Billable hours

Work can be billable in principle and never appear on an invoice. It was out of scope and absorbed. It took twice as long as quoted and the overrun was swallowed. Somebody decided not to charge for it because the client had had a difficult month.

Each of those is a decision, usually a reasonable one, and almost none of them is recorded anywhere. The hours go into the timesheet as billable, the invoice goes out at the agreed amount, and the difference evaporates.

That difference is the write-off, and it is one of the most useful numbers an agency can have. An account billing its full retainer every month while writing off thirty per cent of the hours against it is not a healthy account; it is an underpriced one being subsidised by the team, and every report says it is fine.

Tracking it requires one extra field: whether the hour was invoiced. That is enough to produce a write-off rate per client, and the list it generates is usually the same list the delivery team would have given you if anybody had asked.

Three categories sit on the boundary and get classified inconsistently, which quietly breaks every number downstream.

Rework caused by the agency own error is the first. It is time spent on client deliverables, so it looks billable, and it is not: counting it flatters utilisation and margin at once, and hides the quality problem that produced it. It needs its own category, explicitly unbillable.

Client communication is the second. Meetings, calls and the twenty emails that follow are real work on the account and are usually billable. Agencies that quietly classify them as admin end up reporting a delivery team that spends sixty per cent of its time on something nobody is paying for, which is false and leads to the wrong conclusions about pricing.

Learning on the job is the third and the hardest. A junior taking four hours over something a senior would do in one has produced a billable deliverable at an unbillable cost. Most agencies bill the hour and absorb the difference, which is right, and it should be visible rather than invisible.

What matters less than where the lines are drawn is that they are written down and applied the same way by everybody. Two people classifying differently produces a number that cannot be compared with itself over time.

Time tracking has a bad reputation in agencies for a specific reason: it is frequently introduced as a control measure and immediately understood as surveillance. At that point people start filling it in to look reasonable rather than to be accurate, and everything built on it becomes fiction.

The version that survives is coarse. Account and activity type, in half-hour blocks, logged daily. Precision beyond that produces resistance without producing better information, because nobody genuinely remembers whether a task took forty or fifty minutes.

Daily matters more than the granularity. Timesheets reconstructed on a Friday are estimates dressed as data, and the errors are not random: the memorable work gets over-recorded and the fragmentary work disappears entirely.

And the purpose has to be stated and then visibly honoured. If tracking is for pricing, it should show up in pricing conversations and never in a performance review. The first time somebody hears their utilisation quoted at them in an appraisal, the data quality drops permanently across the whole agency.

Hours are the input to almost every commercial question an agency has, which is why the absence of them forces decisions to be made on feel.

Per-client gross margin needs them, because cost of delivery is people time. Whether a retainer is priced correctly needs them. Whether a service line is worth offering needs them. Whether to take a new account next month needs them, via capacity. Scope creep is invisible without them.

None of those questions requires precise data. All of them require data that is roughly true and consistently collected, which is a much lower bar and a very different exercise from a legal firm billing in six-minute increments.

It is worth saying that to a team, because the resistance usually assumes the harder version. Knowing that an account took about ninety hours rather than the sixty it was priced for is enough to reprice it, and nobody needed to account for their afternoon.