What is gross margin in a services business?

A software services firm bills ₹2,500 an hour. The engineer doing the work costs ₹1,200 an hour fully loaded. On that arithmetic the margin is 52%, and the founder plans accordingly.

Then the year closes and the gross margin in the accounts is 31%. Nothing was stolen and nobody was underpaid. The gap is made up of hours that were paid for but not billable, hours that were billable but written off, a change request that was delivered and never invoiced, and three weeks between projects when two people had nothing to do.

That gap is the whole subject. In a product business the cost of delivery is mostly visible. In a services business it is mostly people's time, and time leaks quietly.

Start with the fully loaded cost of an hour

Almost every services firm starts with a cost per hour that is too low, because it is built from salary divided by working hours.

The real cost includes employer PF, gratuity accrual, ESIC where it applies, insurance, equipment, software licences and a share of the space they sit in. Then the denominator has to shrink: paid leave, public holidays, sick days and training are hours you pay for and cannot sell.

An illustrative calculation. An engineer on ₹12 lakh a year costs roughly ₹13.5 lakh fully loaded with statutory contributions and insurance. Of 2,080 nominal hours, take out about 22 days of leave and holidays and some training, and roughly 1,850 hours remain as paid working hours. That is about ₹730 an hour of cost before a single hour is wasted.

But not all 1,850 hours can be sold. That is where utilisation comes in.

Utilisation: the share of paid hours that are billable

Utilisation is billable hours divided by paid hours. It is the single most important operating number in a services business and the one most often guessed at.

The hours that are paid and not billable are not waste by definition. Internal meetings, proposals, recruitment, training and the time between projects are all necessary, and a firm running at 95% utilisation is usually one that has stopped investing in anything.

What matters is knowing the number. At 1,850 paid hours and 70% utilisation, an engineer produces about 1,295 billable hours a year. The cost per billable hour is therefore not ₹730 but about ₹1,043. The margin at a ₹2,500 rate has already fallen from 71% to 58%, and nothing has gone wrong yet.

Realisation: the share of billable value you actually get

Realisation is what you invoiced and collected as a share of what the billable hours were worth at rate card.

It leaks in five ordinary places.

• ·Discounts agreed at the negotiation, which are visible and usually accepted deliberately.
• ·Write-offs at invoicing, where a project manager decides not to bill hours the client will dispute.
• ·Scope delivered free, which is the largest leak in most firms: the change that was small enough not to argue about, four times.
• ·Fixed-price overruns, where the hours went in and the price did not move.
• ·Unbilled work at the year end, which is often simply forgotten rather than written off.

Continuing the example: at 85% realisation the ₹2,500 rate becomes an effective ₹2,125. Against a cost of ₹1,043 per billable hour, the margin is about 51%, not the 71% the rate card implied.

Firms that have never measured realisation are frequently at 80% or below and have no idea, because the gap never appears as a line in any report. Nobody writes an invoice for the work they did not bill.

The bench, and why it belongs in cost of sales

Between projects, people are paid and not earning. Most Indian services firms put that cost in overheads, where it disappears into a general expense line and never affects the reported gross margin.

It belongs in cost of delivery. The bench exists because of how delivery is resourced, and hiding it makes every project look more profitable than it was.

The effect is not small. A ten-person delivery team with an average of one person on the bench at any time is carrying 10% of its delivery cost against no revenue. On a firm doing ₹4 crore of revenue at a reported 45% gross margin, moving that cost where it belongs typically takes the real figure into the mid thirties.

The point is not that the bench is bad. Some spare capacity is how a firm says yes to a good project starting next week. The point is that it should be a decision, sized deliberately, rather than a number nobody sees.

A worked example, project by project

Illustrative figures for a fictional 22-person agency, one quarter, three projects.

• ·Project A, time and materials. 940 billable hours at ₹2,400, invoiced in full, so ₹22.6 lakh of revenue. Delivery cost at fully loaded rates ₹11.2 lakh. Margin 50%.
• ·Project B, fixed price at ₹18 lakh. Estimated at 620 hours, took 890. Delivery cost ₹10.6 lakh. Margin 41%, and the overrun was absorbed entirely by the firm.
• ·Project C, retainer at ₹6 lakh a month for three months, so ₹18 lakh. Actual hours delivered were worth ₹23 lakh at rate card. Delivery cost ₹13.8 lakh. Margin 23%.

