When Does a Small Business Need Custom Software?

Almost nobody needs custom software. That's the honest starting point, and it's a strange thing to publish on the website of someone who builds it.

But a minority of small businesses genuinely do, and they usually don't realise — because the cost of not having it is spread thinly across everyone's week as "just how we do it". This is about telling those apart.

First: what "custom software" actually means here

Not an app in the App Store. Not a startup. Usually something much duller: an internal tool that handles one process the way your business actually does it. A jobs board. A quote builder. A dashboard showing the four numbers you currently assemble by hand.

Most of the custom software small businesses need has no users outside the business, no logo, and no marketing page. It's a tool.

The eight signals

One of these on its own is normal. Three or four together is when building starts being the cheaper option.

1. A spreadsheet has quietly become your system of record

Not a spreadsheet used for sums — the spreadsheet that *is* your job list, your stock, your bookings. The one where several people edit one file, there's no history of who changed what, and a wrong click in the wrong cell is unrecoverable.

Spreadsheets are extraordinary and I use them constantly. But they have no validation, no audit trail and no concept of who's allowed to do what. When the business depends on one, you don't have a spreadsheet problem — you have an application with no safety rails.

The test: if that file were deleted right now, how long to reconstruct it? If the answer is days, that's your signal.

2. Someone types the same information more than once

A customer's details go into the booking tool, then the invoicing tool, then the spreadsheet. Nobody scheduled that work. It's invisible on any report, and it's often hours a week — plus the errors, which surface later as an argument about what was agreed.

3. Your tools don't talk, so a person is the integration

Four good tools, none of which know about each other, and a human moving information between them. That human is doing a computer's job, and they know it.

This is the one I built my own product around. Zingg Homeschool OS exists because homeschool families end up running a school across a dozen systems — planning in one place, records in another, documents nowhere in particular — and the parent becomes the integration layer. That phrase is the signal, in any business.

4. Your workflow is genuinely unusual

Every owner thinks theirs is unusual, and mostly it isn't — that's why off-the-shelf software exists and is cheap. But some businesses really do have a step, a rule, or an order of operations that no product anticipated, and that oddity is often the thing they're actually good at.

The test: can you name the specific step no tool handles? If you can describe it in one sentence and it's central to how you work, that's real. If it comes out as "we're just different", it isn't yet.

5. Your team has invented workarounds

The status field everyone uses to mean something else. The naming convention in the notes box. The colour code. The second spreadsheet that exists because the software wouldn't do the thing.

Workarounds are documentation. They're your team telling you exactly what's missing, in the most specific possible terms. Write them down before you talk to anyone about building — they're most of a specification.

6. The software almost works, and the gap is expensive

"It does 90% of what we need." The question is what the last 10% costs you. Sometimes nothing — you adapt and move on, which is the right answer. Sometimes it's the part customers see, or the part that takes two hours every Friday.

7. The admin grows faster than the business

Twice the customers, three times the paperwork. That's the clearest sign that a process which worked at ten doesn't work at fifty — and the point of building something is usually to break that link, not to add features.

8. Your customers want something no tool will give them

A portal to see their own job status. A way to submit what you need without three emails. Something that makes dealing with you noticeably easier than dealing with your competitor.

This is the only signal on the list that can make money rather than save it, which makes it the most interesting and the easiest to overestimate. Ask two customers whether they'd use it before you build it for all of them.

When you should not build

Genuinely more important than the list above, because this is where the money gets wasted.

1. 01You haven't tried the boring option. Most problems that feel custom are a setting, a plugin, or a different plan of software you already own. Exhaust that first. It's free and it's often the answer.
2. 02The process is broken, not the software. Building a tool around a bad process gives you a bad process that's now expensive to change. Fix it on paper; sometimes that's the whole project.
3. 03Nobody owns it. Custom software needs one person who decides what it does and says no to things. Without that, it becomes a feature list nobody agrees on and it never gets finished.
4. 04You can't afford the second year. The build is not the cost. Hosting, small fixes, the change you'll want in month four — if there's no budget for that, you'll end up with something you're afraid to touch.
5. 05It's about growth you're hoping for. Building for the business you want in three years is how small companies buy enterprise software they never use. Build for what's breaking now.

What to try, in order

1. 01Change the process. Free. Try it for a fortnight.
2. 02Use what you have properly. A setting, a template, a feature on a plan you're already paying for.
3. 03Buy something off the shelf. Cheap, fast, and supported by someone else. Here's how that comparison actually works.
4. 04Automate the joins. If the tools are right but disconnected, connect them. Fifteen candidates here.
5. 05Build the smallest thing. One process, one screen, the version that's useful in a month rather than perfect in a year.

Most businesses that come to me for step five belong on step two or three, and I'd rather say so. The ones who genuinely belong on five have almost always been through the others already — that's usually how they know.

the goal is the problem going away. sometimes that costs nothing.

If it is a build

Then the thing that decides whether it goes well is scope, not technology. Define the smallest version that removes the worst part of the problem, use it, and add only what turns out to be missing. Everything I build for myself works that way — and it's the reason the ones I've shipped exist at all.

As for what that costs. Custom software, dashboards and automation projects start at $3,000. That is a starting point, not a price: every custom project is scoped and quoted on what it actually involves, and you get that quote in writing before anything begins. There's more on what moves that number.