Billable vs non-billable time: a decision table for agencies

The short answer

Billable time is time spent for a client that your contract lets you charge for, or that counts against a budget the client has paid for. Non-billable time is everything else. In practice you need three buckets, not two:

• Billable: client work you can invoice or draw down from a paid budget.
• Client, not billable: work for a specific client that you will not charge for, such as fixing your own mistake, a warranty fix or an overrun on a fixed fee. It belongs to the client's profitability, so log it against them, but keep it out of invoices.
• Internal: work for the agency itself, such as sales, hiring, training and admin.

The middle bucket is the one most teams skip, and it is where margin disappears. Whether a grey-area activity is billable is decided by your contract, not by a rule of thumb. The table below gives a sensible default for 27 common agency activities, the condition that changes each one, and where the answer comes from.

The decision table

"Contract" means the statement of work, retainer agreement or terms you signed with that client. If the contract is silent, the default applies until you agree otherwise in writing.

Leave and public holidays are not work at all. Record them, but keep them out of all three buckets so they never sit in a utilisation sum.

How to use the table

• Decide once, then write it down. Arguments about whether a client call is billable usually mean nobody decided. Put your answers in the handbook (there is a paragraph to copy below).
• Estimate the way you log. If your estimates include review time and project management, log those as billable. If they don't, either add them to the estimate or log them as client, not billable, and see what it costs you.
• Log first, decide billing later. People should record what they did against the right client. Whether it ends up on an invoice is a decision for whoever owns the account, made with the contract open.

How the contract type changes the meaning

The same hour of work can be billable under one contract and not under another.

Time and materials

You invoice the hours. "Billable" means "goes on the invoice". The main risk is scope creep: work outside the agreed scope that someone does to be helpful. Log it against the client, mark it client, not billable until a change request is approved, and it becomes visible instead of lost.

Fixed fee

The client pays the agreed price whatever the hours. Hours inside the estimate are billable in the sense that the fee pays for them. Hours beyond the estimate are an overrun: still logged against the client, but not billable, because the client will not pay more. Tracking the overrun separately is how you find out which kinds of job you under-quote.

Retainer

With an hours-based retainer, billable time draws down the monthly allowance, so the contract must say what counts. Most disputes are about account management, meetings and reporting. Spell those out in the retainer agreement. With a deliverables retainer, hours are not invoiced at all, but tracking them still tells you whether the fee covers the work. Our retainers guide covers rollover and burn reporting.

Care plans and support agreements

Time on a client site's incidents, updates and small changes usually draws down the plan's hours or falls under its fixed fee. Incidents your own deployment caused are the classic grey area. Decide in advance, and say so in the plan.

Setting up tracking

A structure that works for most agencies:

1. Client: who the work is for, or your own agency for internal time.
2. Project: one per statement of work, retainer or care plan, so each has one contract behind it.
3. Task or ticket: what was done, where your project tool already has it.
4. Billable flag: set by default on the project, so people do not decide it entry by entry.
5. A small set of tags for the grey areas: change-request, warranty, rework-ours, overrun, account-management, pitch.

Write-off reason codes

When you decide not to invoice client time, record why. After a quarter, the totals by reason tell you what to fix.

Billing rounding is not payroll rounding

Whether you bill an hour and whether you must pay for it are separate questions. Classifying time as non-billable never makes it unpaid work.

• US: for non-exempt staff, "work not requested but suffered or permitted to be performed is work time that must be paid for by the employer". Rest breaks of 20 minutes or less count as hours worked, and travel from site to site during the workday is work time. Training and meetings count unless four conditions are all met: outside normal hours, voluntary, not job related, and no other work done at the same time [2]. Rounding to the nearest 5 minutes or a tenth or quarter of an hour is accepted for enforcement purposes only if, over time, it does not fail to pay employees for all the time they actually worked [1]. Your billing can round however your contract allows; payroll rounding has to pass that test on its own.
• Australia: pay follows hours worked and the award, registered agreement or contract, not what you bill. The National Minimum Wage, for example, is an hourly rate: $26.44 an hour from 1 July 2026 [3]. An award or agreement may set higher minimums, penalty rates and overtime.
• UK: for salaried staff, working time for minimum wage purposes includes "time spent training" and "time spent travelling on business" [4]. A pitch week of unbilled long days still counts towards the minimum wage check.

So keep one record of hours worked for payroll and record-keeping, and let the billable flag sit on top of it. Never edit a timesheet's hours to match an invoice.

Turning the data into decisions

Once time is in three buckets, four numbers become easy to read each month:

• Utilisation: billable hours ÷ available hours. See our utilisation guide for how to set a target.
• Client time that is not billable, as a share of all client time, by client. A client at 25% is costing you a quarter of the work you do for them.
• Realisation: what you collected ÷ billable hours at your standard rate.
• Write-offs by reason code, by quarter. One code usually dominates, and it tells you where to start.

Worked example

An account logs 120 hours in a month: 96 billable, 18 client but not billable (10 hours of SCOPE, 8 of REWORK) and 6 internal hours on the team's own pitch for a new piece of work. The standard rate is $150 an hour, in any currency.

• Client time = 96 + 18 = 114 hours. Client time that is not billable = 18 ÷ 114 = 15.8%.
• The client is invoiced for 96 hours: $14,400. Value of all client time at the standard rate: 114 × $150 = $17,100. You collected 84.2% of it.
• The 10 hours of scope work ($1,500 at the standard rate) are the cheapest to fix: a change request next time turns them into revenue.

Five arguments you will have, and how to settle them

1. "Is all that chat with the client billable?" On time and materials or an hours-based retainer, yes, if it is about the work. Log it in sensible blocks against the client's project, not minute by minute. On a fixed fee it is part of the price.
2. "The job took twice the estimate. Do we bill it?" On time and materials, the hours are billable, but tell the client when you see the overrun coming, not on the invoice. On a fixed fee, the extra hours are client time that is not billable; tag them OVERRUN and use them to fix the next estimate.
3. "The client changed their mind after sign-off." That is a change request, and it is billable once approved. If you absorb it to keep the peace, log it with GOODWILL so the cost is visible.
4. "Do we bill for fixing the bug we introduced?" No. Rework caused by your own mistake, and defects inside the warranty period, are client time you do not charge for.
5. "A junior spent six hours learning the client's platform." Bill the time a competent person would have taken for the task, and log the rest as internal training. Tag it LEARNING if it went against the client first, so the write-off has a reason.

Policy snippet for your handbook

Copy and adapt this. Replace the bracketed parts.