The licence is the cheap part of a legacy Windows app

Ask anyone what a legacy Windows application costs and you will get the price of the licence. It is the number on the invoice, it renews on a date and someone in finance can point at it. That is exactly why it is the wrong number to anchor on.

The programs that refuse to die, the ERP client, the accounting package, the vendor tool nobody can replace, tend to be old, specific and deeply embedded. Keeping them alive is rarely expensive in one place. It is expensive in forty small places at once.

What sits below the waterline

Start with installs. Every device that needs the program needs it installed, configured, patched and eventually reinstalled when the laptop dies. The first install takes an hour. The fifth, on a machine with the wrong runtime, takes an afternoon and a favour from someone senior.

Then support. Legacy programs fail in legacy ways: a missing dependency, a database client that no longer matches, a dialog that appears behind another window. The tickets are small and constant, and they go to the people who happen to know the app, who also have other jobs.

Then access. When an outside accountant, a contractor or a new branch office needs the program, someone has to decide how they reach it. Usually the answer is a VPN account, which hands over far more than the job needs, and has to be removed again later by someone who remembers.

Then data. A program that runs on each person's laptop keeps files on each person's laptop. That is a risk nobody invoices, until the day a device is lost and you find out what was on it.

Finally, people. Starting a new hire means installing the app. Ending an engagement means finding every copy. Both take time that nobody logged, because no single task was large enough to notice.

Why the real number stays hidden

These costs land in different places. The installs hit IT. The tickets hit whoever answers them. The VPN hits security. The lost afternoon hits the person who could not open their file. No one owns the total, so no one is ever shown it.

That is the quiet trap of legacy software. Each individual cost is below the threshold that triggers a project. The sum, over years, would have paid for several of them.

A way to count it

You do not need a spreadsheet to start. You need a few honest answers, and it helps to count hours before you count money, because hours are easier to estimate and harder to argue with.

• How many devices have the app installed, and how many are new each year?
• How many support requests mention it in a typical month, and how long does each take?
• Who from outside the company uses it, and how do they get in?
• Where do the files it produces live, and on which devices?
• How long does it take to onboard and offboard one person onto it?

Multiply each by a sensible hourly cost and add the licence last. The answer is usually a multiple of what everyone assumed, and it tells you which line is worth attacking first.

What moving the app actually changes

The common fix is to stop spreading the app across devices. Run it once, on a managed Windows host, and let people reach it from a browser. Installs collapse to one. Updates happen in one place. Outside users get a link to one application instead of a seat on the network. The files live on the host and not on laptops.

It is not free of trade-offs. Programs that need hardware plugged into a local machine, or that lean heavily on graphics, are harder. Licences that tie a program to one computer may not allow it to run elsewhere, and Windows itself still needs a valid licence wherever it runs. Those are things to check before you commit, not reasons to avoid the idea.

Whatever you choose, the exercise is worth doing on its own. Write the five answers down. If the total surprises you, you have found your business case without a single vendor call.