TL;DR
- Most equipment cost per hire figures count the laptop and nothing else, which understates the real number by roughly half.
- Six cost lines belong in it: the device, the peripherals, shipping, setup labour, software seats attached at provisioning, and the recovery cost at the other end.
- The recovery cost is the one always missing, and it is the only line that varies by country rather than by role.
- Build it per role band, not as a single company average. A single average is useless for both budgeting and approval decisions.
- The figure is for forecasting and for pricing a headcount plan. It is not a target to minimise.
- Amortise the device over its refresh cycle if you want a yearly cost. Use the full purchase price if you want a cash forecast. Do not mix them.
The Number That Was Wrong By Half
The finance business partner asked a reasonable question during planning. We are hiring 60 people next year, what is the equipment cost. The answer came back in a day: about £1,300 a head, so roughly £78,000.
The actual spend ran closer to £2,400 a head. The £1,300 was a laptop and a monitor, which is what the asset register could tell you in an afternoon. It did not include the dock, the keyboard and mouse, the chair budget for the people who claimed it, the courier to seven countries, the two hours of setup labour per device, the software seats that got provisioned at the same time, or the return shipping and wipe at the other end for the people who left within the year.
None of those were hidden. They sat in six different cost centres, and nobody had ever added them up, because no single person owned the total.
Best tools for HR Operations
The real issue is not that the first number was wrong. It is that a number produced in an afternoon from the easiest available source becomes the planning assumption, and then every conversation for the next year is anchored on it.
The Six Lines That Belong in Equipment Cost Per Hire
Each of these is a real cash cost attached to one new person starting. Leave any of them out and the figure is wrong in a direction that always favours underestimating.
The device itself. The laptop, and increasingly a second screen as standard rather than as a request. This is the line everybody already counts.
Peripherals and accessories. Dock or hub, keyboard, mouse, headset, cables, laptop stand. Individually small, collectively a meaningful fraction of the device cost, and almost always issued as a matter of course rather than on request.
Shipping, both ways. Outbound to wherever the person is, which for a distributed team means a courier and sometimes customs handling. And the return leg, which belongs in this number even though it happens later, because it is caused by the hire.
Setup labour. Imaging or enrolling the device, configuring it, checking it, packing it. Measured in hours of somebody's time, and real money whether that somebody is internal or a vendor.
Software seats provisioned at the same time. Not the whole stack, which is a separate budget, but the seats that are added because this specific person arrived and that would not exist otherwise.
Recovery and disposition. Return shipping, the wipe, the inspection, and either the resale credit or the disposal cost. Negative in some cases, which is fine. Missing in almost every calculation, which is not.
| Cost line | Typically counted | Why it gets missed |
|---|---|---|
| Device | Yes | It is the one thing the asset register knows |
| Peripherals | Sometimes | Bought separately, often on a different code |
| Outbound shipping | Rarely | Sits in a logistics or facilities budget |
| Setup labour | Rarely | It is somebody's time, not an invoice |
| Software seats at provisioning | No | Lives in an IT software budget, not a hire cost |
| Recovery and disposition | Almost never | Happens months later, so it is not linked to the hire |
Why Recovery Belongs in the Hire Cost
This is the line people argue about, so it is worth making the case properly.
The cost is caused by the hire. If you equip somebody, you will eventually recover or write off that equipment. The recovery is not a separate event that happens to the company. It is the second half of the same transaction, and excluding it means your cost per hire is really a cost per issue.
It varies by country in a way nothing else in the list does. Device and peripheral prices are broadly comparable across markets. Return logistics are not. Recovering a laptop from a market where you have no entity, no office and no local partner can cost a multiple of what the same recovery costs domestically, and sometimes it does not happen at all.
Excluding it makes international hiring look cheaper than it is, which is exactly the decision the number is used to inform. A company comparing a hire in its home market against one three time zones away is making a worse decision with recovery excluded, because that is where the difference actually lives.
What to do: carry recovery as a line even where you cannot price it precisely. A banded estimate by region, reviewed annually against what you actually paid, is far better than a zero. We looked at how the recovery side behaves in practice in IT asset recovery for small teams.
Build It Per Role Band, Not As One Average
A single company-wide average is the most common form of this number and the least useful.
Role bands differ by a factor of two or more. An engineer on a high-specification machine with two external screens and a dock is a different cost from a support hire on a standard laptop. Averaging them produces a figure that is wrong for both and right for nobody.
Three or four bands is usually enough. Standard, high-specification, field or customer-facing, and executive if that is a real category at your company. More bands than that and you are modelling individuals rather than budgeting.
