What is that spreadsheet actually doing?
Every firm of this size has at least one. It might be the sheet that prices a proposal before it goes out, the resource planner that says who is free next month, the WIP tracker partners glance at before deciding what to bill, or the model that turns raw hours into a monthly management pack. Ask around and you usually find three or four, not one.
Each started as a stopgap. Somebody needed an answer that week, opened a blank sheet and worked something out. It did the job, so it was used again. Then a column was added for a new service line, a tab for a second office, a lookup to pull last month's figures from another file. Years later it is load bearing. Nobody decided that. It accumulated.
Why does nobody treat it as a system?
Because it does not look like one. A system has a supplier, a renewal date, a line in the budget and someone whose job description mentions it. A spreadsheet has a filename. It sits in a folder, or on a laptop, and it costs nothing to keep, so it never reaches a management agenda.
Compare the treatment. The practice management system is backed up, patched, supported and, when it fails, escalated to somebody with an obligation to fix it. The sheet that decides what the firm charges gets none of that. No version history anyone reviews, no test that would catch a wrong answer, no written explanation of the logic, and often no second person who understands it. That is a risk position, and no partner ever consciously took it. It is the gap described in the process nobody owns, wrapped in a file that looks harmless.
What does it look like when it breaks?
Partners picture something dramatic: a corrupted file, a stolen laptop, a folder deleted during a tidy-up. Those happen, and IT usually recovers within the day. The failures that cost real money are quieter.
- A formula is typed over with a hard number, because someone wanted to force one answer, and it stays hard for two years.
- A row is inserted just outside the range a total refers to, so one job or one matter silently stops counting.
- A link points at a file that was renamed, and the sheet keeps showing the value it last cached.
- A column is sorted on its own, so names and figures no longer line up.
- Two copies are in circulation, each updated by a different team, and both look current.
What these share is that the sheet carries on producing numbers. It does not throw an error or refuse to open. It gives a plausible answer that happens to be wrong, and the firm acts on it: quotes a fee, staffs a job, raises a bill, reports to the board. Discovery comes weeks later, from a client query or a colleague who notices a figure looking odd.
What happens the morning somebody notices?
Three questions arrive at once. What is the correct number? When did it stop being correct? What else was produced from it in the meantime?
A managed system can answer those. A spreadsheet usually cannot. So the work begins: open old copies, compare them line by line, email colleagues to ask which version they used in March, rebuild the logic from memory, then decide whether anything has to be corrected with a client. That work lands on senior people at no notice.
There is a second problem where the sheet holds personal data, as rota, salary and contact sheets do. Copies saved to desktops, emailed as attachments and left in download folders raise a handling question under UK GDPR and the Data Protection Act 2018, and an unowned file is exactly the one nobody can account for.
Why isn't the backup the answer?
Backups protect the file, not the logic. Restoring last Tuesday's copy only helps if you know the error arrived on Wednesday, and the whole difficulty with a silent failure is that you do not. Restoring also throws away every legitimate change made since.
The deeper gap is knowledge. What makes a sheet fragile is that its reasoning lives in one head: why that rate is uplifted, why that category is excluded, why the third tab is never touched. None of that is backed up by anything. It is the problem set out in when knowledge leaves, concentrated into a single file.
Does this mean buying software?
Not necessarily, and reaching straight for a purchase is how firms end up with a licence nobody opens. Separate two kinds of sheet first.
The first does arithmetic. Somebody enters figures, the sheet adds them up, a person reads the output and judges it. Lost, it could be rewritten in an afternoon. That is a tool, and it is fine.
The second carries a process. It holds rules, exceptions and a sequence several people depend on, it is the only record of how the firm does something, and the steps around it are manual: copy from the case system, paste in, format, export, send. That is an undocumented system, and the answer is to rebuild the job rather than keep patching the file. Fee estimating, board and management reporting and timesheet capture are the three that most often sit here.
What should a firm do about it?
Four questions sort the sheets quickly. Who would fix this by Friday if it broke on Monday? Could a second person explain its rules without opening it? If it produced a wrong number, what would catch it? And how many people maintain their own variant, which is the warning sign in twelve ways of doing one job.
- Name an owner for each sheet that matters, in writing. Ownership is responsibility for whether the thing is still right, not a duty to update it.
- Write down what it does, in a page. The guide to documenting a process sets out how, and the exercise usually surfaces a rule nobody can justify.
- Add one check that would catch a wrong answer: a total that must reconcile, a sense check against another source, a second pair of eyes before the figure goes out.
- Take the one that would hurt most and rebuild that job properly, so the output comes out of the work rather than out of a file someone maintains in the evenings. Choosing which process to take first covers how to decide.
Firms rarely find this through analysis. They find it when the file fails in front of a client. The cheaper route is an hour spent naming the sheets that matter and deciding, deliberately this time, whether the firm is willing to run on them. The audit will point you at the process behind the worst of them.