A full calendar is not proof that you’re making money. It proves that people are spending time. On a fixed fee, that can be the expensive part.
Your control point is realization: the value that survives from planned work to billed and collected value, then to contribution after delivery cost.
Utilization tells you whether eligible capacity reached client assignments. It doesn’t tell you whether the fee still covers the work that arrived there. A calendar can be packed with engagements that are quietly eating the firm.
CRA International reported utilization rising from 75% to 77% in fiscal 2025, with billable hours up 6.0% and revenue up 9.3%. Those results belong to one firm and aren’t a target for another.
They do make the operating point. Utilization, headcount, rates, mix, and engagement pricing are separate levers, and a utilization percentage can’t carry all that freight. It barely fits in the dashboard cell.
Two Measures
Start by separating the two records that too often get one friendly label.
- Utilization records how much eligible capacity reaches client work. Its denominator is a firm policy, not an industry fact. Huron, for example, excludes local holidays and vacation from available hours, reports the measure for professionals who primarily bill hourly, and omits managed-services professionals.
- Billing realization records the portion of recorded work value that survives discounts and write-downs to become an invoice. Collection realization records the portion of billed value ultimately collected. They are different losses at different stages.
- Contribution asks what remains after the actual labor mix and delivery cost. It is the test a fixed fee has to pass after the client has received the work, not when the proposal is celebrated.
Those labels have to sit in a metric dictionary with the population, numerator, denominator, data cutoff, leave and training treatment, currency basis, write-off stage, and policy version.
The SEC’s KPI guidance uses the same discipline in public reporting: define the calculation, explain why it’s useful, and disclose a material method change rather than hiding it behind a familiar name. Internally, freeze the definition for the reporting period before you start comparing practices or paying anyone on it.
The separation matters most when your invoice doesn’t use hours.
On time-and-materials work, entered, billable, and approved time can each affect billing. On fixed-bid work, milestone timing releases the invoice instead.
The effort still reaches project cost, staffing capacity, and the estimate at completion. Your invoice has stopped showing the meter. You haven’t stopped paying for it.
The Risk
A fixed fee is an allocation of scope risk, not a prettier invoice format.
Under the U.S. Federal Acquisition Regulation, a firm-fixed-price contract doesn’t adjust merely because the contractor’s cost experience changes; it places the resulting profit-or-loss risk on the contractor. That’s a jurisdiction-specific federal rule, but it describes the economic bargain in ordinary fixed-fee work too.
The trade is worthwhile when you can bound the work. A repeatable tax filing, payroll run, standard search, or routine conveyance has prior cases that help define normal work and exceptions.
Technical work can be bounded by surveys, drawings, assumptions, review stages, and client-supplied data. Creative work can be bounded by named deliverables, media, and revision rounds.
“Fixed fee” without the boundary is just a donation with project management software.
Before you accept the fee, put three things in the authorization gate:
- The included deliverables, exclusions, dependencies, acceptance criteria, staffing assumptions, and expense treatment.
- The event that changes scope: an added deliverable, a new site condition, missing or corrupted client data, an extra revision round, or a client-requested decision outside the agreed output.
- The authority that can approve the change. A forecast showing more effort is an internal alarm, not permission to invoice more.
Where you can’t estimate extent or duration with reasonable confidence, use a staged boundary. Price discovery or investigation under time-and-materials with labor categories, material treatment, approval gates, and a ceiling, then price the defined deliverable.
The cited FAR model requires fixed labor-category rates, materials at allowed cost, surveillance, and a ceiling exceeded at the contractor’s risk. Private engagements aren’t federal contracts, but the control logic travels. Uncertainty needs a visible boundary.
The Engagement Model
| Work type and engagement model | Billing unit | Scope boundary | Change trigger | Utilization view | Realization view | Contribution view |
|---|---|---|---|---|---|---|
| Repeatable, bounded delivery — fixed fee | Named deliverable, phase, or milestone | Included output, foreseeable variants, exclusions, acceptance criteria, and client dependencies | Contractually valid approval for work outside that boundary | Actual effort against planned staffing and capacity | Fee retained against planned and actual effort; separate billing from collection realization | Actual delivery mix, rework, reuse, and effort against the fixed fee |
| Novel discovery or remediation — time-and-materials | Fixed labor-category rate plus separately treated materials | Approved categories, reporting interval, reimbursables, and ceiling | Ceiling or approved addition before further exposure | Entered, billable, and approved time by category | Recorded value that survives discounts or write-downs, then billed value collected | Rate yield less actual delivery cost; do not mistake more hours for efficiency |
| Uncertain first phase, then defined output — staged hybrid | Time-and-materials discovery followed by a fixed phase or deliverable | Discovery question first; defined findings and output second | Evidence from discovery changes the next authorized scope | Capacity tracked separately for discovery and delivery | Read each phase under its own fee basis rather than averaging the whole job | Whether the discovery reduced estimation risk before the fixed fee began |
That table isn’t a menu of fashionable fee labels. It’s a risk map.
