Is Your Business Ready for Custom Software?

Readiness is not about technology

Clients ask whether they are “technical enough” for custom software. Almost nobody fails on that. The projects that struggle fail on organisational readiness: an unstable process, an absent owner, or no agreement about what success looks like.

The eight questions below are the ones we ask ourselves before quoting. You can run them on your own business in twenty minutes.

The eight questions

1. Is the process stable? Has it worked roughly the same way for six months? Automating a process still in flux means rebuilding it.
2. Can you describe it end to end? If nobody can walk through it in one sitting, it is not one process — it is several, and they need separating first.
3. Does someone own the outcome? A named person whose work improves when this ships.
4. Does the data exist anywhere? Messy is fine. Absent means phase one is capture, not automation.
5. Can you name the number that should move? Hours, error rate, cycle time, conversion. One number.
6. Will the people who do the work be involved? Not just consulted at the end.
7. Is there budget for year two? Software has running costs; a build with no maintenance budget decays.
8. Is the deadline real or aspirational? Both are workable; confusing them is not.

Scoring it honestly

Six or more yes answers means you are ready and a build will very likely go well. Four or five means ready for a small, contained first project rather than a platform. Three or fewer means the next step is not software.

We have never seen a project fail for being too technically ambitious. We have seen several fail because nobody could say what number was supposed to move.

If the process is not stable

Then stabilise it manually first. Write the steps down, agree them with everyone who does the work, and run it that way for a month. It sounds slow and it takes about two weeks of real effort.

The payoff is large: you find out what the process actually is, and you get a specification for free. Half the value of a good discovery is this exercise, and you can do it without paying anyone.

If nobody owns the outcome

Find the person whose job gets better and make them the sponsor, even if they are junior. Sponsorship by a senior person who is not affected produces a project that is approved but not adopted.

If nobody's job gets better, that is genuinely useful information. It usually means the problem is smaller than the proposed solution, or that it belongs to a different department than the one commissioning the work.

If you cannot name the number

Pick one, even roughly. “Order processing takes about 90 minutes a day and we want it under 20” is enough. It does not need to be instrumented today; it needs to be agreed today.

• Hours a week spent on the manual step
• Error or rework rate, even estimated
• Cycle time from request to completion
• Enquiries handled per person per day
• Percentage of orders needing manual intervention

Without a number, nobody can say whether the project worked, and a project that cannot be shown to have worked does not get a phase two.

Frequently asked questions