Leaving Excel behind: when a custom internal tool pays for itself

Many Swiss companies run an essential part of their operations on shared Excel files. Site planning, stock tracking, customer databases, invoicing. And it works: which is worth saying, because the engineer's reflex is to sneer at spreadsheets when they in fact solved a real problem, fast and with no budget.

So the question is not whether Excel is good or bad. It is spotting the moment it becomes more expensive than the alternative.

The six signals

If you tick three, the arithmetic is worth doing:

1. The file has a keeper. One person is the only one who knows how it works, and everyone waits for them to come back from holiday.
2. Parallel versions circulate. planning_v4_final_OK_edited.xlsx by email.
3. You re-key data. The same value is typed twice, into the spreadsheet and into the invoicing software.
4. Nobody knows who changed what. A row changed; there is no history.
5. Mistakes cost money. A broken formula has already caused a wrong order or a wrong invoice.
6. You cannot reach it from the field. Your teams take notes on paper, then somebody types them up in the evening.

The arithmetic, simply

Take the time actually spent each week on this handling: re-keying, correcting, hunting for the right version, asking the file's keeper.

Five hours a week at CHF 80 fully loaded is roughly CHF 20,000 a year.

A targeted internal tool (one that replaces one process, not ten) generally lands between CHF 15,000 and 40,000. The payback is measured in months, not years. And that calculation still ignores the cost of errors, which is the most painful and the hardest to quantify.

The classic mistake: replacing everything

The project that fails is almost always the one aiming to replace the whole information system at once. Eighteen months of development, a budget that doubles, and a go-live nobody dares trigger.

The approach that works is the opposite:

• One process first: the one that hurts most, and only that one.
• In production in eight to twelve weeks, with real users.
• The other files stay in place meanwhile. You migrate them afterwards, one at a time.
• Excel export stays available. Your teams need it for ad-hoc analysis, and taking it away creates pointless resistance.

What to insist on

A badly built internal tool is worse than a spreadsheet, because you can no longer work around it. Three safeguards:

• Data stays exportable. No exceptions. That is your insurance against lock-in.
• The tool connects to what exists. If your accounting software has an API, use it; re-keying must disappear, not move.
• End users take part in scoping. The people filling the spreadsheet today know the edge cases nobody else suspects. Without them, the tool will be right on paper and wrong in reality.

The right moment

It does not arrive when the company grows. It arrives when the same person performs, for the hundredth time, an operation a machine would do better, and nobody finds that surprising any more.