Every small business hits this fork eventually. There's a process that needs software, and there are two roads: subscribe to something that exists, or build something that's yours.
The usual advice is useless because it comes from people selling one of the roads. The subscription company says custom is expensive and risky. The dev shop says subscriptions are inflexible and you'll outgrow them. Both are telling you half the truth and pocketing the other half.
I build custom software for a living, so watch me argue against my own interests for a minute. Here's the actual tradeoff.
When off-the-shelf wins, and it often does
If your problem is common, buy the common solution. Accounting, email, scheduling for standard appointments, point of sale. Thousands of businesses have your exact problem, so mature products exist, and their monthly fee buys you a support team, security updates, and years of other people's mistakes already fixed.
The rule of thumb I give clients: if you can describe your need without ever saying "except" or "but we do it differently," buy, don't build. QuickBooks doesn't care that you're special, and for accounting, you aren't. None of us are.
Off-the-shelf fails you in a specific, predictable way: the last 20 percent. The product does most of what you need, and then you spend years fighting it over the rest. Workarounds pile up. You export to a spreadsheet to do the thing it won't do, and now you're running two systems. Per-seat pricing grows with your team. And the workflow bends to fit the software instead of the other way around.
When custom wins
Custom earns its cost in exactly one situation: when the process is part of how you win. Not how you operate, how you win. The quoting method that closes jobs your competitors lose. The scheduling logic specific to your trade. The customer experience nobody else offers because no off-the-shelf product imagines it.
Generic software makes you operate like everyone else, by design, because it's built for everyone. If a process is genuinely yours, renting the generic version means sanding off your own edge.
Custom's failure mode is just as predictable: building it for the wrong reasons. Building because subscriptions "feel like waste" (a $50 subscription is cheaper than any custom build for years). Building the deluxe version before validating the simple one. Or building with someone who disappears after launch, leaving you holding software nobody maintains.
The math nobody does out loud
Take the subscription price times 36 months, add the hours your team spends on workarounds, priced at what those hours cost you. That's the real price of off-the-shelf. Compare it against a custom build quote plus modest hosting and maintenance. In 2026, with AI-assisted development cutting build costs to a fraction of what they were, this comparison flips to custom far earlier than most owners assume. Small focused tools now get built in days.
Sometimes the math still says subscribe. Good. Subscribe with a clear conscience.
The hybrid most people miss
The best answer is often both: keep the off-the-shelf products for the common stuff and build one small custom piece that connects them or covers your weird 20 percent. A little bridge between your booking system and your invoicing. A custom quoting front-end feeding a standard back office. Highest return per dollar of anything I build.
If you're staring at this fork right now, tell me about your process and I'll tell you which road I'd take, including when the honest answer is "just subscribe." Grab 30 minutes at ctfdesigns.com/book. Bring the weird 20 percent.