Consulting Utilisation Rate: Track It Weekly and Price Projects With AI

Consulting utilisation is billable hours divided by available hours for the same period. Track it weekly per person and for the team, separate bench time from internal work, and use a break-even rate built from your utilisation to price projects. Let AI write the weekly commentary, but calculate the numbers in a spreadsheet.

The number is simple and the arguments are about the definitions. Whether holidays count, whether business development is billable, whether a part-timer's available hours are 24 or 40: each choice moves the percentage by several points. So the first job is to write down your definitions, once, in the header of the sheet. After that the weekly roll-up is a ten-minute task, and the number starts to be useful for what it is really for: deciding what a project must cost.

Follow me on Instagram@sagnikteaches

The definitions to fix before you track anything

Use these four, and change them only on purpose.

Connect on LinkedInSagnik Bhattacharya

• Available hours: contracted hours in the week, minus planned leave and public holidays. A full-timer on a 40-hour week with one day off has 32 available hours.
• Billable hours: hours charged to a client project, whether or not you have invoiced them yet. Write-offs are tracked separately, below.
• Internal hours: necessary work that no client pays for directly: selling, proposals, training, admin, tooling, management.
• Bench time: available hours with no billable or internal task assigned. It is the gap, and it is the number that tells you whether you have the right number of people.

The identity to check every week is billable + internal + bench = available. If a person's rows do not add up, a timesheet is missing hours, and you have found a data problem before it distorts a price.

Subscribe on YouTube@sagnikteaches

Utilisation = billable hours ÷ available hours. Some firms divide by hours worked instead. That flatters people who work longer weeks and hides overload, so I would use available hours and show hours worked as a separate column. The utilisation calculator tutorial covers building the formula from raw timesheets if your data is untidy, and the agency version looks at spotting overload in a team that bills by deliverable. This one is about a consulting practice, where the same figure drives project pricing.

A weekly roll-up with numbers

Here is one illustrative week for a five-person consultancy. The figures are made up, but the structure is the one to copy.

Check the sums: 132 + 32 + 28 = 192. Team utilisation is 132 ÷ 192 = 68.75 percent, shown as 69. The table says four things at a glance. Person D is at 85 percent, so there is no slack for a delay on D's projects. Person E has 12 hours of bench, which is where the next project should go. Internal hours are 17 percent of available (32 ÷ 192), which is worth watching: if it creeps upward week after week, the firm is spending more time on itself than on clients. And the principal's 75 percent is high for someone who also sells, so check where proposals are being written at night.

Getting the roll-up from timesheets with AI

Most timesheet tools let you export a CSV with person, date, project, hours and a billable flag. The AI job is to turn that export into the table above plus a short note. The safe way to do it is in two steps. First, calculate the table with formulas, for example in Excel:

=SUMIFS(Hours, Person, $A2, Type, "Billable", Week, $H$1) / $B2

Here Hours, Person, Type and Week are named columns on the data tab, H1 holds the week you are reporting and B2 holds that person's available hours. Then hand the finished table, not the raw export, to an assistant for commentary. The reason is that language models can slip on arithmetic across many rows, and a wrong percentage in a weekly note erodes trust in the whole routine. If you do want the model to work from raw data, ask it to show the sums and then check two rows by hand. The ChatGPT and Excel guide covers how to get a model to write and explain the formulas without doing the sums itself.

A prompt that works on the finished table (use first names or initials, and remove client names):

Here is this week's utilisation table for a five-person consultancy
(hours). Definitions: utilisation = billable / available; bench =
available hours with nothing assigned.

[paste table]

Write a 120-word note for the partners. Say: team utilisation and how
it compares with last week's [paste last week's figure]; anyone above
80 percent or below 55 percent; total bench hours and who has them.
Do not recommend hiring or firing. Do not add any number that is not
in the table.

A likely output, labelled illustrative: "Team utilisation was 69 percent, up from 66 percent last week. D is at 85 percent, above the 80 percent threshold. E is at 45 percent, below the 55 percent threshold, with 12 bench hours; A has 4, B has 8, C has 2, D has 2. Total bench time is 28 hours." Check: 4 + 8 + 2 + 2 + 12 = 28. Correct. If the note had said "total bench time is 30 hours", you have caught a typical slip. That is why the last instruction, no number not in the table, is in the prompt, and why you still read every figure.

What to do with bench time

Bench is not a failure; it is information. Two hours a week per person is normal friction between tasks. Twelve hours for one person for a second week running is a scheduling problem or a pipeline problem. A simple decision rule:

• Under 10 percent of available hours on the bench: leave it.
• 10 to 20 percent for a team member for two weeks: assign internal work with a purpose, such as a proposal, a reusable template or a case study.
• Over 20 percent for three weeks: it is a pipeline issue. Look at the sales funnel, not the person.

These thresholds are starting points, not rules. Change them to match how lumpy your work is. A practice with short projects will swing more week to week than one with long retainers, so use a four-week rolling average when you judge a trend.

