A professional services firm sells capacity that expires daily. An hour not sold on Tuesday is not held in stock for Wednesday. Yet most firms in this sector run finance systems designed around inventory, purchase orders and periods, and then wonder why the monthly close explains what happened to margin six weeks after anyone could have changed it. The gap is not reporting. It is that the system's unit of account and the firm's unit of economics are different things.
In a product business the cost sits in the warehouse. In a services business it walks out of the building at six o'clock whether it was sold or not
That single asymmetry drives everything that follows. Capacity cannot be stored, discounts cannot be recovered through volume, and a slow week is gone permanently. Decisions therefore have to be made weekly at engagement level, which is not what a general ledger is built to support.
Four numbers, and the ways firms mislead themselves with each
Utilisation. Chargeable hours divided by available hours. The manipulation lives entirely in the denominator: excluding annual leave, excluding training, excluding business development, excluding anyone with a title above a certain grade. Every exclusion raises the number and reduces its meaning. Pick one definition, apply it to everybody, publish it, and never change it mid-year. A firm reporting eighty-five per cent utilisation and losing money is usually measuring against a denominator it invented. Realisation. Value billed against value recorded at standard rates. This is where the discounting, the scope creep and the partner-level goodwill get quantified. Firms that track utilisation without realisation routinely discover that their busiest teams are their least profitable ones. Engagement margin. Revenue less the actual cost of the people who delivered it, at the engagement rather than the service line. Averages across a practice hide the two projects destroying the quarter. Capacity coverage. Signed and committed work expressed in weeks of team capacity, not in currency. Currency backlog tells you about revenue; weeks of capacity tells you whether to hire, and it is the only one of the four that is forward-looking.
| Measure | What it describes | Watch for |
|---|---|---|
| Utilisation | Chargeable hours divided by available hours | An altered denominator can inflate the number |
| Realisation | Value billed against value recorded at standard rates | Discounting, scope creep and write-offs |
| Engagement margin | Revenue less actual delivery-person cost at engagement level | Practice averages can hide individual loss-making projects |
| Capacity coverage | Committed work expressed in weeks of team capacity | Currency backlog is not the same as staffing coverage |
Qualitative summary of this article's source text, not a measured outcome or performance estimate.
Why generic ERP disappoints here
The ledger is organised by period and account. The firm is organised by engagement and person-week. When the system cannot carry engagement as a first-class dimension, four workarounds appear, and every firm reading this will recognise at least three: time capture lives in a separate tool that reconciles monthly; work in progress and unbilled revenue live in a spreadsheet maintained by one person; revenue recognition is a manual journal computed after the fact; and the engagement leader sees actuals when the close finishes rather than while the work is running. None of those is a reporting failure. They are the consequence of a data model in which the project is a text field.
Revenue recognition is where this stops being an internal matter
Services contracts are generally recognised as the work is performed, which means someone must estimate completion. That estimate drives revenue, margin and work in progress, and it is only as good as the discipline behind it. The practice that separates well-run firms from the rest is unglamorous: the engagement leader updates the estimate to complete every week, in the system, before invoicing. Where that estimate is refreshed only at month end, margin is discovered rather than managed, and the discovery is nearly always unpleasant — an overrun that has existed for five weeks surfaces on the day it becomes irreversible.
Sequence the implementation properly
In order, because the order is where these projects fail. Daily time capture with an enforced weekly cut-off. Engagement as a structural dimension in the ledger, carried through to invoicing and cost. Rate cards with a documented write-off approval path, so discounts have an author. Weekly estimate-to-complete owned by the person running the work. Only then, dashboards. Firms that begin with dashboards build a beautiful view of unreliable data, and the fee earners learn within a month that the numbers can be ignored.
The hard part is not the software
It is getting a senior fee earner to record time within twenty-four hours and to own a completion estimate in writing. That is a management problem with a systems component, not the reverse. Two governance mechanisms do most of the work: no invoice is raised without a current completion estimate, and timesheet compliance is reported at partner level alongside financial performance. Both are unpopular. Both are cheaper than the alternative, which is a write-off discovered at year end.
Practical Guidance for Services Operations Assessment
- Fix one utilisation definition and apply it to every grade.
- Track realisation beside utilisation, never alone.
- Make engagement a ledger dimension, not a description field.
- Update estimates to complete weekly, owned by the engagement leader.
- Require an approver for every write-off and report them by partner.
- Express backlog in weeks of capacity, not only in currency.