Band by equipment need, not by seniority. This matters. Seniority correlates with equipment spend but does not determine it, and banding by level produces the wrong answer for a senior person on a standard machine and a junior engineer on a workstation.
And carry a location dimension separately. Resist the temptation to multiply bands by countries, which produces a matrix nobody maintains. Hold the band as the base cost and the location as a shipping and recovery adjustment on top.
| Band | What drives the cost | Planning note |
|---|---|---|
| Standard | One laptop, one screen, standard peripherals | The majority of hires at most companies |
| High specification | Higher-spec machine, second screen, dock | Engineering, design, data roles |
| Field or customer-facing | Portability, accessories, sometimes a phone | Shipping and damage exposure both higher |
| Executive | Specification plus accessories by expectation | Small volume, so a rough figure is fine |
What to Do With the Number Once You Have It
Three uses, and they need the figure built differently, which is where most calculations fall over.
Headcount planning and budgeting. You want cash in the period, so use full purchase prices and include the outbound shipping and setup. Recovery for this year's leavers, not this year's joiners, which is a subtlety worth getting right because the two populations differ.
Approving an individual request. You want the band figure and the exception rule, not the company average. The useful question at approval time is whether this request sits inside its band or outside it, which is a conversation we set out in full in equipment approval workflow.
Comparing hiring locations. You want the base band plus the location adjustment, and here the recovery line does most of the work. This is the use case that justifies carrying recovery at all.
What to do: state which of the three a given figure is for, on the figure itself. The same calculation used for all three without adjustment is the reason finance and HR produce different numbers from the same data and each believe the other is wrong.
Amortised or Full Cost, and Why Mixing Them Breaks Everything
A device lasts several years. Whether to divide its cost by that lifespan depends entirely on what question you are answering, and the common error is answering both at once.
Full purchase price is what you need for a cash forecast. You are buying 60 laptops next year and you will pay for 60 laptops next year, regardless of how long they last.
Amortised over the refresh cycle is what you need for a cost-per-employee-per-year figure, or for comparing owning against a device-as-a-service arrangement where the vendor's monthly fee is already spread.
Mixing them is the actual failure. A table that shows the device amortised over three years and the shipping at full cost produces a per-year figure that is too low on devices and too high on logistics, and it is almost impossible to spot by reading the output. It only shows up when somebody reconciles against actual spend and the variance makes no sense.
What to do: build one number at full cost and derive the amortised version from it with the arithmetic written out. Never maintain the two independently, because they will drift and the drift is invisible.
A Worked Calculation
Arithmetic makes this concrete. The figures below are illustrative shapes, not benchmarks, and you should replace every one with your own actuals.
Standard band, domestic hire. Laptop plus one screen plus peripherals as the device and accessory base. Add outbound shipping at the courier rate you actually pay. Add two hours of setup labour at your internal cost rate. Add the software seats provisioned on day one. Add the return leg and the wipe. The total typically lands between 1.6 and 2 times the laptop price alone, which is the headline finding of doing this exercise at all.
Same band, a hire three time zones away with no local entity. Everything above, plus a higher outbound courier cost, possibly customs handling, and a recovery line that may be several times the domestic figure. The device and accessory base has not changed. The total has, materially.
High specification band, domestic. The device and accessory base rises substantially. Shipping, setup and recovery barely move, because a more expensive laptop ships in the same box. So the ratio between total cost and device cost falls, which is counterintuitive and worth knowing when you are explaining why the expensive band looks relatively efficient.
What this tells you. The fixed costs in equipping somebody are a bigger share of the total at the cheap end, and the variable costs are dominated by geography rather than by specification. That is the single most useful output of building this number properly, and a laptop-only figure cannot show it.
Cost Per Hire Versus Total Cost of Ownership
These are two different numbers that get conflated constantly, and the conflation produces arguments that cannot be resolved because the two sides are computing different things.
Cost per hire is an event cost. It answers what it costs to get one new person working, and it is bounded by the provisioning event plus the eventual recovery of what you issued. It is the right number for a headcount plan, because headcount plans are lists of events.
Total cost of ownership is a stream. It answers what one equipped person costs per year on an ongoing basis, and it includes things cost per hire deliberately excludes: the support tickets, the repairs, the mid-life replacements, the ongoing software seats rather than only the day-one ones, the insurance if you carry it.
The overlap is the device, and that is where the confusion starts. Both numbers include a laptop. In cost per hire it appears at full purchase price as a one-off. In total cost of ownership it appears amortised across the refresh cycle alongside the running costs. Somebody comparing the two numbers sees the same laptop twice at different values and concludes one of the calculations is wrong.