Turning utilisation into a break-even rate

This is the part that links the weekly number to pricing. A consultant costs you a certain amount a year, and only some of their hours are billable, so each billable hour has to cover the whole cost. The break-even hourly rate is:

Break-even rate = fully loaded annual cost ÷ (available hours × target utilisation).

An illustration with invented figures. A consultant's fully loaded annual cost is $120,000: salary, payroll costs, software, equipment, a share of rent and management time. Available hours are 1,760 a year (220 working days at 8 hours, after holidays and leave). At a target utilisation of 70 percent, billable hours are 1,760 × 0.70 = 1,232, and the break-even rate is $120,000 ÷ 1,232 = $97.40 an hour. To earn a 25 percent margin on price, divide by 0.75: $97.40 ÷ 0.75 = $129.87, so about $130.

Now the cross-check against your rate card. Say your published rate is $150. If utilisation drops to 60 percent, billable hours fall to 1,056 and the break-even rate rises to $120,000 ÷ 1,056 = $113.64. Margin at $150 is (150 − 113.64) ÷ 150 = 24.2 percent. At 70 percent, margin at $150 is (150 − 97.40) ÷ 150 = 35.1 percent. The same rate card earns eleven points more margin when utilisation is ten points higher. That is why a weekly utilisation figure belongs in the pricing conversation, and why discounting a rate card when you are quiet is expensive: it lowers price just as utilisation is already working against you.

If you do not know the fully loaded cost, estimate it, label it an estimate in the sheet and refine it when the accounts are in. A rough figure you update is far better than none.

A project price, worked end to end

Say a client asks for a six-week process review, and you have to price it. Use hours, not a gut feel, and let utilisation set the rate.

1. Estimate delivery hours by task. Interviews 10, analysis 30, workshops 24, report 36, handover 8, project management 12. That is 120 hours.
2. Add a contingency. 15 percent of 120 is 18, so 138 hours.
3. Split by seniority. 20 hours at a principal rate of $220 and 118 hours at a consultant rate of $140. That is $4,400 + $16,520 = $20,920.
4. Check the blended rate. $20,920 ÷ 138 = $151.59 an hour. Compare it with the break-even rate for the people doing the work. If the blended rate is below break-even at your target utilisation, the project loses money before any overrun.
5. Add the unbilled front end. Proposal and scoping took 8 hours that the client will never see on an invoice. These are internal hours, and they are why a firm's rate has to be above its break-even rate.
6. Quote a fixed price. Round to $21,000 and state the assumptions and the change process in the engagement letter.

The engagement letter tutorial has wording for the scope limits that make a fixed price safe, and the pricing models tutorial compares fixed fee, retainer and value pricing for when the hours-plus-margin sum is not the right basis. Hours-based pricing is a floor, not a ceiling: if the work is worth far more to the client than the hours suggest, a value-based fee may be fairer to you.

How much a few points of utilisation are worth

A sensitivity table makes the point better than a paragraph. Using the same illustrative $120,000 cost, 1,760 available hours and a $150 rate card:

Read down the right-hand column: moving from 50 to 60 percent adds 15 points of margin, while moving from 70 to 80 adds only 8. The first ten points matter most, which is why a practice at 50 percent should fix its pipeline before worrying about its rate card, and a practice at 80 percent should look at price and capacity rather than squeezing more hours out of people.

A solo consultant, with the same sums

For a one-person practice, utilisation is lower by design, because the same person sells, delivers and does the admin. Say a solo consultant needs $80,000 a year from the business to cover personal income and costs, has 1,760 available hours, and realistically bills 55 percent. Billable hours are 968, and the break-even rate is $80,000 ÷ 968 = $82.64. A published rate of $100 gives a margin of (100 − 82.64) ÷ 100 = 17.4 percent. If the same person improves billing to 65 percent, billable hours rise to 1,144 and break-even falls to $69.93.

The solo case shows why many one-person consultants price per project rather than per hour: a project price of $12,000 for 90 hours is $133 an hour, which is the number to compare with the break-even rate, and it rewards finishing faster. Keep a log of estimated against actual hours for every project, so each quote improves the next.

Realisation: the number that sits next to utilisation

Suppose a consultant logs 40 billable hours on a client and invoices for 30, because the client pushed back or the scope drifted. Utilisation counts 40. Revenue reflects 30. The ratio, 30 ÷ 40 = 75 percent, is realisation. A team at 70 percent utilisation and 75 percent realisation is effectively earning at about 52 percent of potential (0.70 × 0.75 = 0.525), which changes the break-even sum completely. If you track only one of the two, you will think the firm is healthier than it is. The weekly roll-up should include a monthly realisation line, and the AI commentary should mention it whenever it falls below your target, say 90 percent. Persistent low realisation is a scope or estimating problem, and it connects directly to the change-request process in your engagement letter.

