Healthy Write-off Conversations using Xero and Link

Most accountants were never told how a write-off happens

You go to university, you work at a firm, you move to another firm. Along the way nobody explains what WIP is, how a write-off is calculated, or how the number ends up next to your name. All you know is that write-offs are bad and you don't want one. That gap is why these conversations go badly. Correcting it is the fastest thing you can do, and the correction is simple: the #1 cause of write-offs is pricing.

Write-offs are a management accounting tool

They're not good and they're not bad. Any firm's write-offs could disappear tomorrow by resetting everyone's billable rate to their actual recovered rate. You'd make less money, because every hour of hourly work would now be sold at $220 instead of $250. Good firms aim for a small negative number across most of the team, which keeps rates meaningful and gives people something real to deliver against. Write-offs are also asymmetric. Things go wildly wrong more often and by more than they go wildly right, so expect more write-offs than write-ups. And they aren't profits or losses, because the WIP equation runs on billable rates, not cost rates.

Pricing, then process, then people

Pricing comes first because the moment you know least about a job is the moment you price it, and if the engagement hasn't been repriced in three years that number is the most stagnant thing about the work. Process is second: the sequence, the tools, the complexity that turned up after the price was set. People is last, and the maths says so. A full-time accountant does around 1,500 hours of client work a year. Four years in, that's 6,000 hours, and learning is logarithmic, so most of the curve is behind them by year three. If someone has been doing this for eight years, it probably isn't a them problem.

Six questions that change the conversation

"You wrote off $2,800 last month, sharpen up" tells someone nothing except that they don't want the conversation again. Try these instead. Was there anything unusual about this job? Was it more complex than we anticipated? Should we be charging more for this work? Should we have charged separately for part of it? Is this likely to take 12 hours again next year? Could we have delivered it differently? All open, all inquisitive, and all of them produce something you can act on.

One good conversation per person per month

Sort your team by net write-ups, drill into the biggest number, and find the one job. It's remarkable how a single job can wipe out a month of work that went to plan. That's your conversation. Have it in person or on a call, not over email, and within a month of the write-off while the job is fresh. Give the job to one owner. If two people are meant to be having these conversations, nobody is. And don't spend your time apportioning write-offs to individuals. Use the default apportionment and treat the result as management information, because picking who wears a write-off creates a different problem entirely.

Then move the conversation earlier

Everything above is retrospective, which is the easiest place to start and the wrong place to finish. The same questions work better while the job is still open. Put a budget on every deliverable, use last year's actual hours as the starting number, and watch remaining time in the workflow. When someone is nine hours into a ten-hour job and still on work papers, that's the moment to ask what changed and whether it needs repricing.

Every answer your team gives you is a "because" for next year's pricing conversation. That's what these questions are really for.