TL;DR
- Unreturned equipment is hardware issued to somebody who left and never came back. Most companies have some, almost none can say how much, and nobody owns the number.
- If you have had three leavers in two years and all three handed their laptop over in person, this is not your problem. Do not go looking for a business case you do not need.
- Four separate costs sit behind each missing device, and the hardware is the smallest of them. The others are the replacement you had to buy, the data you cannot prove was destroyed, and the record you can no longer defend.
- The reason the number is invisible is accounting rather than carelessness: the loss is absorbed into depreciation and nobody is asked to report it.
- Producing the figure takes an afternoon and is the only thing that reliably releases budget to fix the cause.
- The number is almost always larger than the team's estimate, and the gap between them is itself the finding.
The Write-Off That Was Never Written Off
A 400-person company was preparing for a funding round. During diligence an analyst asked a routine question about fixed assets and the finance team produced a schedule showing laptops purchased over four years, depreciated on a three-year straight line, with a small residual carrying value.
The analyst asked a second question. How many of those devices does the company physically hold today. Nobody knew. A week of work produced a rough answer: of 612 machines bought in that period, the company could account for 538. Seventy-four were unaccounted for, mostly issued to people who had since left. At an average purchase price of just over 1,100 pounds, that was around 81,000 pounds of hardware that had quietly left the business.
What made it awkward was not the money. It was that the figure had never appeared anywhere. Each device had depreciated to zero on schedule, the schedule had never been reconciled against reality, and no single quarter had ever shown a loss. Unreturned equipment does not show up as a cost because depreciation hides it perfectly: the asset was always going to reach zero, and nobody asks whether it reached zero in your possession or in somebody's flat.
Best tools for Device Management
This is why the number is worth producing deliberately, and why almost nobody has.
When You Don't Actually Need to Count This
When the manual way is genuinely fine
Small team, one office, leavers who hand hardware over on their last day. If you can name every device you have issued and where it is, the exercise will confirm what you already know and cost you an afternoon. Spend it elsewhere.
When friction starts appearing
The signal is a guess. Somebody asks how many laptops are missing and the answer is a shrug with a range attached. A range is fine as an answer to a casual question and is not fine as an input to a decision about whether to spend money fixing the cause.
When it becomes a liability
Once an outside party asks. Diligence, an insurer, an auditor, or a customer security review. Each wants a defensible number, and the honest absence of one is a finding in itself. This is also the moment the exercise stops being optional and starts being urgent, which is the worst time to do it.
The edge case that forces it
A redundancy round. Several leavers at once, often remote, often without a warm relationship, is the scenario that converts a slow leak into a visible loss. Companies that have been through one tend to produce this number afterwards, having not produced it before.
Five Questions Finance Asks
"How many are missing?" A count, not a proportion. The proportion is useful later and the count is what gets attention.
"What did they cost?" Purchase price, not written-down value, because the written-down value is the thing hiding the problem.
"Is this in our accounts anywhere?" Almost certainly not as a loss. That is the point.
"Are we insured for it?" Possibly, and a policy requiring you to demonstrate reasonable recovery efforts is a policy you may not be able to claim under.
"What would fixing it cost?" Less than the leak, usually, and the comparison is only possible once the leak has a number.
The Four Costs Behind One Missing Laptop
Counting the hardware alone understates this by a wide margin. Each unreturned device carries four separate costs and they are not all financial.
The asset you no longer hold. Purchase price, less whatever it was genuinely worth at the point it went missing. A machine lost in year one is a far larger loss than one lost in year four, which is why the age profile of your missing devices matters as much as the count.
The replacement you had to buy. This is the cost people forget, and it is frequently larger than the first. A device that comes back and is reissued avoids a purchase. A device that does not come back means the next starter needed a new machine. The missing laptop did not only cost you its own value, it cost you the purchase it would have prevented.
The data you cannot prove was destroyed. You cannot evidence the erasure of a device you do not hold. For most companies this is a gap in a record rather than an incident, and it becomes material the moment somebody asks the question formally. It is also the one cost that does not diminish with the age of the device.
