When we tell clients not to build the software
A discovery engagement that ends in "do not build this" is a success. Three cases where the honest recommendation cost us the project.
Consultancy has an obvious conflict of interest: the firm advising you on whether to build software is usually the firm that would like to build it.
We handle that by separating the two commercially — discovery is a fixed-price engagement with its own deliverable — and by being willing to reach the unprofitable conclusion. Three examples.
The ERP that was a training problem
A distributor wanted to replace an ERP their staff described as unusable. Two weeks of watching people work made the actual problem clear: the system did almost everything they needed, but three critical workflows had been configured incorrectly at implementation four years earlier, and nobody had been trained on the rest.
The recommendation was reconfiguration and a training programme, at roughly eight per cent of the replacement budget. We did not get a build project. We did get the reconfiguration work, and a client who now brings us everything.
The marketplace with no supply
A startup asked us to scope a two-sided marketplace. The demand side was well researched. The supply side rested on an assumption that a particular category of vendor would sign up because the economics were attractive.
We suggested spending four weeks and a small fraction of the build budget testing that assumption with a landing page and fifty phone calls before writing any code. Eleven of fifty vendors were interested, and their reasons for hesitating were structural rather than fixable by software.
The founders pivoted. That was the right outcome, and it was not a project for us.
The automation that would have automated waste
A manufacturer wanted to automate a reporting process consuming about thirty hours a week. Mapping it revealed that two of the four reports produced were not read by anyone — they had been requested by a manager who had since left.
We recommended killing the two reports and automating the remaining two. The engagement shrank by roughly half.
Why this is not altruism
Every one of these clients came back. Two of them have since referred others. The arithmetic works out in the long run, and it is a great deal easier than maintaining the alternative position.
The test we apply internally: would we still recommend this if we were not the ones being paid to build it?