Blended, the three produce ₹58.6 lakh of revenue and ₹35.6 lakh of delivery cost, a 39% margin. Add ₹4.1 lakh of bench cost for the quarter and it falls to 32%.

The retainer is the finding. It looks like the safest work in the firm, it is the easiest to sell, and it is the least profitable, because scope grew over two years and the fee never did. That is the single most common pattern in agency and IT services economics, and it is invisible without time tracking against a rate card.

What a rate card actually has to cover

Most firms set rates by looking at what competitors charge and what clients have accepted. Building the number from the bottom up is a useful exercise even if you then decide to price differently.

The rate has to cover the fully loaded cost of the hour, divided by utilisation, plus a share of everything that is not delivery: sales, admin, finance, the founder's time, the office, and the margin the firm intends to keep.

Worked through with the earlier figures: ₹730 of fully loaded cost per paid hour, at 70% utilisation, is ₹1,043 per billable hour. If overheads outside delivery run at 25% of revenue and the firm wants a 20% operating margin, the rate needs to be around ₹1,900 before allowing for realisation. At 85% realisation, the card rate has to be nearer ₹2,250 for the firm to actually land where it intended.

That last step is the one most firms skip, and it is why a rate card that looks adequate produces a year that is not. You price at the number you want to earn and collect less than it, every month, without the gap ever appearing as a line anyone reviews.

Client concentration and the margin you dare not defend

In services businesses, margin and concentration tend to fail together. The largest client negotiated the hardest, absorbs the most senior people, and is the one nobody wants to have a pricing conversation with.

It is worth calculating margin by client each quarter and looking at the largest three specifically. A client at 45% of revenue and 22% margin is doing two things at once: subsidising nothing, and making it very difficult to say no to scope.

The answer is rarely to walk away. It is to know the number before the next renewal, to price the next piece of work properly rather than extending the old rate, and to spend the intervening quarters building enough other revenue that the conversation becomes possible.

What to measure, minimally

This does not require a consulting-grade system. Four numbers, monthly, get most of the value.

1. 01Utilisation by person and by team: billable hours divided by paid hours.
2. 02Realisation by project: invoiced value divided by billable hours at rate card.
3. 03Margin by project: revenue less the fully loaded cost of the hours delivered.
4. 04Bench cost for the month, shown separately rather than buried.

Time tracking is the precondition, and it is the part teams resist. The way to get it adopted is to be clear that it is about project economics rather than about individual performance, to keep the categories few, and to actually show the team what the numbers revealed. A firm that collects timesheets and never reports back gets worse data every quarter.

Fixed price, time and materials, and retainers

Each model fails differently, and knowing how changes how you price it.

Time and materials transfers the risk of overrun to the client, so the margin is roughly stable and the exposure is utilisation: if the client pauses, you carry the people. The discipline is a notice period in the contract and a realistic ramp-down.

Fixed price keeps the risk with you. The margin is entirely decided by estimation quality, and the honest response is to track estimated against actual hours on every project and adjust the contingency you build in until the average project lands where you intended.

Retainers drift. Scope expands gradually, the fee is anchored to what was agreed at the start, and nobody wants to reopen a comfortable relationship. The defence is a stated inclusion in hours or deliverables, a quarterly review against actual delivery, and an annual price conversation that happens whether or not anyone is uncomfortable.

What to do about a thin margin

1. 01Fix realisation before touching rates. Billing the work you already do is faster and less contentious than charging more for it.
2. 02Put a change request process in place, however light. Most free scope is given away because there is no mechanism to price it, not because anyone decided to.
3. 03Review your retainers against actual delivered hours. This is where the largest single gap usually sits.
4. 04Improve utilisation through resourcing rather than pressure. Better visibility of who is free next week does more than asking people to bill more hours.
5. 05Reprice the work that consistently loses money, and be willing to lose some of it. A client at 20% margin who absorbs your best people is costing you the projects you cannot take.
6. 06Only then raise rates, on new work first, where there is nothing to renegotiate.

The underlying shift is small and hard: a services firm has to think of its people's time as inventory that expires. An hour not sold this week cannot be sold next week, which is why utilisation and realisation deserve the attention that founders in product businesses give to unit economics.

About Simplify

Simplify is a finance clarity and investment readiness practice working with founders across India, built on six years inside startups. We write about the questions founders bring before a decision, not after it.