Asset Tracking ROI: How to Build the Business Case (With a Model)

The usual business case for asset tracking leads with the equipment you'll stop losing. It's the hardest number in the model to defend, and in the worked example below it doesn't cover the running cost on its own. The case that survives a finance review is built from inputs you can measure from your own records: repair invoices, purchase history, a short time log and a physical count.

This guide sets out the formula, the five inputs, how to cost staff time using published wage data, and a worked model with a payback period and a sensitivity table. The figures in the model are hypothetical and labelled as such. Replace them with yours.

The formula

Keep it simple enough that someone in finance can rebuild it from your spreadsheet in five minutes.

• Annual benefit = losses avoided + warranty and contract claims recovered + duplicate purchases avoided + staff time saved + leaver equipment recovered.
• Annual running cost = subscription + ongoing admin time.
• One-off cost = labels, any hardware, setup hours and the first labelling count.
• Payback period (months) = one-off cost ÷ (monthly benefit − monthly running cost).
• ROI over a period = (total benefit − total cost) ÷ total cost.

Present the payback period first. "Pays back in nine months" is easier to test and harder to argue with than a percentage. If you need a full cost picture, how much asset management software costs breaks down the one-off and running costs in more detail.

Five inputs, ranked by how well you can evidence them

Missed warranty and contract claims

Pull last year's repair and replacement spend. For each line, check whether the item was under a manufacturer warranty, an extended warranty or a service contract on the date the work was authorised. The ones that were covered and paid for anyway are pure recoverable cost, and they come from invoices, which finance already trusts.

If you can't tell whether an item was covered on a given date, record that as a finding. It says your records can't support a claim even when you have one. There's more on the different kinds of cover in warranty vs service contract vs extended warranty.

Duplicate purchases

Cross-reference last year's equipment purchases against what you already owned at the time. Look for items bought by one team or site while an identical one sat unused elsewhere. The number is usually modest, but it's specific and slightly embarrassing, which makes it persuasive.

Staff time

Ask a handful of people who handle equipment to log, for two weeks, every time they go looking for something or chase who has it, and how long it took. Add the time spent reconciling the spreadsheet. Then scale to a year and to everyone in a similar role.

Cost the hours at a loaded rate, not a salary. The US Bureau of Labor Statistics puts employer compensation costs for private industry workers at $46.89 per hour worked in June 2026, of which $32.82 was wages and $14.07 benefits. For IT staff specifically, BLS reports a median annual wage of $61,860 for computer user support specialists in May 2025, before benefits. In the UK, the ONS reports median hourly earnings of £19.67 for full-time employees in April 2025, excluding overtime, which is a wage figure before employer costs. Your own payroll data beats all of these.

Then apply a realistic reduction. Software doesn't remove searching; it removes the searches caused by not knowing who has what. I'd claim half and be ready to explain why.

Leaver equipment

Compare the number of people who left last year with the devices formally returned. Gaps are common when a leaver works remotely or the return isn't part of the offboarding checklist. Each recovered laptop is a replacement you don't buy. Our guide to remote employee equipment return covers the process side.

Losses avoided

Use it, but put it last and keep the assumed reduction modest. Your baseline understates the problem, because items nobody registered can't appear in a write-off report. Say that openly; it makes the rest of the model more credible.

Published data can frame the size of the risk, as long as you're clear about its age. A 2010 study by the Ponemon Institute, sponsored by Intel, surveyed 329 US organisations that lost 86,455 laptops in a year: an average of 263 each, of which just 12 were recovered. It put the loss rate at 7.12% over a laptop's useful life and found 43% were lost off-site and 33% in transit. These figures are old and vendor-sponsored; laptop loss and theft statistics goes through what each study measured and what current breach data shows.

Cost a lost device as more than its replacement price. Add the IT time to revoke access and set up a new machine, and the employee's lost working time. The cost of lost equipment calculator does this for you, and the real cost of lost equipment explains the layers.

