. A discount that looked like eight points is frequently fourteen. You do not have to fire that client. You stop treating the discount as free.
When you do not need software for this
Under roughly twenty billable people, one office, one rate card: a carefully built workbook and a disciplined weekly close will beat a practice management platform, and it will beat it in week two rather than month seven. We say that as a firm that builds software. The problem in a small practice is essentially never the tool. It is that three people define billable differently and nobody has ruled on which is correct.
The tipping points that genuinely justify a system are more than one rate structure live at once, revenue recognized on percentage of completion, more than roughly forty timekeepers, or a second location where two people must edit the same view and the workbook starts arriving by email with a name ending in _v4_FINAL_JT. If you do shop, know that most failed implementations we are asked to clean up failed on rate-card modelling: the platform holds an opinion about how rates, roles and discounts relate, your rules differ, and the gap becomes a side spreadsheet that is the real system within a year. Write your actual rules on one page first, ugly exceptions included, and make every vendor show you those rules working.
What we get called in to fix
- Utilization on a 2,080 denominator compared against a benchmark computed on 1,880
- Write-offs with no cause code, so every year's losses look like weather
- A fixed-fee portfolio with no hours tracked against fees, and no idea which project types are underpriced
- Rates unchanged for three years while salaries were raised twice
- Subcontractor cost buried in an expense line, so project margin looks fine and firm margin does not
- A dashboard refreshed nightly from timesheets that close on the fifth
- One blended rate priced on a team mix nobody actually staffed
- Bench time coded to an internal project so utilization reads healthy while nothing is sold
A four-week instrumentation pass
Getting to numbers you can act on
Step two is the one people skip and the one that decides whether any of this gets used. Nobody acts on a new number until they trust the history behind it, and restating the last year on agreed definitions is what buys that trust.
Before you change the rate card
- The denominator is written down in one sentence and has not changed this year
- Every write-off carries one of five cause codes
- Effective rate is computed by role, by client and by project type
- The rate floor comes from measured loaded cost, not from last year's card plus three percent
- Fixed fees carry a stated contingency and are tracked against actual hours
- Timesheets close weekly and a named manager approves them
- Utilization targets differ by role group and everyone knows theirs
- Committed hours for the next eight weeks are visible to whoever sells
Bottom line
The chain runs from available hours to utilization to rate realization to hours realization to effective rate, and every link multiplies the one before it. Define the denominator once and publish it. Put a cause code on every unbilled hour. Build the rate floor from real loaded cost and price above it on purpose rather than by habit. Track hours against fixed fees even though no contract requires it. Do those four and you will not need a better dashboard, because you will already know where the money went and who decided it should go there.
Frequently asked questions
It depends on the role and the denominator, which is why published benchmarks are close to useless. On a time-off-adjusted denominator near 1,880 hours, delivery staff commonly target 60–75%, staff-augmentation placements 75–85%, account-carrying managers 40–60%, owners and principals 20–40%. A single firm-wide target measures the wrong thing for most of the firm.
Utilization measures capacity: how much of the time you pay for gets pointed at client work. Realization measures price and collection. Excellent utilization with poor realization is the most common way a busy firm has a bad year.
Estimated hours times a target effective rate, plus a stated contingency — ten to fifteen percent for work whose shape you have delivered before, twenty-five to forty for a first of its kind. Then keep recording hours against the fee. That record is the only thing that improves the next estimate, and it is the first thing firms stop doing.
When more than one rate structure is live at once, when revenue is recognized on percentage of completion, above roughly forty timekeepers, or when two locations must edit the same schedule at once. Below that, a good workbook and a disciplined weekly close usually win. Write your real rate rules on one page before taking any demo.
Almost always because the extra hours were sold below the rate the firm needs, or the team mix drifted senior on work priced at a junior blend. Read utilization next to effective rate, never alone.