Using AI to sanity-check an estimate

An assistant is useful as a second pair of eyes on your task list, not as the estimator. Paste the task list and hours, with the client name removed, and ask: "Which tasks here are most likely to run over, and why? Which common tasks for a process review are missing?" You might get "no time for a kick-off meeting" or "report review cycle not included". Add the ones that are real. Reject the ones that are not, since the model does not know your scope.

Then compare with history. If your last four reviews took 142, 150, 131 and 160 hours, an estimate of 138 is optimistic, and you should find out why before you quote. That comparison, estimate against actuals, is the best use of the timesheet data you are already collecting, and the roll-up can include a monthly line for each completed project: estimated hours, actual hours and the difference.

Four weeks of roll-ups, and what they showed

Illustrative: a small practice tracks team utilisation for four weeks and gets 69, 64, 71 and 58 percent. The mean is (69 + 64 + 71 + 58) ÷ 4 = 65.5 percent. Against a target of 70, it is below, and the break-even sum says each point under target costs margin. The drop to 58 in week four coincided with a project ending and no start date for the next. The note to partners reads "bench 52 hours this week, 38 of them from the project that closed on Tuesday; two proposals out, no start before week seven". Now the pipeline conversation is specific: two proposals, how likely, and how soon.

A different case: a firm sees 88 percent for three weeks. It looks wonderful. Look closer and the internal hours are 3 percent, no proposals have been written, and the pipeline for week eight is empty. High utilisation today with no selling time is next quarter's bench. The weekly note should report the two numbers side by side, utilisation and internal hours, for exactly this reason.

Where the numbers go wrong

Missing time. Timesheets completed on Friday afternoon from memory understate hours, usually the awkward unbillable ones. Ask for daily entry and check the identity (billable + internal + bench = available) every Monday.

Write-offs hidden in billable. A timesheet tool that marks every client-coded hour as billable hides the hours you later write off. Add a write-off column, filled in when the invoice goes out, so the realisation line above has data behind it. Low realisation at high utilisation means the work is being given away.

Mixed definitions. One person counts a client lunch as billable and another does not. Put a one-line rule beside each category on the timesheet form.

The part-timer problem. A person on 24 hours a week has 24 available hours, not 40. If the sheet uses 40, their utilisation looks like 40 percent for a week of full billing.

Distinct from an agency retainer. A consultancy that sells time against scoped projects can price from hours and utilisation. A retainer-based agency often prices on deliverables and monthly allowances; if that is you, the agency pricing tutorial fits better, though the break-even sum above still tells you what the retainer must cover.

A ten-minute Monday routine

1. Export last week's timesheet CSV and paste it into the Data tab.
2. Check every row of the table satisfies billable + internal + bench = available. Chase missing time before doing anything else.
3. Read the formulas' results for team and person utilisation and bench hours.
4. Paste the finished table into your assistant with the prompt above and read the note against the table, number by number.
5. Add one line of your own: what changed, and what you will do about it.
6. Once a month, update the break-even rate with current costs and compare it with your rate card.

The aim is a short, accurate note that partners read, and a pricing habit that starts from your own costs. If you want a second pair of eyes on how your data is set up, that is a good topic for a call, but most of this you can do with one spreadsheet and a consistent definition of an hour.

Questions about consulting utilisation

What is a good utilisation rate for a consultancy?

There is no single right number. It depends on seniority, how much selling and delivery overhead a person carries, and how fast work is replaced. Set a target from your own cost base using the break-even sum below, then compare actual weeks against it, rather than copying someone else's percentage.

Should leave and public holidays count as available hours?

Usually not. Available hours are the hours a person could do billable or internal work, so subtract planned leave and holidays. Counting them as available makes utilisation look worse in holiday weeks for reasons nobody can fix. Pick one definition and write it in the sheet header.

Can I let AI calculate the percentages?

Let a spreadsheet calculate them and let AI explain them. Language models sometimes slip on arithmetic, especially across many rows. Give the model the finished table and ask for commentary, or ask it to write the formula and check the result against a manual sum.

Is high utilisation always good?

No. A person at 90 percent every week has no room for selling, learning or delivery slip, and tends to burn out or miss follow-on work. Watch the trend in pipeline and quality alongside the figure, and treat a sustained very high number as a warning, not a win.

Further reads

• AI in the Consulting Industry: Package a Fixed-Scope Diagnostic to Sell — A fixed-scope offer built on the pricing sums here.
• How to Negotiate an AI Consulting Quote Without Cutting Corners — How buyers push on a quote, so you can price with room.
• Should You Pay an AI Consultant on Results? Success Fees Explained — When pricing on results rather than hours makes sense.
• AI Consulting Report Writing: From Findings to a Client-Ready Draft — Cut the unbillable hours spent on report drafting.
• What a Typical AI Consulting Engagement Looks Like, Week by Week — A realistic week-by-week shape for planning hours.
• Virtual CFO Pricing: Build a Monthly Package With AI-Prepared Reports — How a firm can price a monthly virtual CFO package: three tiers, the reports in each, review time, tool costs and add-on rules.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: standard professional-services utilisation definitions; all figures in the worked examples are illustrative and invented for this tutorial.