Time-and-materials transfers much of the quantity risk toward the buyer, but it doesn’t create a positive profit incentive for labor efficiency or cost control. Fixed fees put the overrun on you, while letting you keep the gain when a reusable method, automation, or better delegation reduces effort.
That’s why the same team can become more productive and less profitable if its pricing model is attached to the wrong layer of work.
Locate the Loss
Use the symptoms to locate the broken stage. “People are busy” is a weather report, not a diagnosis.
Everyone is booked, invoice totals are stable, and profit per employee falls → the fee-to-effort stage is broken
Reforecast actual hours plus remaining effort against the contracted fee. Don’t use utilization to explain a fee that no longer covers the staffing mix.
Fixed-fee jobs overrun repeatedly → the scope gate is broken
The proposal didn’t price an uncertainty, name an exclusion, identify a client dependency, or define the event that authorizes a change. Separate the work that’s genuinely outside scope from rework you own.
Partner hours rise while junior utilization is strong → the leverage stage is broken
Senior review has replaced the lower-cost mix you assumed in the proposal, or delegated work is coming back as error correction. Leverage is successful transfer of work, not a headcount ratio printed after year-end.
Average realization looks stable while a few projects feel cursed → the segmentation stage is broken
Sort results by client, matter class, work type, fee model, and staffing mix. Under fixed fees, effort against estimate, reuse, rework, change-order capture, and contribution by matter class expose losses that a practice average folds into wallpaper.
A milestone invoice releases on schedule while contribution keeps falling → the billing stage is being mistaken for earnings
Hours, invoicing, and accounting revenue are three progress records. Under IFRS 15, over-time recognition requires an appropriate measure of progress, and your payment schedule isn’t automatically that measure.
A manager removes hours from the file to protect realization → the evidence stage is broken
Keep recorded, supervisor-approved, billable, invoiced, written-down, and collected statuses separate. For covered U.S. employment records, specified payroll records are preserved at least 3 years, while supporting time and earnings records generally have a 2-year period.
Preserve the original entry, editor, timestamp, reason, and approving authority even where different retention rules apply.
The Reforecast
The corpus’s invented Northline engagement makes the arithmetic visible without pretending to be a benchmark.
It sold for CAD 48,000 over 8 weeks, was estimated at 320 labor hours, and invoiced 30% at signing, 40% after the workshop, and 30% on acceptance. The original shadow fee was CAD 150 per planned hour.
When the estimate at completion rose to 390 hours, the shadow fee fell to about CAD 123 before the client invoice changed a cent.
By closeout the engagement used 386 hours: 26 partner, 80 manager, and 280 analyst. The contract, invoices, and collections still each totalled CAD 48,000, but the effective fee per hour was about CAD 124, and partner hours exceeded plan by 30%.
The calendar was full, the invoices were healthy, and the delivery economics were worse. That isn’t a billing problem. It’s an estimating, scope, and staffing record waiting to be used.
Run that record for every fixed-fee segment at period close:
- Compare planned hours, approved actuals, remaining effort, estimate at completion, contracted fee, invoiced value, collected value, and actual staffing mix.
- Code the source of variance as valid scope change, avoidable rework, client dependency, estimation error, or deliberate relationship investment. Don’t let a general “overrun” code teach sales nothing.
- Feed the observed task mix and exception pattern into the next proposal. Hours in a fixed-fee practice move from invoice construction to estimating and capacity control; they don’t retire.
Don’t average the results too early. A routine matter class can carry an unbounded one for a long time, especially when the unbounded one comes from a large client or a charismatic partner.
Segment first. Then decide whether the next quote needs a different boundary, a different staffing mix, a different change trigger, or a different fee model.
Utilization tells you where capacity went. Realization tells you what value survived. A fixed fee keeps the efficiency gain. It also keeps the overrun.
Price the boundary, not the hopeful estimate. Record the hours, then segment the loss. Let contribution decide whether the full calendar was worth it.