We are a software development firm, so treat what follows with appropriate suspicion. But a meaningful share of the projects we are asked to quote for should not be built at all, and saying so early has cost us less business than you would expect.
The question that decides it
Not “can we build this?” — you almost always can. The question is:
Is the way we do this differently from our competitors the reason customers choose us?
If yes, build. That process is your advantage, and forcing it into a product designed around somebody else’s assumptions will slowly erode the thing that makes you worth choosing. If no — if it is payroll, or accounting, or email, or helpdesk ticketing — buy it. There is no prize for a bespoke expense-claim form.
When buying is clearly right
- The problem is universal. Thousands of companies have it in nearly identical form. Someone has already solved it more thoroughly than your budget allows.
- Compliance is the hard part. Payroll and tax software carries the burden of keeping current with regulation. That is a subscription worth paying.
- You need it this quarter. Custom software has a lead time. Sometimes the honest answer is to buy something adequate now and revisit in a year.
- Nobody internally will own it. Software needs a custodian. If no one has the time or authority to own the thing after launch, it will decay regardless of how well it was built.
When building is clearly right
- The workflow is genuinely yours. You have tried three products and each required you to work in a way that made you slower. That is a real signal, not stubbornness.
- Integration is the actual job. You need four existing systems to agree. No vendor sells the seam between them; that seam is where the value is.
- The per-seat maths has turned. Licence costs scale with headcount while a built system does not. At a certain size the arithmetic simply flips.
- The data is the asset. When the information you accumulate is itself the competitive advantage, keeping it inside a vendor’s schema is a strategic risk, not just a technical one.
The option people forget
It is rarely a binary. The most cost-effective answer is frequently to buy the commodity parts and build only the seam: an off-the-shelf accounting package, an off-the-shelf CRM, and a modest custom layer that makes them behave as one system and produces the reporting neither can produce alone. That layer is often ten percent of the cost of a full custom build and delivers most of the benefit.
What to ask a vendor
- “What should we not build?” A partner with judgement will have an answer ready. One who says everything is worth building is quoting, not advising.
- “What happens if we leave?” Ask who owns the code, where it is hosted, and how the handover works. Hesitation here is the most informative response you will get all meeting.
- “What will this cost to run in year three?” The build price is the smaller number. Hosting, maintenance and change requests are the real total.
Custom software is a good investment when it is aimed at the part of your business that is actually distinctive. Aimed anywhere else, it is an expensive way to recreate something you could have bought on a card.