A spreadsheet is fast, flexible and free, and it's usually the correct place to start. The problem is that it never announces the moment it stops being suitable. It just gets slower, more fragile, and more dependent on one person who knows which cells not to touch.
None of the signs below mean "replace everything". Most are fixed by a small tool that does one job well and leaves the rest of your setup alone.
1. Someone re-types the same data every week
A supplier statement, a bank export, a timesheet, an order list — pasted in, cleaned up, reformatted, then keyed into something else. This is the clearest signal of all, because it's pure cost with no judgement involved.
Smallest fix: an import-and-transform tool. You drop the file in, it applies your rules, and it outputs clean data — or writes it straight into your accounting or job system. Automations like this are typically the cheapest work we do and the fastest to pay back.
2. Version confusion has become normal
Files called rota_v4_FINAL_JAN_updated.xlsx. Two people editing two copies. Someone working from last week's numbers in a meeting. If your team has developed etiquette rules about who may open a file, the file is now a shared database being used without any of the protections a database provides.
Smallest fix: move that one process into something with a single live record and proper access — everyone sees the same state, and history is kept. Our AdminRota project started exactly here: a rota that several people needed at once, in a spreadsheet that only tolerated one.
3. One person is the system
The formulas make sense to exactly one member of staff. When they're off, the process stops or degrades. This is a business-continuity risk, not an inconvenience — and it also traps a capable person in maintenance work instead of the job you hired them for.
Smallest fix: encode the rules where they can be read and tested rather than reverse-engineered from nested IF statements. The measure of success is that a new starter can run it on day two.
4. You can't answer a simple question quickly
"How many live jobs are we running?" "What's outstanding this month?" "Which client is most profitable?" If answering means half an hour of cross-referencing tabs, you don't have a reporting problem — you have a data-structure problem. Decisions get made on feel because the facts are too expensive to retrieve.
Smallest fix: a single small dashboard over the data you already keep. Often no process change at all, just a reliable read.
5. Errors are being caught by customers
A wrong figure on an invoice, a missed booking, a duplicate order. Spreadsheets will cheerfully accept a date typed into a price column, a broken reference, or a formula dragged one row short. Nothing validates anything.
Smallest fix: validation at the point of entry. A form that refuses impossible input prevents the whole downstream cleanup. Reconciliation tools that flag mismatches automatically are the same idea applied after the fact.
6. The process has quietly grown a second job
It started as a list. Now it also tracks approvals, holds notes, calculates commission and feeds three other files. Each addition was sensible; the total is a system nobody designed. Fragility rises faster than the feature count.
Smallest fix: separate the one part everyone depends on daily, and give that part real structure. Leave the rest in the spreadsheet — ad-hoc analysis is what spreadsheets are genuinely best at.
7. It doesn't work where the work happens
Your team is on site, in a van, or in a client's office, and the file lives on an office PC. So updates happen "later", which means information is always a few hours out of date, which means someone rings someone to check.
Smallest fix: a small private web app — one screen, works on a phone, updates the same live record everyone else sees.
Don't rip out the spreadsheet. The best result is usually a hybrid: structured software for the process everyone depends on, and spreadsheets kept for the modelling and one-off analysis they're excellent at. We've never had a client regret keeping Excel for the thinking.
How to work out whether it's worth fixing
Pick the single most annoying process and estimate three numbers: hours per week spent on it, how often it goes wrong, and what going wrong costs. Six hours a week at a £20 fully-loaded rate is around £6,000 a year — that's your budget, and it recurs whether or not you spend it.
If you'd like to sanity-check the maths before committing to anything, our guide on bespoke software vs off-the-shelf has a five-minute decision test.
Tell us about the spreadsheet that's slowing you down.
Send over what the process does and where it hurts. You'll get a straight answer on whether it's worth automating and a fixed price if it is.