- Capture delivering entity and location on every time entry.
- Sequence time capture before dashboards, always.
The Regional Angle
Three factors reshape this for firms operating across the Gulf. The first is that the timesheet is quietly becoming a tax record. Value-added tax already creates timing questions on continuous supplies and milestone billing, and corporate tax arrives in the Emirates for financial years beginning from the middle of next year. For firms that deliver work across entities — a consultant employed by the Dubai company working on a Riyadh engagement, a shared partner supervising both — intercompany charging will need to be evidenced rather than assumed, and the evidence is the time record. This has a concrete systems implication that most regional firms have not yet made: every time entry needs to carry the delivering entity and the physical location of the person delivering the hour, alongside the client and engagement. Retrofitting that dimension after a year of records is painful; adding it before the first tax period is trivial. Firms that ignore it will spend next year reconstructing transfer pricing support from calendars and travel bookings. The second is that bench cost behaves differently here. In markets with short notice periods, utilisation shortfalls can be managed by flexing headcount. In this region staff are recruited internationally on sponsored visas with real mobilisation cost, notice periods are contractual, and end-of-service entitlements accrue, so a decision to reduce capacity is expensive, slow and partly irreversible. The planning consequence is that bench cost should be modelled as fixed for at least two quarters, which changes how aggressively a firm should chase marginal work: a discounted engagement that covers salary cost for otherwise idle staff is a better decision here than the same discount would be in a market with flexible labour. It also makes secondment between group entities — rather than hiring and releasing — the first lever to reach for, which again depends on the entity and location fields being in the time record. The third is that realisation and cash diverge more sharply here than the textbooks allow. Government and semi-government clients dominate the pipeline in several regional markets, and their invoicing requires certified deliverables, portal submission and an acceptance step before any payment clock begins. A firm can bill at full rate, report healthy realisation, and still wait five months for cash — because the bottleneck is certification, not the payment run. Track days to certification as a metric in its own right, separately from days sales outstanding, assign it to the engagement leader rather than to finance, and put it in the same weekly review as the completion estimate. The firms that manage this well treat obtaining the client's sign-off as a project deliverable with a named owner and a date, not as an administrative step that happens after the work is done.
The objection worth taking seriously
The strongest objection is that professional services firms have run for decades on spreadsheets, partner judgement and a monthly review, and that services automation projects have a poor record precisely because fee earners will not comply with them. The discipline matters far more than the tool; a firm with weekly partner scrutiny and a shared spreadsheet will outperform one with an expensive platform and no habit. Buying software to impose management attention is a recognisable way of avoiding the harder conversation. For a single-office firm of thirty people, that is simply right, and a platform purchase at that scale usually buys administration rather than insight. Where it breaks down is at the point most growing regional firms have already passed: several entities, more than one country, work delivered by people employed by a different company from the one invoicing, and tax authorities that will soon want to see the basis of intercompany charges. At that scale the spreadsheet stops being a lightweight alternative and becomes an unauditable dependency on one person. The discipline argument also cuts the other way — weekly habits are easier to sustain when the clock is enforced by a system that closes the week than when it depends on a partner chasing colleagues by email every Friday.
Common Questions
What utilisation target is right?
There is no universal figure, because the number is meaningless without its denominator. Define the denominator, measure consistently for two quarters, and set targets from your own data rather than from a benchmark computed differently.
Do we need a dedicated services platform or can our ERP do this?
Many mainstream systems handle project accounting adequately if engagement is a real dimension and time capture is integrated. The question to ask a vendor is not whether it supports projects, but whether an engagement leader can update a completion estimate and see live margin without finance intervening.
How do we get fee earners to submit time on time?
Make it short, mobile and weekly; make compliance visible at partner level; and tie invoicing to it. Firms that solve this treat late time as a management issue rather than an administrative one.
What should we expect over the next twelve months?
Expect services automation vendors to add AI features this year, mostly suggesting time entries from calendar and email activity and drafting engagement narratives — genuinely useful for compliance, not transformative. Expect regional corporate tax to force entity-level project reporting into scope within eighteen months, which will catch firms that treated entity as a reporting afterthought. Expect utilisation targets to come under pressure as salary costs rise faster than billing rates. And expect the firms that can price by outcome rather than by hour to begin separating from the rest, because they are the only ones for whom a productivity gain increases margin instead of reducing revenue.
Services Operations Assessment — we fix the four numbers that run the firm, get engagement into the ledger properly, and make sure your time records will survive the first corporate tax audit.