Three vendors, three opposite answers
The vendor says their product covers everything, the agency says you need custom, the integrator proposes heavy configuration. Each honestly describes what they sell.
By situation
In most cases you shouldn't build custom software. Here's how to know if you're in the minority.
The decision rests on four questions: is your process a competitive advantage or mere necessity, does an existing product cover it 80% after configuration, does the deciding data live elsewhere, and is the three-year calculation favourable. One "yes" isn't enough. When in doubt, buy — buying wrongly can be corrected, building wrongly is paid for over years.
What brought you here
You've already paid once for needless custom work
A system built to reproduce what a standard subscription product already did. The resulting distrust is justified, and deserves better than a pitch.
The real question isn't being asked
It isn't "which software", it's "which part of how we work deserves protecting". A company that hasn't answered that will buy badly, whichever route it takes.
What we actually do
The 80% rule, applied honestly
If an off-the-shelf product covers 80% of the need after configuration, buy it and adapt your 20% — unless that 20% is precisely what sets you apart from competitors. The whole decision lives in that exception.
The calculation runs three years, not on purchase price
Per-seat licences times projected headcount growth, against a one-time build cost plus maintenance. Custom often loses over twelve months and sometimes wins over thirty-six.
The hybrid is almost always the right answer
Keep accounting and payroll off the shelf, build what carries your difference, and make the two talk. It's what we recommend most often, and it's also the smaller engagement.
When this isn't for you
This page sells nothing and shouldn't try: if its framework leads you to buy off the shelf, it did its job. We maintain it because an owner who makes the right decision against us this year comes back the time it's the right decision with us.
No sales page writes this section. That's precisely why it's here: it saves you a thirty-minute call, and saves us an engagement we'd decline anyway.
Questions we get on this
Is custom riskier?
Yes, on two specific points: scope creep and dependence on the builder. Both are handled by contract rather than technology — closed scope and fixed price for the first, code ownership and handed-over documentation for the second. A vendor refusing both clauses is telling you something important.
What if we get it wrong?
Getting it wrong by buying costs a cancelled subscription and a migration. Getting it wrong by building costs the development, plus the years spent maintaining it. That asymmetry is what justifies the rule: when in doubt, buy.
Neighbouring situations
Fifteen minutes to find out whether we can be useful.
You leave with what we saw, in writing, even if we never work together.