The record you can no longer defend. Once a meaningful share of your register is wrong, every number derived from it is soft: the insurance schedule, the security questionnaire answer, the asset note in a diligence pack. The cost here is the credibility of the data rather than the data itself.
| Cost | How to value it | Grows with |
|---|---|---|
| Asset lost | Purchase price less genuine residual at the time | Devices lost early in life |
| Replacement bought | Full price of the machine you bought instead | Every single occurrence |
| Unprovable destruction | Not financial until somebody asks | Number of devices, not their value |
| Register credibility | Not financial until diligence | Proportion of records wrong |
So the rule of thumb that holds up: the real cost of an unreturned laptop is roughly the purchase price plus the replacement, which means the arithmetic most companies do understates it by about half before you have considered the two non-financial costs at all.
Producing the Number in an Afternoon
The exercise is not difficult and it is almost never done, because it has no natural owner. IT assumes finance tracks it and finance assumes IT does.
Start with leavers, not with devices. Pull everybody who left in the last 24 months from your HR system. This list is reliable, which is more than can be said for the asset register, and it is the correct denominator.
For each leaver, answer three questions. Were they issued a company device. Did it come back. If it came back, was it reissued or is it sitting somewhere. You will not be able to answer all three for everybody, and the proportion you cannot answer is a finding in its own right.
Value what is missing at purchase price. Not written-down value. The written-down value is an accounting construct that is specifically designed to reach zero, and using it makes the loss disappear, which is how the loss disappeared in the first place.
Add the replacement purchases. For each missing device, did you buy a machine you would not otherwise have needed. In a growing company the honest answer is usually yes.
Then produce three numbers. Devices unaccounted for, their purchase value, and the replacement spend attributable to them. The third number is the one that moves a budget conversation, because it is money that actually left the business in a quarter somebody can point at.
| Finding | Typical interpretation | What it supports |
|---|---|---|
| Under 5 per cent unaccounted | Normal leakage, process is working | No action beyond monitoring |
| 5 to 15 per cent | A recovery process that depends on memory | A case for tooling or an owner |
| Over 15 per cent | No process, or no ownership of it | Fix ownership before buying anything |
| Cannot determine for many leavers | A register problem beneath the recovery problem | Rebuild the register first |
And write down the date you produced it. This number only has value as a series, and the second measurement, taken a year later, is what tells you whether anything you did worked.
The Age Profile, and Why It Changes the Answer
The count tells you how many. The age profile tells you what to do, and it is the second cut almost nobody takes.
Sort your missing devices by how long they had been in service when the person left. The shape of that distribution points at a different cause and a different fix.
Mostly devices lost in year one. These are expensive, because the machine had most of its life ahead of it, and they usually indicate a problem with short-tenure leavers. People who leave within twelve months are the hardest group to recover from: the relationship is thin, the offboarding is often abrupt, and nobody feels much obligation in either direction. If your losses cluster here, the fix is in how quickly and warmly the request goes out, not in logistics.
Mostly devices lost in year three or four. Lower value, and usually a different cause: the machine was old, nobody judged it worth the effort of retrieving, and the decision was made informally by somebody who had no authority to make it. This is less a recovery failure than an absent policy on what is worth recovering, and the honest response may be to write that policy rather than to chase harder.
Evenly spread. The most common pattern, and it means the step simply is not happening regardless of circumstances. No targeted intervention will help; this is the ownership problem.
| Age at loss | What it costs you | Likely cause | Fix |
|---|---|---|---|
| Under 12 months | Most of the purchase price | Short-tenure leavers, thin relationship | Faster, warmer first contact |
| 12 to 24 months | Substantial, worth reissuing | Process gaps | A triggered return |
| Over 36 months | Little as an asset, same as a data gap | Nobody judged it worth retrieving | A written threshold |
| Evenly spread | Varies | The step does not happen | Assign an owner |
And the age profile also settles an argument you will otherwise have. Somebody will suggest that old devices are not worth recovering, which is financially reasonable and ignores that the data obligation does not depreciate. A four-year-old laptop is worth almost nothing and still holds whatever was on it. If you decide not to pursue older hardware, decide it explicitly and write down how you handle the data question, rather than letting it be decided case by case by whoever is busy.
