What This Means

• If a spreadsheet has run a process for more than a quarter, it’s infrastructure, whether anyone labeled it that way or not.
• Infrastructure needs an owner, version control, and validation. Most “temporary” spreadsheets have none of the three.
• The word “temporary” is doing a job: it’s letting everyone avoid the work of designing something real.

This piece is the opinion companion to From Spreadsheets to Systems: How HR Leaders Can Escape Broken, Disconnected Tooling, which lays out the practical fix. This one argues for why the fix is overdue.

The Argument

I’ve sat across from HR leaders who describe their spreadsheet the same way every time: “it’s just temporary, until we get the real system in place.” I’ve heard some version of that sentence for years, from teams where the spreadsheet has been running payroll adjustments for three or four years. Nobody’s lying when they say it. They believe it. That belief is the problem.

Every growing business has one: the person who also does HR, and the spreadsheet that holds it together. That person built the spreadsheet under deadline pressure, it worked, and calling it temporary let them avoid admitting they’d just built a piece of core infrastructure without meaning to. A temporary tool doesn’t need documentation. It doesn’t need a named owner. It doesn’t need validation rules. A permanent one does. The label “temporary” is what’s letting a critical process run without any of that.

I don’t think this happens because HR teams are careless. I think it happens because nobody schedules the moment a workaround graduates into infrastructure. There’s no calendar reminder that says “this spreadsheet just crossed the line from stopgap to system of record, treat it accordingly.” The label just never updates, and the tool keeps running exactly the way it did on day one, minus whatever controls it would have gotten if anyone had built it on purpose.

Research backs up how common this is. HR.com’s 2025 State of Today’s HR Technology and Integrations survey found 62% of organizations run between two and four separate paid HR solutions, and only 39% describe them as usefully integrated. That gap between owned tools and connected tools is exactly where a spreadsheet steps in to bridge the difference, and exactly where it becomes permanent without anyone deciding it should.

The cost isn’t hypothetical. A manual re-entry error, the kind a spreadsheet with no validation layer can’t catch, moved a $103K compensation figure into a system as $130K, a $27K overpayment that only surfaced when an employee questioned their own paycheck. That’s not a story about carelessness. It’s a story about a tool that was never given the safeguards a permanent system needs, because everyone kept insisting it was temporary.

Research from University of Hawaii professor Ray Panko, cumulated across years of field audits, found error rates in operational spreadsheets running as high as 88 to 94%. That’s not a fringe statistic. It’s the accumulated finding of one of the most cited bodies of research on spreadsheet reliability that exists. Treating that same tool as your system of record for payroll or benefits data isn’t a temporary decision. It’s a standing risk you’re renewing every day you don’t replace it.

Counterarguments

The most common pushback I hear is that replacing the spreadsheet is expensive and disruptive, and the current setup, however fragile, is at least familiar. I take that argument seriously. Migration is real work, and a poorly executed migration can be worse than the workaround it replaces.

But familiarity isn’t the same as safety, and the cost comparison usually leaves out the ongoing cost of the workaround itself. HR Cloud’s research, citing EY’s 2025 data, puts the cost of a single manual data entry at $4.86 for an HR professional. That cost repeats every single time someone re-types a number the spreadsheet should have synced automatically. The “cheap and familiar” option is rarely actually cheap once you count the recurring labor it consumes without anyone tracking it.

The second pushback is that the spreadsheet works fine, nothing has broken yet, so there’s no urgent reason to change it. I’d push back on “yet.” A spreadsheet with no validation layer isn’t safe because it hasn’t failed. It’s unmonitored, which is a different thing entirely. The absence of a visible failure is not the same as the absence of risk.

What to Do Differently

Stop calling it temporary. Call it what it is: the current system of record for whatever process it’s running, with all the responsibility that label implies. That single shift in language changes the conversation from “we’ll get to this eventually” to “this needs an owner and a plan right now.”

Run the audit. Map every process the spreadsheet is actually handling, not just the one it was originally built for. Spreadsheets accumulate scope without anyone approving it, and most have grown well past their original purpose by the time anyone questions them.

Set a real timeline, not an open-ended one. “Someday” is how a stopgap survives for four years. A specific quarter, with a named owner accountable for hitting it, is how it actually gets replaced.

For the practical audit that turns this argument into a ranked action list, see How to Audit Your HR Tech Stack for Manual Data Bridges. For what it looks like once the spreadsheet is finally retired, see How to Retire an HR Spreadsheet Without Losing Institutional Knowledge.