“We could just build that” is said in every company with engineers. Sometimes it is right. The way to find out is not a debate about features; it is four numbers and one division.

The four numbers

  1. What you pay now. The monthly subscription, at your real seat count and your real billing terms.
  2. What it costs to build. Engineering hours for a version that covers what your team actually uses — not the vendor’s whole feature list — multiplied by what an hour of engineering costs you.
  3. What it costs to keep. Hours a month to maintain it (bug fixes, dependency updates, small requests), plus hosting.
  4. How long you would keep it. Internal tools are long-lived. Three years is a reasonable horizon to compare against.

The division

Subtract upkeep from the subscription. That is what you save each month once the tool exists. Divide the build cost by that monthly saving and you get the break-even month.

break-even month = build cost ÷ (subscription − upkeep)

Worked example with round numbers: a tool costs $1,000 a month. A replacement takes 200 hours at $60 an hour, so $12,000 to build. Upkeep is 6 hours a month plus $20 hosting, so $380. The monthly saving is $620, and the build pays for itself in month 20.

If the subscription is smaller than the upkeep, the answer is simple: it never pays back, however cheap the build looks.

Where building tends to win

The division favours tools that are simple and expensive at your scale:

  • The part you use is narrow — a board, a form, a single workflow — even if the product does much more.
  • The price grows with every seat or every task, and your count keeps rising.
  • The data is yours and the tool is mostly storage plus a screen on top.

Automation glue and simple boards are the usual examples. Our catalog scores tools from 1 (trivial to build) to 5 (don’t build); scores of 1 and 2 are where this most often works out.

Where it almost never does

  • Infrastructure-grade products. Chat, video, password management, file sync. Reliability and security are the product; your version would be worse and cost more to run.
  • Network-effect products. If customers or partners use the tool with you, an internal copy cuts them off.
  • Small teams. At a handful of seats, the subscription is usually smaller than the upkeep, and the division above returns “never”.

Questions to ask before you start

  • Who maintains it in a year, when the person who built it has moved on?
  • What happens to the data and the workflows if you later go back to a vendor?
  • Is there a cheaper plan, or a consolidation, that gets most of the saving without building anything?

The last question matters most. Downgrading a plan or merging two overlapping tools saves money the same week; a build saves money from its break-even month onward. Do the cheap moves first, then build where the arithmetic is still clearly in your favour.

Each tool page on this site shows the build score, an estimated build range and the break-even month for a reference team, with a calculator for your own seat count.