Where the Devices Actually Went
"Missing" covers four quite different situations, and separating them changes what you should do next. Running this split costs an extra half hour on top of the main exercise and it is the difference between a number and a diagnosis.
Never returned by the leaver. The classic case. The device went to somebody, they left, it stayed with them. A recovery arrangement addresses this directly and it is the category platforms are built for.
Returned, then lost internally. It came back, went into a cupboard, and nobody recorded it. This looks identical in the register and has nothing to do with recovery: it is a goods-in and triage problem, and buying a recovery platform will not touch it.
Deliberately left with the person. Somebody agreed, informally, that a leaver could keep a machine. Frequently nobody wrote it down, so it shows up as a loss. This is a policy and authority question, and the fix is stating who may make that decision and recording it when they do.
Never properly recorded in the first place. The device was bought outside the process, often on a card, and never entered the register. It may well still be in use by somebody currently employed. This is not a loss at all; it is a register gap presenting as one.
| Category | Looks like | Actually is | Right response |
|---|---|---|---|
| Never returned | A loss | A loss | Recovery process or platform |
| Returned, lost internally | A loss | A storage and triage gap | Goods-in process, not a vendor |
| Deliberately left | A loss | An undocumented decision | State who may authorise it |
| Never recorded | A loss | A register gap, possibly still in use | Reconcile against purchasing |
So split the list before you present it. A company that reports 70 missing devices and then discovers that 20 are in a cupboard, 9 were given away by a manager and 14 were never recorded has a very different problem from one where all 70 walked out of the door, and only the first category justifies the purchase most people reach for.
Why the Number Never Appears on Its Own
Four structural reasons, and understanding them is what stops the exercise being repeated and then forgotten again.
Depreciation absorbs it. A three-year straight line takes every laptop to zero whether or not you still have it. No ledger entry ever asks about physical possession, so the loss never surfaces as a loss.
It is distributed across quarters. One or two devices at a time is beneath anybody's threshold for investigation. The same total arriving in a single quarter would trigger a review, which is why the slow version persists for years.
Nobody owns the question. It sits between IT, finance and People. In most companies the only person who would naturally produce it is whoever maintains the asset register, and that person is usually already aware the register is incomplete.
And the register is the thing being measured. Asking an incomplete system how incomplete it is produces an optimistic answer. This is why the exercise above starts from the leaver list in HR rather than from the asset system, and that single methodological choice is what makes the number credible.
What to do:
- Run the exercise from the HR leaver list, never from the asset register.
- Value at purchase price so the figure is not pre-absorbed by depreciation.
- Include the replacement spend, which is the number finance responds to.
- Give the figure a named owner and a review date before you present it.
- Present the three numbers together, because any one alone is easy to dismiss.
Turning the Number Into a Decision
A figure with no comparison is trivia. Two comparisons make it actionable.
Against the cost of fixing the cause. Most recovery arrangements cost meaningfully less than the leak they prevent, once replacement purchases are counted. If the leak is small, that comparison honestly favours doing nothing, and that is a legitimate outcome of the exercise rather than a failure of it.
Against your own trend. The second measurement matters more than the first. A company moving from 11 per cent to 4 per cent has evidence that something worked. A company that produced the number once, acted, and never measured again has a belief.
The common mistake is presenting the loss and asking for budget in the same meeting. The stronger sequence is to present the number, agree that it will be measured again in twelve months, and then propose a response. It turns a request into a programme with a success criterion, which is a considerably easier thing to approve.
Presenting It So It Gets Funded
The exercise fails more often at the presentation than at the arithmetic. A number delivered badly reads as a complaint about other people's processes, and gets handled accordingly.
Take it to finance first, not to your own leadership. Finance has the depreciation schedule and the purchasing records, and bringing them a reconciliation they could not previously produce makes them an ally rather than an audience. A number that arrives jointly from IT and finance is treated very differently from one that arrives from IT alone.
Lead with the replacement spend. Devices unaccounted for is interesting and easy to argue with. Money spent on machines you would not have needed is concrete, recent, and sits in a budget somebody recognises. Open with that and the other two numbers land as supporting detail.
Show the split, not just the total. The four categories above. Volunteering that a quarter of your apparent losses are actually a cupboard problem establishes that the number is honest, which is what makes the remainder credible.
