The standard way to buy business software is to buy the platform, then spend six months rolling it out. Everybody has watched this go wrong at least once. What is less obvious is that it goes wrong in a specific and predictable way, and that the fix is structural rather than a matter of planning harder.
How platform rollouts actually fail
They rarely fail technically. They fail because the decision to buy was made before anyone knew which parts would be useful, and by the time that becomes clear, too much has been spent to admit it. The project acquires a defender. The defender is not lying — they genuinely cannot see a version of the story where stopping is acceptable.
That's a design flaw in the purchase, not a failure of the people. Any commitment large enough to need defending will get defended.
What changes when the unit is one job
Make each step small — one repetitive job, one fixed price, three days to first look — and several things happen at once:
- Stopping becomes normal. Nobody has to defend a decision that cost three days.
- The second decision is better informed than the first. You have watched the first one run in your own office.
- Nothing waits for everything. The invoice checks don't sit idle while the meeting notes get argued about.
- One manager can approve it. No steering committee means no six-week gap between deciding and starting.
The obvious objection
Doesn't this produce a mess — a pile of disconnected little tools? It would, if each job were built on something different. It doesn't, because they all sit on the same base: the same account, the same connections to your systems, the same rule that a person approves before anything is final. The second job is faster than the first for exactly that reason.
What we deliberately don't do is promise you where it ends up. We don't know, and neither does anyone else honestly. What we can say is what the next job costs and what it will do.
The uncomfortable part, for us
This is a worse business model than selling platforms. The contracts are smaller, every one has to earn the next, and a customer can stop after the first three days having paid us very little. We build this way anyway, because the alternative is asking a company to bet on something neither of us has seen working yet.
Pick the job that wastes the most hours. Start there. Decide about the second one later.
Start with one job
A free one-hour walkthrough at your office picks the job. Three days later you see it running on your own files.
Start with a walkthrough →