Build the case on four inputs: equipment you stop losing, warranty and contract claims you stop missing, time spent looking for things, and duplicate purchases of items you already own. The last three are easier to evidence than the first and are frequently larger.
Every business case I've seen for asset tracking leads with the same number: the value of equipment we'll stop losing.
It's the weakest input in the model. You're estimating a counterfactual — what wouldn't have gone missing — using loss data that's incomplete, because the items most often lost were never registered. Anyone competent in finance will ask about that, and they'll be right to.
There are three better inputs, and in my experience at least one of them is larger.
Input 1: equipment recovery
Use it, but use it carefully and put it last.
Formula: (annual write-offs × expected reduction). Be conservative on the reduction — 30 to 50% is defensible where you're introducing named custody and signatures for the first time, and anything above that invites scrutiny you don't need.
The honest caveat, which you should state rather than hide: your baseline understates the problem, because unregistered items don't appear in it. Saying so up front makes the rest of your model more credible, not less.
Input 2: missed warranty and contract claims
This is the one people forget and it's usually easy to evidence.
Pull last year's repair and replacement spend. For each line, check whether the item was inside a manufacturer warranty, an AMC or a CMC at the time. You will find some that were and weren't claimed, because nobody checked at the moment the repair was authorised.
|
Step |
Where the data is |
Typical finding |
|---|---|---|
|
List last year's repairs and replacements |
Finance, purchase ledger |
More lines than expected |
|
Match each to the asset's cover at that date |
Purchase files, supplier emails |
Cover status often unknown |
|
Count those that were covered and unclaimed |
The overlap |
Usually non-zero, sometimes substantial |
If matching cover to date is impossible with your current records, that's itself the finding, and it's a stronger argument than any projection.
Input 3: time spent looking for things
Soft, but defensible if you measure rather than assume.
Ask five people to log, for two weeks, every time they went looking for a piece of equipment and how long it took. Multiply by loaded hourly cost and headcount. Two weeks of real data beats any industry statistic you could quote, and it's specific to you, which is what makes it survive a finance review.
Input 4: duplicate purchasing
Buying something you already own, because nobody could find it or nobody knew it existed. Easy to find and genuinely embarrassing, which makes it persuasive.
Cross-reference last year's equipment purchases against your existing register. Look for items bought in one department that already sat unused in another. In multi-site organisations this is reliably the second-largest number in the model.
Putting the model together
|
Input |
Evidence strength |
Effort to produce |
Where it usually lands |
|---|---|---|---|
|
Missed cover claims |
Strong — actual invoices |
Half a day |
Often the largest single figure |
|
Duplicate purchasing |
Strong — purchase records |
Half a day |
Large in multi-site organisations |
|
Search time |
Moderate — measured, not assumed |
Two weeks of logging |
Surprisingly large at scale |
|
Recovery |
Weak — a counterfactual |
An hour |
Use as supporting, not lead |
Against that, put the total cost including labels, migration and configuration time. If the case only works on the recovery number, it's a weak case and I'd reconsider whether you have the problem you think you have.
A worked example, so the shape is clear
A 180-person organisation, roughly 1,400 tracked items across four sites. The numbers below are illustrative, but the method is the point.
|
Input |
Basis |
Annual value |
|---|---|---|
|
Missed cover claims |
6 repairs last year that were under warranty or AMC |
£7,200 |
|
Duplicate purchasing |
11 items bought that already existed elsewhere |
£4,800 |
|
Search time |
Measured: 3.2 hrs/person/month across 40 people at £28 |
£43,000 |
|
Recovery |
£19k written off, 40% reduction assumed |
£7,600 |
|
Gross benefit |
|
£62,600 |
|
Software, labels, migration |
Three-year total, annualised |
−£9,400 |
|
Net |
|
£53,200 |
Notice which row dominates. Search time, measured rather than assumed, is usually several times the recovery figure — and it's the one nobody puts in the business case because it doesn't appear on an invoice.
Notice also that recovery, the number everyone leads with, is the smallest quantified input and the weakest evidenced.
Presenting it without losing the room
- Lead with the measured inputs, not the modelled ones. Invoices and logged hours first; counterfactuals last.
- State your baseline's weakness openly. Loss data understates the problem because unregistered items aren't in it. Saying so pre-empts the obvious challenge and makes everything else more credible.