Name what you are not asking for yet. Present the finding, propose the second measurement date, and explicitly say that a spending decision comes after the split is understood. Asking for budget in the same breath as reporting a loss invites the response that IT should have been doing this anyway.
And own the part that is yours. If the register was never maintained, say so plainly and early. The alternative is somebody else noticing, which turns the whole exercise into a discussion about accountability rather than about the fix.
| Element | Include | Leave out |
|---|---|---|
| Opening figure | Replacement spend, last 24 months | A percentage |
| Supporting | Count and purchase value | Vendor quotes |
| Honesty marker | The four-way split | Caveats about data quality |
| Ask | A date for the next measurement | A budget request |
So keep the first conversation to one page and one quarter's worth of consequence. The budget conversation is easier a month later, once the number has been absorbed and nobody is hearing it for the first time while also being asked to approve something.
How to Choose: Five Questions Before You Spend Anything
Is the leak large enough to justify a vendor? Compare the annual loss, including replacements, against what a recovery arrangement would cost. If the leak is three devices a year, the honest answer is a process change and a named owner rather than a platform.
Where are the missing devices? If they are concentrated in one country, the problem is that country rather than your process, and a regional specialist solves it more cheaply than a global arrangement.
Is the real problem recovery or the register? If you could not establish what was issued for a large share of leavers, no recovery tool will help, because it will be pointed at an incomplete list. That is the register problem wearing a recovery costume and it is cheaper to fix.
Who will own the number after today? The exercise has no value as a one-off. Name the person and put the next measurement in a calendar before the current one is presented.
What does your insurer require? Some policies expect demonstrable recovery efforts. Worth reading before assuming unreturned hardware is covered, and worth asking your broker rather than inferring from the wording.
Six Options Worth Knowing
These are the arrangements that address the cause once you have the number. Each was checked against its own pricing page on 6 and 7 October 2026 and none publishes a figure, which means the cost side of your comparison requires a sales conversation while the loss side you can calculate yourself in an afternoon. Two describe a model without numbers: GroWrk names a per-order option alongside a subscription tier, and allwhere refers to pay-as-you-go and fixed rates.
RemoAsset
Disclosure: RemoAsset is owned by the same people who publish HROpsLab. It appears here because it addresses the cause this article measures, and its limitations are stated in the same detail as everything else on this page.
Best for: companies whose losses come from devices issued to remote staff who then leave, which is the dominant pattern behind most of these numbers.
Why companies choose it: the device is recorded against a person and address from purchase, and offboarding triggers the return automatically rather than depending on somebody remembering. Recovered machines can be stored and reissued, which addresses the replacement cost as well as the asset cost, and the replacement cost is usually the larger half.
Where it struggles: it publishes no price and requires a demo, so the cost side of your comparison takes longer than the loss side. It works from the register it holds, so devices bought before adoption need importing and then carry the same drift as any other record. It is not an MDM and not a certified disposition vendor, so neither device control nor end-of-life certification comes with it.
Workwize
Best for: multi-region fleets where storing recovered devices regionally makes reissue viable.
Why companies choose it: regional warehousing is what converts a recovered laptop into an avoided purchase, which is the part of the loss most worth attacking.
Where it struggles: no published price, and more capability than a two-country company needs.
Deel IT
Best for: companies already employing internationally through Deel.
Why companies choose it: the leaver event already exists in the system holding the employment relationship, which is where recovery processes usually fail.
Where it struggles: quote-based, and most compelling alongside an existing Deel relationship.
Firstbase
Best for: companies whose losses include furniture and peripherals as well as machines.
Why companies choose it: broader scope than laptops, which matters if your unreturned list includes monitors and desks.
Where it struggles: no published price, and furniture is rarely worth recovering, so the broader scope may not reduce your number.
GroWrk
Best for: losses concentrated in Latin America or parts of Asia.
Why companies choose it: in-country presence where others subcontract, which decides whether a retrieval happens at all in those markets.
Where it struggles: publishes nothing, naming its models without attaching figures.
A named owner and a written process
Best for: companies whose leak is small, or whose register is the actual problem.
