Custom software can be a genuine advantage, and it is also one of the easiest things to over-buy. Before you commission a bespoke build, it is worth understanding the odds, and applying a short, honest test for whether you actually need one.
The honest odds of a custom build
Building software is hard, and the data is blunt about it. Across decades of Standish Group CHAOS research, only about 31% of software projects fully succeed; 19% fail outright and the rest are “challenged”, late, over budget, or short on scope. And of the features that do get built, around 64% are rarely or never used.
It gets starker at scale. A study by McKinsey with the University of Oxford found that large IT projects, on average, run 45% over budget and deliver 56% less value than predicted, and 17% go so badly they threaten the company’s existence. None of this means do not build. It means build deliberately, and only when the problem justifies it.
Build versus buy
Most needs are met by configuring existing tools and connecting them well. Custom software earns its place where off-the-shelf genuinely cannot fit how you work, or where the way you work is itself the advantage.
Often the real fix is the workflow, not the software, which is why it pays to check the process before you code.
| Lean towards off-the-shelf when… | Lean towards a custom build when… |
|---|---|
| A common tool already does 80% of the job | Your process is your edge, and no tool fits it |
| The process is standard (accounting, CRM, email) | The workflow is unique or a genuine differentiator |
| You need it working in weeks, cheaply | Integration and data are the hard part, not the screens |
| Requirements will keep changing | You have a clear, repeated, high-value problem to solve |
If you do build, de-risk it
When a bespoke build is the right call, the failure modes above are avoidable. We scope and build software by proving the riskiest part first with a prototype, then delivering in priced phases with clear milestones, in-house or with a trusted partner. You should never be asked to fund the whole thing on faith, and you should own everything at the end.
The five-minute test
Before you commission anything, answer these honestly:
- Can you describe the specific problem, and what it costs you each month, in one sentence?
- Have you tried to solve it by combining existing tools plus a little automation?
- Is this workflow a genuine differentiator, or just admin everyone has?
- Could you prove the riskiest assumption with a small prototype first?
- Do you know what “good” looks like, and how you will measure it?
If you are answering “not sure” to most of these, a short scoping conversation will usually save you far more than it costs.
Thinking about a build? We help you decide honestly, then build it properly if it is the right move. See how we run and build software, or scope a build with us.