Most businesses should buy the software, not build it.

If your business runs the way most businesses in your industry run, buy the software. Configure an existing platform, accept the ninety per cent fit, and spend the money you saved on the part that actually makes you money.

Custom is worth it in one situation: the way you work is the thing you are better at, and no product on the market is shaped like it. That is narrower than it sounds, and we say so as a studio that sells custom builds.

Why every article about this says "build"

Search this question and almost every result is written by a software development company. They conclude that custom wins. We develop custom software and we are telling you it usually does not, which should tell you something about the incentive behind the rest.

The honest version has three parts, and only one of them is a yes.

When off-the-shelf is right

Your processes are normal. Quotes, invoices, stock, payroll. If you can describe your operation to a peer in your industry and they nod, a product already exists that fits it. The ninety per cent fit you resent is cheaper than the hundred per cent you pay for and then maintain.

Nobody on your team can own software. A custom system is an asset with an owner. If nobody is going to hold it, a vendor's support contract is doing a job you cannot do yourself, and that job is worth paying for.

You are not sure what you need yet. Buy something, run it for a year, and let it show you where it does not fit. Those gaps are a much better brief than anything you could write today.

When custom is right

The repeat work is eating a person. Not "takes a while" — eating. If someone's week is mostly moving the same information between the same places, that is a system shaped like your business, and no product knows the shape.

You already pay for five tools that do not talk. The subscriptions are rarely the cost. The cost is the person who reconciles them, and that person is doing it because nothing on the market sits in the middle of your particular five.

The way you work is the advantage. If your process is why clients choose you, configuring someone else's product to approximate it is slowly giving the advantage away.

If none of those three is true of you, stop here and go and buy something.

What we actually did for one client

An Australian business. Confidential, so there is nothing here to show, which is the right way round: it is their system, not our portfolio piece.

We studied how the business ran before proposing anything. Where the hours went, what repeated, and what a person still had to decide. That study is also the honest test — about half the time it finds the answer is a product, and we say so before anyone has spent money on a build.

We put the whole system on paper first. Every step, every approval. Paper is the only point in the process where changing your mind is free, and we would rather use that up than discover a misunderstanding in week six.

We built one system, not a stack. AI inside the steps that repeat, a person kept wherever a decision actually matters. Not AI deciding and a human rubber-stamping — the approvals are real, and the system stops and waits for them.

It is in use inside the business. That is the whole result, and it is the only one we can show you.

The question that settles it

Not "what features do we need". Everyone can write that list and every product will seem to tick it.

Ask instead: what does someone here do every week that no product has ever heard of? If the answer is nothing, buy. If it takes you ten minutes to explain and the person explaining it is the one holding the business together, you have found the case for building.

What I would tell someone deciding

Price the person, not the licence. Comparing a subscription to a build cost is the wrong sum. The right one is the subscription plus the hours somebody spends compensating for it, every month, forever.

Make whoever you hire study you first. If a studio quotes a build before it understands how your week works, it is quoting a guess, and you are the one carrying the risk in it.

Get the whole system on paper before any of it is built. It is the only stage where changing your mind costs nothing, and the only one people try to skip.

Ask who owns it afterwards. A custom system with no owner inside the business becomes a dependency on whoever built it. Agree the handover, the source and the documentation before the first line is written, not after.

Be suspicious of anyone who always says build. Including us, if we ever do.