Why companies choose it: it costs nothing, and for a company losing a handful of devices a year it will outperform any platform on a cost basis. Prompt contact, prepaid packaging and somebody who notices on day ten resolves most of the gap.
Where it struggles: it does not scale past a certain volume or geographic spread, and it depends entirely on one person continuing to care. The failure mode is that the owner changes role and the process degrades silently.
What Each One Published
| Option | Published price | Unit | Addresses |
|---|---|---|---|
| A named owner and a process | Nothing | n/a | Small leaks, register problems |
| RemoAsset | Not published, demo required | n/a | Automated return, storage, reissue |
| Workwize | Not published | n/a | Regional storage and reissue |
| Deel IT | Not published | n/a | Leaver-triggered recovery |
| Firstbase | Not published | n/a | Full setup recovery |
| GroWrk | Not published, models named only | n/a | Hard-to-reach markets |
Checked against each vendor's own page on 6 and 7 October 2026.
The Decision Table
| Situation | Scale | Setup | Primary Pain | Recommended Starting Point |
|---|---|---|---|---|
| 3 leavers in 2 years, all handed devices over | Under 40 | Office | Nothing is leaking | Do not run the exercise |
| Cannot say how many are missing | Any | Any | No number exists | Run the exercise from the leaver list |
| Under 5 per cent unaccounted | Any | Any | Normal leakage | Monitor annually, change nothing |
| 5 to 15 per cent, devices in homes | 50 to 500 | Remote | Recovery depends on memory | A recovery platform |
| Over 15 per cent | Any | Any | Nobody owns the step | Assign an owner before buying |
| Cannot establish what was issued | Any | Any | The register, not recovery | Rebuild from purchasing first |
| Losses concentrated in one country | Any | Remote | Coverage in that market | A regional specialist |
| Diligence or an insurer has asked | Any | Any | A defensible figure, quickly | Run the exercise this week |
The row most companies land in is the third or fourth, and the distinction between them is worth the afternoon it takes to establish, because one of them justifies spending money and the other does not.
What Getting This Wrong Costs
The expensive version of getting this wrong is not losing laptops. It is never producing the number, and therefore never being able to argue for fixing the cause. A loss that is invisible cannot be compared against the cost of preventing it, so the comparison never happens, and the leak continues for years because no individual quarter ever looked bad enough to investigate.
The second cost arrives during diligence, and it is about confidence rather than hardware. An acquirer or an investor asking a simple question about fixed assets and receiving a vague answer draws a conclusion about operational discipline that extends well beyond laptops. The hardware number is small. The impression is not, and it is formed in the room.
The third is the insurance position. Policies vary, and some expect a business to have taken reasonable steps to recover its property. A company with no recovery process, no record of attempts and no figure may find that cover is less useful than assumed. That is a question for your broker rather than for this page, and it is worth asking before you need the answer.
So ask the diagnostic question plainly. Is the problem that devices leave, that nobody chases them, or that you never knew what you had? Devices leaving is normal and is addressed by a recovery arrangement. Nobody chasing is an ownership problem that software will not fix. And not knowing what you had is a register problem that sits underneath both, and it has to be solved first or everything built on top inherits it.
When You're Ready to Act on It
The signals are simple. Somebody asked how many devices are missing and nobody could answer. A diligence process, an insurer or an auditor is coming. Or you have just run a redundancy round and the usual slow leak arrived all at once.
Run the exercise first and present three numbers: devices unaccounted for, their purchase value, and the replacement spend they caused. Then agree when it will be measured again, before proposing any response. That sequence turns the finding into a programme rather than a complaint, and it is the version that gets funded.
If the cause is devices issued to remote staff who then leave, which is what most of these numbers turn out to be, RemoAsset addresses it by triggering the return from the offboarding event and holding recovered machines for reissue, which attacks the replacement cost as well as the asset. It publishes no price, so expect the comparison to take a demo. And if your leak is three devices a year, the honest recommendation is a named owner and prepaid packaging, which costs nothing and will serve you better.
Whatever you choose, put the second measurement in the calendar. A number produced once is a story. A number produced twice is evidence.
