Every quote you have been given is probably right. That is the confusing part. One studio says twelve thousand, another says ninety, and both are describing something they could genuinely deliver. They are just not describing the same thing.

So before comparing numbers, it helps to know what the numbers are made of.

What you are actually paying for

A system that runs a business is rarely dominated by the screens you imagined. In our experience the build itself is somewhere between a third and a half of the first-year cost. The rest sits in five places.

  • Discovery. Writing down how the work is done today, including the parts nobody documented and the spreadsheet one person maintains. This is where projects are saved or lost.
  • Integrations. Accounting, payments, equipment, a supplier portal, an existing database. Each one is a small project with someone else's schedule attached.
  • Data migration. Moving years of history out of whatever holds it now, and reconciling it once it lands. Almost always underestimated.
  • Rollout and training. The system is not delivered when it is deployed. It is delivered when people stop using the old spreadsheet.
  • Running it. Hosting, backups, monitoring, and the small changes that never stop.

Ask any studio to price those five lines separately. A quote that covers only the build is not cheaper — it is incomplete, and the difference will find you in month four.

Where budgets actually break

Three things account for most overruns, and none of them are the interface.

The first is integrations with systems you do not control. If the estimate assumes a documented API and the reality is a nightly file export, that is not a small correction.

The second is data. Old records are messy in ways nobody remembers until they are moved. Budget for cleaning, not just copying.

The third is the second year. Custom software has no licence fee, which makes it look free once it is built. It is not. Plan for a modest ongoing budget from the start, and the system stays useful. Skip it, and in three years you will be paying to replace something that only needed maintaining.

When buying beats building

We turn down custom work fairly often, because in these three cases a product you can subscribe to today will serve you better.

  • Your process is genuinely standard. If what you do looks like what thousands of other companies do, someone has already built it better than a first version of yours will be, and they have been fixing it for a decade.
  • The requirement is regulatory and shared. Accounting, payroll, tax reporting. Rules change, and keeping up with them is somebody's full-time job. Let it be theirs.
  • You are still discovering what you need. If the process is changing every month, a rigid custom system will freeze a shape you are about to outgrow. Use something flexible and cheap until the shape holds still.

There is a simple test. If you removed the software entirely, would your competitive position change? If not, buy it.

When building is the cheaper answer

Custom starts winning when the process is the business.

We built a production and operations system for a manufacturer whose entire margin came from how they sequenced work. No product on the market modelled that sequence, and the ones that came close would have required them to work like everyone else, which is exactly what they were being paid not to do.

The other case is arithmetic. Per-seat pricing is comfortable at ten people and painful at two hundred. If a subscription is set to cost more over five years than a build plus its maintenance, the decision makes itself — just be honest about the maintenance side of that comparison.

And there is a third, quieter reason: leverage. When the tool is yours, a competitor cannot buy the same one.

How to decide in a week

You do not need a long evaluation. You need four conversations and one document.

Write down the process as it actually runs today, with the workarounds included. Then shortlist two or three existing products and get a demo using your data, not theirs. Ask each vendor what they cannot do. Finally, ask a studio to price a custom build against the same document.

If the products cover eighty percent and the missing twenty percent is not where you make your money, buy. If the missing twenty percent is exactly where you make your money, build that part and integrate the rest.

The worst outcome is not choosing wrong. It is choosing without writing the process down first, because then nobody — you, the vendor, or the studio — is looking at the same thing.

If you are at that point and want a second opinion on the numbers, that conversation is free and usually takes half an hour.