Write-Offs and What They Tell You

The gap between hours recorded and hours billed is the most informative number in a professional services practice and the most commonly misread.

Where write-offs come from

Fixed fee exceeded. The work took longer than the price assumed. A pricing finding.

Scope not agreed. Work done that the client did not authorise. A scoping finding.

Rework. Something done twice. A quality or brief finding.

Junior time on work that should have been done once, properly. A staffing finding.

Client relationship management, where partners write down to preserve goodwill. A commercial decision, not a performance one.

Genuine inefficiency. Present, and the smallest category in most analyses.

Six causes, five of which are about how work was set up rather than how it was done.

Why misreading it is expensive

Attributing write-offs to the person who recorded the time punishes whoever was handed the badly scoped matter.

It teaches people not to record, which removes the evidence of the underlying problem.

And the underlying problem — the price, the scope, the brief — continues, because the signal that would have surfaced it has been suppressed.

A practice with low recorded write-offs and recurring margin problems has a recording problem, not a delivery one.

Analysing them properly

By matter type, which finds the work that is systematically underpriced.

By client, which finds the relationships where scope creeps.

By phase, which finds where estimates break down.

By cause, which requires a reason code that people can apply consistently and that is short enough to be used.

Never by individual as a headline. Where an individual pattern genuinely exists it will show up in the matter-level analysis with that person's name attached to the matters, and it is a conversation rather than a metric.

The reason code

Keep it to five or six options or the first one gets chosen every time.

Pricing, scope, rework, staffing, commercial, other.

Review the "other" rate. A dominant "other" means the codes do not fit the reality and should be rewritten with the people who select them.

Review annually, because the mix changes.

Using the finding

Reprice the work that is systematically written off. This is the whole point and it is frequently deferred because it means a conversation with a client.

Fix the scoping process where scope is the dominant cause.

Feed it back into estimates, which has its own note.

Report the change, because a practice that surfaces an underpricing finding and does nothing teaches everyone that the analysis is decorative.

What to report

Realisation by matter type, trended.

Write-off by cause.

Estimate against actual for completed matters.

Not: individual realisation percentages, which measure the matters someone was given.

The reason code that works

Six options, short enough to be used and distinct enough to be informative.

Pricing: the fee assumed less work than it took.

Scope: work done that was not authorised.

Rework: something done twice.

Staffing: the wrong person did it.

Commercial: written down deliberately for the relationship.

Other.

Watch the "other" rate. A dominant catch-all means the codes do not fit, and they should be rewritten with the people who select them.

Acting on the finding

The analysis is worthless without the uncomfortable conversation that follows.

Reprice the work that is systematically written off, which means a client discussion.

Fix the scoping process where scope is the dominant cause.

Feed the pattern into estimates.

Report the change, because a practice that surfaces an underpricing finding and does nothing teaches everyone that the analysis is decorative.

Where the decision is to keep the price and accept the write-off, say so explicitly as a commercial choice rather than leaving it as an unexplained margin problem.

Connect policy to configuration

The practical choices behind this note can be compared with the service-desk workflow. Keep the organisation's written purpose in control of the setup, enable only the data needed for that purpose and review the result with affected users.