Skip to main content
DeepGrade
Note31 August 20266 min

When a spreadsheet stops being enough.

Five signs, and none of them is that the spreadsheet looks unprofessional. Two people need it at once. The same number is typed more than once. There is a rounding error nobody can find. Nobody can say quickly where a lot went. And only one person dares change it.

We build software for a living and we will still say this plainly: a spreadsheet is a genuinely excellent tool, and most operations should keep using one for longer than a software vendor will tell them.

Why the sheet usually wins

It is free, it is already there, everybody in the building can read one, and it fits the operation exactly because the operation built it. That is a real product with a real advantage, and any piece of software that replaces it has to be better than that rather than merely newer.

The pitch you will hear from most vendors is that spreadsheets are unprofessional or error prone. That is not an argument, it is a mood. The honest version is narrower: a spreadsheet has specific structural limits, and it is worth knowing which one you have actually hit before you spend money.

  1. 01

    Two people need it at the same time

    The wharf wants to add a landing while the office is reconciling last week. One of them waits, or somebody emails a copy and now there are two versions and a conversation about which is right.

    This is the cheapest one to fix and it is often not software: a shared sheet in a proper cloud spreadsheet with real concurrent editing solves it for a lot of operations, permanently.

  2. 02

    The same number gets typed more than once

    A weight written on the wharf, keyed into the sheet at nine at night, typed again into the purchase slip, and once more into the settlement. Four copies of one fact and three chances to differ, with no way afterwards to know which was the original.

    This is the real signal, and it is the one that justifies a build. Not because typing is slow but because a number that exists in four places has no owner. Capture it once where it happens and let everything downstream read it.

  3. 03

    There is a rounding error you cannot find

    The shares add up to a cent less than the pool. It has been happening for a while. Nobody can find it because each share was rounded on its own, which is exactly what a spreadsheet does and exactly what nobody notices.

    A spreadsheet divides in floating point and rounds for display, and those two facts together make a penny go missing. Money math needs to work in whole cents and allocate the remainder deliberately. That is a hundred lines of tested code, and it is not something to build inside a sheet.

  4. 04

    You cannot answer where a lot went in under an hour

    Somebody asks which customers got product from a particular landing. The answer exists, across a receiving sheet, a production tab, a shipping log and somebody's memory, and finding it is an afternoon.

    Worth being honest about urgency here. If this has never actually come up, it is not yet a reason to buy anything. If it has come up twice, or a buyer has started asking, the answer needs to live in the records rather than in the afternoon.

  5. 05

    Only one person can change it

    The sheet has grown a structure that works and nobody else dares touch it. When that person is on the water, or leaves, the operation has a single point of failure it did not choose and cannot see.

    This one is about risk rather than efficiency, and it is the one most likely to be dismissed until the week it matters. At minimum, write down how it works. Better, move the parts that other people need into something they can use without fear.

One sign that is not on the list

Volume. People assume the trigger is getting bigger, and it usually is not. Plenty of operations run a great deal of fish through a sheet perfectly well. What breaks a spreadsheet is not the number of rows, it is the number of hands and the number of places the same fact has to live.

What to do if only one of them is true

Fix that one, and keep the sheet. Move to a properly shared cloud sheet. Write down how it works so it is not one person’s. Take the money math out of the sheet and into something that gets it right, and leave the rest where it is. A build that replaces everything at once is more expensive and more disruptive than one that replaces the part that is actually broken, and a studio that will not tell you that is not worth hiring.

When three or four of them are true at once, that is the point where the sheet is holding the operation together rather than serving it, and where a fitted build stops being a luxury.

If the rounding error is the one that got you, the crew share calculator is free and does that part properly. If it is the typing, that is Dockside.