Most businesses run their finances on off-the-shelf software, and for good reason. Tally, Zoho, QuickBooks, and Xero solve the vast majority of problems for the vast majority of companies. If your needs are standard, you should buy standard software. Nobody should build a custom general ledger.
But every growing business eventually hits a workflow that the standard tools don’t cover — a process held together by spreadsheets, email threads, and one person’s memory. That’s the moment the “build vs. buy” question actually becomes worth asking. This guide is about how to answer it honestly.
The Signs You’ve Outgrown Off-the-Shelf
Custom software isn’t about wanting something fancier. It’s about a specific process that’s costing you real time, money, or risk, and that no product on the market fits. A few honest signals:
- The “spreadsheet that runs the business.” There’s a workbook so critical that if it broke, work would stop — and only one or two people understand it. That’s a system pretending to be a file.
- Re-keying the same data into three places. When information gets copied from an email to a spreadsheet to your accounting software by hand, you’re paying salaries for work software should do — and inviting errors at every hop.
- A workflow no product matches. Your approval chain, your compliance clock, your way of matching documents — you’ve tried to force it into an off-the-shelf tool and spend more time fighting the tool than doing the work.
- You’re paying for software you barely use. Sometimes a stack of five subscriptions, each used for 10% of its features, costs more and works worse than one system built for exactly what you do.
The test isn’t “could custom software be nice?” It’s “is a specific, repeating process costing us enough — in hours, errors, or risk — that fixing it would pay for itself?” If you can’t name the process and the cost, you’re not ready to build.
Build vs. Buy: An Honest Framework
Buying is almost always the right first answer. Building is right only when the process is both core to how you operate and genuinely unserved by the market. Here’s how the two compare on the factors that actually matter:
| Factor | Off-the-shelf | Custom system |
|---|---|---|
| Setup speed | Fast — sign up and go | Slower — weeks to a first working version |
| Fit to your workflow | You adapt to the software | The software adapts to you |
| Cost shape | Per-seat, forever, rising with headcount | Higher upfront, lower to run at scale |
| Best for | Standard, common processes | A core process nothing on the market fits |
The mistake most businesses make isn’t choosing wrong — it’s choosing custom for a problem that a product already solves, or choosing to suffer with a product for a problem that clearly needs a system. Match the tool to the problem, not to the ambition.
The Real Risk of Custom Software (and How to Kill It)
Custom software has a bad reputation for good reasons: projects that run for a year, blow the budget, and ship something nobody uses. Almost every one of those failures comes from the same root cause — a big specification written up front, built in the dark, and revealed only at the end.
The fix is to build in thin, working slices. Ship the smallest version that does something real, put it in front of the people who’ll use it, and refine on their actual data. You find out in week three whether it fits — not month twelve. Two things make this work in practice:
- Build with people who understand the domain, not just the code. A finance workflow built by developers who’ve never closed a book will miss the things that matter — the edge cases, the compliance rules, the reconciliation that has to tie out. Domain understanding is what separates a tool that gets used from one that gets abandoned.
- Insist on ownership and maintenance. Software that’s built and handed over goes stale fast. The system should be hosted, maintained, and improved by the team that built it — so it keeps working as your business changes.
What This Looks Like in Practice
Custom doesn’t mean exotic. The best custom systems are usually unglamorous — they take one painful, repeating process and make it disappear. A shared inbox that auto-sorts and routes hundreds of enquiries a day. A records vault that turns “where’s that voucher?” from an afternoon into a search box. A dashboard that reads broker statements and computes profit-and-loss automatically instead of by hand.
None of those needed to be invented from scratch as concepts — they needed to be built for one business’s exact workflow, because no product matched it closely enough to be worth the friction. That’s the whole game: not novelty, but fit.
Have a workflow that needs a system?
Tell us about the process that’s stuck in spreadsheets and email. We’ll tell you honestly whether it’s worth building — and if it is, how we’d do it.