When you should not buy software from us

Software encodes a way of working. If the way of working is still being invented, encoding it means rebuilding the encoding every few weeks, and the business ends up serving the system.

Run the process manually until it stops changing. Then automate the version that survived.

Every successful implementation has one person inside the business who cares whether it works. Not a champion in the motivational sense, just somebody whose job is easier when it is right and harder when it is not.

Without that person, adoption decays quietly. The system stays in use for billing and stops being used for everything else, and within a year the data is unreliable enough to be worthless.

A spreadsheet that one person maintains, that everyone understands, and that has not broken is a legitimate system. It fails at multi-user editing, at history, at scale and when its author leaves, and if none of those are pressing then there is nothing to fix.

Replacing a working spreadsheet with software costs real money and real disruption to buy capabilities the business is not yet using.

This is the food stall case. If entering a sale takes longer than serving one, and the daily close takes longer than counting the cash box, the arithmetic does not work and no feature list changes it.

There is a real threshold below which a business is better off without this, and pretending otherwise sells something that will be abandoned in a fortnight.

Some difficulties look like data problems and are not. Staff not following an agreed process, a supplier who is consistently unreliable, a pricing decision nobody wants to make. Software will document these accurately and change none of them.

It is worth being clear about which kind of problem is being solved before spending anything, because a system bought to fix a management issue produces a very well documented management issue.