A note on tax

Tax changes the after-tax cost of a loss, which is a reason to keep your ROI model pre-tax and consistent. In the US, IRS Publication 547 says the deductible loss on stolen or destroyed business property is its adjusted basis, minus any salvage value and any insurance or other reimbursement. A three-year-old laptop that has been largely depreciated has a small adjusted basis, so the deduction is small while the replacement cost is the full price of a new machine. In the UK, HMRC's Business Income Manual treats periodical software payments as deductible unless they are instalments of a capital sum. Neither point changes the logic of the case; both are worth a line so finance can see you've considered them.

Worked model: a 120-person organisation

Everything in this section is hypothetical. The organisation has 120 staff and 400 tracked assets, including 150 laptops at a replacement cost of $1,400 each. Staff time is costed at $47 an hour, the BLS figure rounded up. The subscription is an assumed $3,000 a year, not any vendor's price.

Results

• Net benefit in year one: $12,600 − $4,692 − $2,607 = $5,302, an ROI of 73% on year-one cost.
• Three-year net benefit: $21,119, an ROI of 127% on three-year cost.
• Payback: 4 months if benefits start in full straight away. Assuming benefits run at half for the first six months while labels go on and habits change, payback moves to month 9. I'd present the second figure: it is the one that survives a finance director's questions.

Look at which rows carry the case. Measured search time and recovered cover claims together are about 44% of the benefit. Losses avoided, the input most business cases lead with, is under 30%.

Sensitivity: what the case depends on

A sensitivity table shows finance you've already asked the awkward questions. Each row changes one assumption from the base case and leaves everything else alone. Payback here assumes full benefits from the first month, so add a few months for ramp-up as in the base case.

Two rows matter most. If the benefits come in at half your estimate, the system still pays for itself over three years, slowly. And if the case rests on losses alone, it doesn't pay back at all. That's the strongest argument I know for doing the half day of invoice work and the two-week time log before you write anything.

Presenting it

1. Lead with measured inputs. Invoices and logged hours first; modelled loss reductions last.
2. State the baseline's weakness. Loss data understates the problem because unregistered items aren't in it. Saying so pre-empts the obvious challenge.
3. Give the payback period with a ramp-up, and show the sensitivity table beside it.
4. Keep unquantified benefits separate. Being able to produce a verified count for an auditor or insurer is real value, but adding a guessed figure for it to the total is what gets a case picked apart.
5. Name who will own the system. A case with no owner is a case for buying software nobody maintains.

Measuring after the fact

A business case is a prediction. Measuring the outcome is what makes the next one credible, and it's the step that's usually skipped once the money is approved.

1. Baseline before you start: a discrepancy rate from a scoped physical count, last year's repair spend, and the two-week time log.
2. Re-measure at six months using the same method on a fresh sample.
3. Report the difference, including anything that didn't improve. A report that says two of five inputs moved is trusted more than one claiming all five did.
4. Adjust. If the discrepancy rate hasn't moved, the usual cause is that recording a handover is still harder than skipping it.

Our guide on how to do an asset audit covers running the scoped count.

Set expectations against that table. A sponsor expecting the loss rate to fall in month two will conclude it failed, when it's simply too early to tell.

If you use Stackroom, several of these measurements are standard reports: cover never claimed, repairs claimed on cover, lost and stolen, and missing at audit. Whatever tool you choose, check before you buy that it can produce the numbers you promised to report.

What weakens a case

• Industry benchmarks in place of your own numbers. A statistic lifted from a vendor page invites the question of who paid for it.
• Loss reductions above 50%. A defensible figure beats an impressive one.
• Soft benefits presented as hard ones. "Better visibility" is not a line item.
• Leaving out setup and admin time. Finance knows nothing is free, and the omission costs more credibility than the number would have.

Sources

• Employer Costs for Employee Compensation, June 2026 — US Bureau of Labor Statistics, 2026
• Computer Support Specialists, Occupational Outlook Handbook — US Bureau of Labor Statistics, 2026
• Employee earnings in the UK: 2025 — Office for National Statistics, 2025
• The Billion Dollar Lost Laptop Problem — Intel, research by Ponemon Institute, 2010
• Publication 547, Casualties, Disasters, and Thefts — Internal Revenue Service
• BIM35810: Computer software, general considerations — HMRC Business Income Manual