And if the exercise comes back showing that very little is missing, publish that too. Teams are quick to report problems and slow to report the absence of one, which means the only numbers anybody ever sees are bad ones. A clean result is worth circulating precisely because it establishes that somebody checked, which is the thing an auditor, an insurer or an acquirer is really asking about.
Frequently Asked Questions
What counts as unreturned equipment?
Unreturned equipment is any company-owned hardware issued to somebody who has since left and which the business no longer holds. In practice it is dominated by laptops, with phones and monitors making up most of the remainder. It is distinct from equipment that was returned and then lost internally, which is a storage and register problem rather than a recovery one, and distinct again from equipment deliberately left with a departing employee as part of an agreed arrangement. Keeping those three categories separate matters, because they have different causes and only the first is addressed by a recovery process.
How much does an unreturned laptop actually cost?
More than its purchase price, and the usual approximation is roughly double. The device itself is the first cost, valued at what you paid rather than at its written-down value, because depreciation is precisely what hides this. The second and frequently larger cost is the replacement machine you had to buy for the next starter, which a recovered and reissued device would have avoided. Two further costs are real and not financial until somebody asks: you cannot evidence that data on a device you do not hold was destroyed, and a register with a material error rate undermines every figure derived from it.
What is a normal rate of unreturned equipment?
Under 5 per cent of devices issued to leavers is typical for companies with a working process, and that level is usually best treated as normal leakage rather than as a problem to spend money on. Between 5 and 15 per cent suggests a process that depends on somebody remembering, which is the point at which tooling or a named owner starts paying for itself. Above 15 per cent generally means nobody owns the step at all. The more useful comparison is against your own figure a year later rather than against any benchmark, because the composition of fleets and the geography of staff vary too much for cross-company comparison to mean much.
Why does unreturned equipment never show up in our accounts?
Because depreciation absorbs it completely. A laptop on a three-year straight line reaches zero carrying value whether it is sitting on somebody's desk or in a former employee's flat, and no routine accounting process asks which. The loss is also distributed, arriving one or two devices at a time, which keeps it below the threshold that would trigger an investigation if the same total appeared in a single quarter. Add to that the absence of a natural owner, since the question sits between IT, finance and People, and the result is a cost that is entirely real and structurally invisible.
How do we calculate how much equipment we have lost?
Start from the leaver list in your HR system rather than from the asset register, because the register is the thing whose accuracy is in question and asking it how incomplete it is produces a flattering answer. For every person who left in the last 24 months, establish whether they were issued a device, whether it came back, and whether it was reissued. Value anything missing at purchase price, then add the replacement machines you bought as a consequence. Present three numbers together: the count, the purchase value, and the replacement spend. The third is the one that gets attention, because it is money that visibly left the business.
Can we claim unreturned equipment on insurance?
Sometimes, and it depends entirely on your policy, so it is a question for your broker rather than one with a general answer. The point worth knowing in advance is that some policies expect a business to have taken reasonable steps to recover its property, which means a company with no documented recovery process, no record of contact attempts and no figure may be in a weaker position than it assumes. Establishing what your cover actually requires is worth doing before you need to rely on it, and the answer may itself be an argument for keeping records of recovery attempts.
Should we pursue former employees for unreturned laptops?
What you can do, and whether any financial recovery is permissible, depends on your jurisdiction, on the terms the person agreed to and on the circumstances, and that is a question for your own legal advice. What the evidence consistently suggests operationally is that pursuit is the least effective lever available. The companies with the lowest loss rates are not the ones that chase hardest; they are the ones that ask promptly while the relationship is still intact, pay for the return themselves, and make complying easier than ignoring the request. Treating recovery as an enforcement problem produces worse results than treating it as a convenience problem.
How often should we measure this?
Annually is sufficient for most companies, and the second measurement matters more than the first because the trend is the only thing that demonstrates whether a change worked. Run it more frequently if you have just implemented a recovery arrangement, where a measurement at six months will tell you early whether it is working, or after any event that produces a cluster of departures. Put the next date in a calendar at the moment you present the current figure, because the common failure is producing the number once, acting on it, and never establishing whether the action helped.
HROpsLab takes no vendor money and publishes no paid placements, which is why one of the options on this page costs nothing and is recommended for the companies it suits.