Rounding billable time is a pricing decision, not an accounting one

Almost everyone who bills by the hour rounds. Very few people have decided how.

The rule usually arrives by accident: it is what the last agency did, or what the invoicing tool defaulted to, or a habit picked up from whoever taught you to freelance. Then it runs, unexamined, across every entry for years.

That is worth a few minutes of attention, because across a fragmented year the rule is not a detail. It is a price change you never announced.

The three common rules

Round up to the nearest unit. A four-minute call becomes six minutes, or fifteen. This is standard in legal and much of professional services, and the justification is real: an interruption costs more than its duration, because the work it interrupted has to be picked back up.

Round to the nearest unit. Four minutes becomes zero, eleven becomes twelve. Symmetric, and over a large number of entries it approximately cancels out.

Do not round at all. Bill the measured figure. Increasingly common where the tracking is automatic and the client can see the timeline.

None of these is dishonest. All three are defensible. They just produce different numbers.

The size of the difference

Take a consulting day made of genuinely fragmented work: eighteen separate pieces, most of them short. Emails that required thought, two calls, a handful of small tasks, one longer block.

Under a fifteen-minute round-up, each of those eighteen pieces is billed at a minimum of a quarter of an hour. Under round-to-nearest, the short ones mostly disappear and the medium ones move slightly up. Under no rounding, you bill what happened.

The spread between the highest and lowest of those three, on the same day of the same work, is not marginal. Now apply it to two hundred and thirty working days.

That is the point. Not that any rule is wrong, but that choosing between them is a decision about your annual income, and it deserves to be made deliberately rather than inherited.

What makes a rule defensible

A rounding rule survives a client conversation when three things are true.

It is stated up front. In the contract or the engagement letter, before the first invoice. A rule the client discovers by auditing your invoice is a rule you are now defending rather than applying.

It is applied consistently. The same rule for every entry, including the ones where it works against you. A rule that only ever rounds in your favour is not a rounding rule, and a client who notices will describe it accurately.

It is applied to something measured. This is the one most people miss. Rounding a figure you measured is arithmetic. Rounding a figure you estimated is estimating twice, and the two errors compound in whichever direction you were already inclined.

That last point is why rounding gets much less fraught when the underlying tracking is automatic. If the raw entry says 43 minutes because 43 minutes is what the record shows, the rounding rule is a visible, explainable transformation applied on top. If the raw entry says “about 45, I think”, the rule is decoration on a guess.

Round the total, or round the entry?

A quieter question, and it changes the answer more than people expect.

Rounding each entry up to fifteen minutes and adding them gives a different total from adding the raw minutes and rounding once at the end. On a day of eighteen short entries, dramatically different.

Per-entry rounding is the norm in professional services, and the argument for it is the interruption cost mentioned above: each piece of work carries a real switching overhead, so each piece earns its minimum. That is coherent.

What is not coherent is doing per-entry rounding while describing it as “we round to the nearest quarter hour”, which a client will hear as applying to the invoice total. If your rule is per-entry, say per-entry.

A note on the arithmetic underneath

One technical point that matters more than it sounds.

If you compute money from hours, round the hours, and then add up the money, you will sometimes get a different total from adding the raw minutes and computing money once. The differences are cents. On a single invoice nobody notices.

On a year of invoices, reconciled against an accounting system, those cents become a discrepancy somebody has to explain, and the explanation is unsatisfying.

Sum in whole minutes, convert at the end, and compute money in whole cents per row. It costs nothing to do it correctly and it removes a category of conversation entirely.

What to actually do

Spend ten minutes on this once.

1. Work out what rule you are currently using. Look at your last five invoices rather than at what you think you do.
2. Decide whether it is the rule you want, given how fragmented your work actually is. Highly fragmented work makes round-up much more consequential than continuous project work does.
3. Write the rule into your contract in one sentence, including whether it applies per entry or to the total.
4. Apply it to measured time rather than remembered time. If you are estimating the underlying figure, the rounding rule is the smaller of your two problems.

Then leave it alone. The value here is in having chosen, not in optimising. A rule you picked deliberately and stated clearly will not cost you a client. A rule nobody agreed to, applied to numbers nobody recorded, eventually will.