We build software for small businesses. So it is a strange thing to say, but here it is: quite often the right answer is to build nothing.
A business comes to us because a process is painful. Quotes take too long to go out. The team cannot see which jobs are stuck. Someone spends every Monday morning rebuilding the same report by hand. The instinct, and often the request, is for an app that fixes it.
Sometimes an app is the right call. But the base rates are not encouraging. BCG's 2020 study of 825 executives found that around 70 percent of technology transformations fall short of their objectives, and that is leaders grading their own projects, not an independent audit. The Standish Group, which has tracked IT project outcomes since the 1990s, finds the same thing every year: small, contained projects succeed far more often than large ones. Its method is argued over, but the size effect is not.
So before we can know whether to build, we look at how the work happens now, step by step. And when we do, the cause is frequently something else.
It is a setting. The business is paying for Microsoft 365, or a tool it already owns, and using a fraction of what it covers. Across the companies in Zylo's SaaS Management Index, about half of paid software licenses go unused. The feature someone wants built is often already there, switched off, because nobody was shown it.
It is a step that should not exist. A form gets printed, signed, scanned, and filed. The printing and scanning are there because that is how it was done fifteen years ago, not because anyone needs the paper.
It is a sequence. Two teams are each waiting for the other to go first, and a five-minute conversation about who owns what removes a delay everyone had assumed was a software problem.
Laying new software over a process nobody has examined does not fix the process. It makes the same process run faster, which means the same confusion, the same duplicated work, and the same bottleneck, now with a monthly license attached and a thing to maintain.
So we say what we find. If the answer is a setting you already pay for, you hear that. If it is a step you can delete, you hear that. Occasionally this costs us the work, because the honest recommendation is small and does not need us. We think that is the right trade. A client who gets a real fix for very little money is a client who calls back when there is something genuinely worth building.
None of this means we talk businesses out of building things. When a process needs an app, it needs one, and half-measures waste more time than they save. It means the recommendation comes after the diagnosis, not before it. And when we do build, it is in small stages, for the same reason the research gives: the smaller the thing, the better the odds it works.
The order matters more than the tool. Understand the work first. Then decide what, if anything, to build.