Ask a software engineer about technology debt and they'll describe legacy code. Ask me, and I'll describe a spreadsheet: built in an afternoon years ago as a quick fix, now version 47, undocumented, owned entirely by one person's memory, linked to nothing, and quietly load-bearing for pricing, rostering or job tracking. Every SME runs on a layer of these — the workarounds, the manual bridges between systems, the "temporary" tools that achieved permanence by stealth. That layer is technology debt, and like financial debt it has an interest rate: paid daily, in hours, errors and fragility, by people who've stopped noticing the repayments. Here's how to find yours, price it, and pay it down without a big-bang rebuild.
What SME tech debt actually looks like
The recognisable species: the load-bearing spreadsheet (pricing models, rosters, job trackers — powerful, invisible to controls, one corrupted file or one resignation from disaster); the swivel-chair integration — a human re-typing data between systems that don't talk (the quote into the invoice, the timesheet into payroll), which we've flagged elsewhere as the purest symptom; the abandoned half-implementation — the system bought, partially configured, and now used at 30% while the old way limps alongside it, doubling the work it was meant to halve; the founder-era tool that fit ten people and now strains at twenty-five, surviving on workarounds; and the undocumented automation — the clever macro, the integration flow, the script Dale wrote — running silently until the day it doesn't, with no one alive who knows how. Individually, each was a rational shortcut. The debt is the accumulation: a business where every process carries two or three of these taxes simultaneously.
Price the interest, not just the principal
The reason debt persists is that the repair cost is visible while the interest is diffuse. So make the interest visible — for each item, estimate three lines: hours (the weekly human time the workaround consumes, at loaded rates — the swivel-chair jobs alone usually shock), errors (the reworks, the mispricings, the payroll corrections traceable to the manual step), and fragility (what happens the week the owner of this thing is sick, resigns, or the file corrupts — score it like the single-person dependencies it is). A genuinely typical SME finding: fifteen-plus hours a week of pure interest payments across the debt portfolio, which at loaded cost is a part-time salary spent maintaining shortcuts. Against that, most repairs are cheap — which is the entire argument, and it can't be made until the interest has a number.
Pay it down like debt: highest interest first, no consolidation loans
The repayment strategy borrows directly from finance. Rank by interest rate, not by size or annoyance — the debt costing twelve hours a week outranks the ugly one costing one. Pay one down at a time, properly: replace the spreadsheet with the system feature that already exists (audit your current platforms first — a remarkable share of SME tech debt duplicates capability you're already paying for), connect the two systems that should talk, finish the half-implementation or formally kill it (a decision either way beats the zombie state), document the automation and give it a second owner. Each repayment follows the systems-upgrade rules we've covered: a slice, a parallel run for anything touching money, an owner who isn't the busiest person in the company. And beware the consolidation loan — the all-new everything-platform that promises to retire the whole debt portfolio in one migration. Occasionally right, usually the SME equivalent of refinancing credit cards onto a bigger card: the failure mode of most eighteen-month implementation stories began as a debt-consolidation dream.
Then stop new borrowing, or at least borrow knowingly. Tech debt is created in sentences like "just track it in a spreadsheet for now" — which is sometimes the correct call (speed matters; not everything deserves a system). The discipline isn't prohibition; it's a register: every "for now" solution gets a line — what it is, who owns it, what it should eventually become — reviewed twice a year alongside the systems-health questions we've posed elsewhere. Debt you've listed, priced and scheduled is leverage; debt you've forgotten is just decay with a formula bar. The businesses that handle this well aren't the ones with no spreadsheets. They're the ones where every spreadsheet knows whether it's a tool or a tenant — and nothing load-bearing is held up by one person's memory and a file called FINAL_v47_USE_THIS_ONE.
The load-bearing spreadsheets and swivel-chair jobs show up fast in our free Business Health Check. Five minutes to see what your business is really running on.
If this is the problem you are working on, Valont’s Finance hub runs accounts receivable and bookkeeping as part of a connected back office, and the back-office cost calculator shows where to start.
About the author
Nick Lucock
Chief Executive Officer, Valont
Nick Lucock is the Chief Executive Officer of Valont and leads its day-to-day operations across Finance, People, Operations and Growth. He runs the team that does the work Valont publishes about: the bookkeeping, BAS and payroll runs, the award interpretation on every pay cycle, the IT and security layer, and the advisor cadence that keeps an owner's back office current. Nick writes the practical side of Valont's library, the processes, systems and tools that keep Australian SMEs compliant and growing, and he is the author of most of the operational articles on the site. He works with owners across construction and trades, health, hospitality and professional services.
LinkedIn →