- Give a payback period, not a percentage. "Pays back in five months" survives scrutiny better than an ROI figure nobody can reconstruct.
- Separate the unquantified risks — audit readiness, insurance position — into their own section. Mixing them into the total is what gets a business case picked apart.
Measuring after the fact, which almost nobody does
A business case is a prediction. Measuring the outcome is what makes the next one credible, and it is the step that gets skipped once the money is approved.
- Baseline before you start. Discrepancy rate from a real count, last year's repair spend, and a two-week search-time log. Thirty minutes of work that you cannot reconstruct later.
- Re-measure at six months using the same method on a fresh sample.
- Report the difference, including anything that did not improve. A report claiming everything worked is trusted less than one that says two of four things did.
- Adjust. If the discrepancy rate has not moved, the problem is that recording a movement is still harder than skipping it — not the software.
The numbers that move first
|
Metric |
When it moves |
Why |
|---|---|---|
|
Claim capture rate |
Immediately |
Expiry reminders work from day one |
|
Search time |
Within weeks |
Requires labels and accurate locations, not culture change |
|
Duplicate purchasing |
Next purchase cycle |
Depends on people checking before buying |
|
Loss rate |
Two counts, so 6–12 months |
Needs a full cycle to be visible at all |
|
Audit readiness |
First count |
Binary — you can produce the document or you cannot |
Set expectations against that table when you present. A sponsor expecting the loss rate to move in month two will conclude it failed, when what is actually true is that it is too early to tell.
What weakens a case
- Industry benchmarks in place of your own numbers. Everyone recognises a statistic lifted from a vendor page.
- Loss reduction assumptions above 50%. Defensible is better than impressive.
- Soft benefits presented as hard ones. "Improved visibility" is not a line item.
- Omitting implementation effort. Finance knows nothing is free, and the omission costs you more credibility than the number would have.
The benefit you can't model, and should still mention
Audit readiness. The value of being able to produce a verified count on request is impossible to quantify until the request arrives, at which point it's the only number that matters.
Don't try to put a figure on it. Put it in as a risk reduction, name who has asked you for one before, and let it sit alongside the quantified inputs rather than pretending it's one of them.
Key takeaways
- Lead with missed warranty and contract claims — actual invoices are the strongest evidence you have.
- Duplicate purchasing is easy to find by cross-referencing purchases against the register.
- Measure search time for two weeks rather than quoting an industry statistic.
- Put recovery last and be conservative; it's a counterfactual built on incomplete loss data.
- State the weakness in your baseline openly — it makes the whole model more credible.
Frequently asked questions
How do you calculate ROI on asset tracking software?
Total four benefits — recovered equipment, warranty and contract claims you stop missing, time no longer spent searching, and duplicate purchases avoided — then subtract total cost including labels, migration and configuration time. Express it as payback period rather than a percentage; it's easier to defend.
What is a realistic payback period for asset tracking?
Where the problem is real, most organisations model somewhere between three and twelve months, and the variance is driven almost entirely by how much they were losing to missed cover claims and duplicate purchasing rather than by the software's price.
How do I prove equipment loss to finance?
Run a scoped physical count of one location or category. The discrepancy rate is measured rather than estimated, and it's far more persuasive than any projection. It also gives you a baseline to measure improvement against later.
Is asset tracking software worth it for a small team?
If equipment changes hands and somebody would struggle to say who has what, usually yes — but start on a free tier and prove it before paying. If nothing is issued to anyone and items live in one place, a spreadsheet may genuinely be enough.
What's the biggest hidden benefit?
Audit readiness. You can't model the value of producing a verified count on demand until somebody asks for one. Include it as a risk reduction alongside the quantified inputs rather than trying to price it.
How long does asset tracking take to pay back?
Where the problem is real, most models land between three and twelve months. The variance comes almost entirely from how much was being lost to missed cover claims and duplicate purchasing, not from the software price — which is why those are the inputs worth measuring first.
What if we cannot measure our current losses?
That is itself the finding, and a persuasive one. Run a scoped count of one location, and the discrepancy rate gives you a measured baseline in an afternoon. It is far more convincing than a projection and it gives you something to measure improvement against.
Should the business case include staff time saved?
Yes, but measure it rather than assuming it. Ask five people to log every search for two weeks. Two weeks of your own data survives a finance review; an industry benchmark lifted from a vendor page does not.