The Software Budget Line Nobody Writes Down

It's budget season, which means somewhere in your draft for next year there's a line that says 'Software,' holding last year's number plus three percent. It's one of the most carefully reviewed budgets in the company, and one of the least accurate, because the line only contains the part of the cost that arrives as an invoice.

Where the real spend hides

The licences are the visible part. The real spend is the labour around the seams: the places where two systems don't talk and a person has been quietly hired, or quietly repurposed, to be the integration.

• Re-keying: every number typed from one system into another, times every person who does it, times fifty weeks a year.
• Month-end chasing: the controller's week spent hunting numbers across spreadsheets and inboxes instead of reviewing them.
• Spreadsheet maintenance: the load-bearing workbooks that take real hours to feed, reconcile, and repair when a formula breaks.
• Error correction: the invoice that went out wrong because the field ticket was re-typed, and the hour spent making it right.
• The quick exports: 'just pulling it into Excel' a dozen times a week, per department, forever.

None of this shows up as a line item, because nobody sends you an invoice for it. It's spread across a dozen payrolls in amounts too small to question, which is exactly why it compounds for years without a decision ever being made about it.

A ten-minute budgeting exercise

This year, write two lines instead of one: 'Software' and 'Working around software.' For the second line, you don't need a study. Count the handoffs: every place a number moves between systems by hand, estimate the hours per week it takes, and multiply by a loaded hourly rate. Be conservative. The number will still be uncomfortable, and in most companies we've sat down with, it's larger than the first line.

Why the second line matters

The point of the exercise isn't guilt. It's that the two lines behave differently. The first line is a cost you negotiate occasionally. The second is a cost you can actually engineer away, because it exists only where systems meet. Consolidate the systems, or connect them properly, and the hours come back without anyone working harder.

It also reframes the conversation about replacing software. Measured against licence cost alone, a modern ERP looks expensive next to the pile of tools it replaces. Measured against both lines, the pile is usually the expensive option, and it gets more expensive every year as volume grows and the seams multiply.

If the number scares you

That's not a budgeting problem. That's a roadmap. The companies that fix this rarely do it in one heroic project; they retire the most expensive seam first, usually the one between operations and accounting, and let the recovered hours fund the next step. One connected system, like Acumatica, exists precisely to remove the seams rather than staff them. If you want a second opinion on where your own second line is coming from, that conversation is free, and it's the kind we have at Summit every week.