Spreadsheets are not the problem — they are being used past their design
Most companies' first "system" is a spreadsheet, and it lasts longer than people expect. A well-built workbook plus one or two experienced hands can carry a business a long way.
So the question is not whether spreadsheets are bad. It is when they turn from an asset into a risk. The more of the following signals you see, the more urgent the change.
Signal 1: only one person understands the file
Nested formulas, hidden columns, links across workbooks — and exactly one person knows what must not be touched.
That is an operational risk, not an efficiency problem. If that person is away, leaves, or simply sorts a range wrongly once, you will make decisions on bad numbers and not notice immediately.
A test: send that person on a two-week holiday and see whether month-end still closes.
Signal 2: the same number has more than one version
Sales says stock is 50, the warehouse says 42, accounting has 47. Nobody is lying — they are reading three copies from three moments in time.
When "how much do we actually have" requires a meeting, the data is no longer one record.
Signal 3: month-end close takes more than three days
A close should be reconciliation, not reconstruction. If every month means re-sorting transactions and patching gaps, the data was not recorded correctly when it was created.
Signal 4: you cannot calculate per-item margin
Ask what you earn on one unit and the answer is "let me work it out" — and it takes three days.
Cost is usually scattered: one price on the purchase order, freight on another document, channel commission on a third report. Because combining them is manual, nobody does it often, and loss-making items keep selling.
Signal 5: more channels means more people
Each new sales channel adds a headcount's worth of work: listing, stock reconciliation, settlement.
That is a linear cost — double the channels, double the work. Flattening that line is exactly what a system is for. If your first thought about opening another channel is "we will need to hire someone", this signal is already clear.
Signal 6: people bypass the process to keep up
Shipping is urgent, so goods go out first and paperwork follows. The count does not match, so it is overwritten with the physical number and the cause is never investigated.
Once bypassing becomes normal, the data is permanently wrong and you lose traceability — later you will not be able to answer "where did this batch go".
How many signals mean it is time
- 0–1: no rush. Keep the spreadsheet tidy, backed up and version-controlled.
- 2–3: start evaluating, but fix the most painful one first — usually stock or channel sync. A full ERP is not required yet.
- 4 or more: you are already paying a hidden cost in overtime, rework and decisions made on wrong numbers. That cost is usually higher than the system.
One thing to settle before you switch
A system will not improve your process by itself. It will freeze your current process in place — including the parts that do not make sense.
So before choosing a system, write down how things are *actually* done today, not how they should be done. That document decides how smoothly the rollout goes, and it is the first thing we ask for when scoping a project.