Which one you need depends on the decision. Approving a hiring plan is a cost-per-hire question. Choosing between buying outright and a monthly device arrangement is a total-cost-of-ownership question, because the vendor's fee is a stream and you have to compare a stream to a stream. Setting a device standard sits in the middle and genuinely needs both.
What to do: name the number in the file name and in the title of the output. "Equipment cost per hire, FY planning, full cost" is a longer label than anybody wants and it prevents the single most common dispute in this area. Keep them in separate sheets, derived from one shared set of inputs, so the inputs cannot drift between them.
The Lines People Argue About
Three cost lines sit on the boundary, and reasonable people disagree about them. What matters is that you decide and record the decision rather than letting each calculation make its own call.
Replacement within the first year. Some proportion of devices fail, get damaged, or turn out to be wrong for the role, and a replacement is issued. Whether that belongs in cost per hire depends on whether you are forecasting cash or costing an event. For a cash forecast it does belong, as an expected rate rather than a per-person line, because you will pay for those replacements in the period. For an approval decision it does not, because the request in front of somebody is for one device.
Damage and loss. The same logic, with an extra complication: the handling of damage involves a conversation about responsibility that we deliberately treat as a separate subject in device damage liability, because it is a people question before it is a cost question. For the purposes of this number, carry an expected rate at the portfolio level and keep it out of the per-person band.
The chair, the desk and the home-office allowance. Sometimes these sit in an equipment budget and sometimes in a benefits budget, and it genuinely does not matter which as long as it is one of them consistently. What breaks is when they sit in benefits for budgeting and in equipment for a cost-per-hire exercise, so the same spend appears in two totals and the company believes it is spending more than it is.
What to do: write a one-page scope note saying which lines are in, which are out, and where the out ones live instead. Revisit it annually. The note is worth more than the precision of any individual figure, because it is what lets two people reproduce the same answer.
How to Present It So Finance Accepts It
A correct number that gets rejected is worth nothing, and these calculations get rejected for predictable reasons.
Show the lines, not the total. A single figure invites a challenge to the figure. Six lines invite a challenge to one line, which is a conversation you can have. It also means a disagreement about setup labour does not invalidate the other five.
Label every estimate as an estimate, with its basis. "Setup labour: 2 hours at internal cost rate, estimated from a sample of 12 provisioning tasks" is accepted. An unexplained number in the same cell is not, and the reason is reasonable rather than obstructive.
Reconcile against actual spend for one past period. This is the step that converts the exercise from a model into evidence. Take last year's hires, build the number as if forecasting, then compare against what was actually paid across the six budgets. The variance tells you which lines your estimates are weak on, and presenting the reconciliation alongside the forecast removes most of the objections before they are raised.
Give a range on the lines that deserve one. Shipping and recovery vary by destination, so a single number for them is false precision. A banded figure with the assumption stated is more credible, not less, and it protects you when the mix of hiring locations shifts.
What to do: present the reconciliation first and the forecast second. A model that has been tested against a past period is a different kind of artefact from one that has not, and the order of presentation is what communicates that.
The Refresh Cycle Interaction
Cost per hire and the device refresh cycle are connected in a way that is easy to miss and that distorts planning when it is missed.
A new hire does not always get a new device. If you run a refresh cycle, you have a pool of devices coming back from people who have left and from refreshes, some of which are perfectly good for a standard role. A hire equipped from that pool has a materially lower equipment cost, and a plan that assumes every hire buys a new laptop overstates the cash requirement.
But the pool is not free, and treating it as free is the opposite error. A device issued from the pool has a shorter remaining life, so it will need replacing sooner, which moves cost into a later period rather than removing it. The honest treatment is a lower cost now with the replacement date brought forward, not a zero.
The pool is also unreliable as a plan input. It depends on attrition and on refresh timing, neither of which you control precisely, so a hiring plan that depends on the pool covering 40 per cent of hires is a plan with a hidden assumption about attrition. Worth stating explicitly if you are going to rely on it.
What to do: forecast at full new-device cost and treat pool reissues as a favourable variance rather than a planning assumption. It is the conservative direction, and it is the one that does not produce an unfunded mid-year surprise when attrition runs lower than expected and the pool is empty.
What Changes at 500 People
The calculation does not change in shape at scale, but three things about it do.
The bands stop being estimates and become measurements. Below about 150 hires a year, your band figures are built from a handful of examples and sensibly carry a range. Above that you have enough volume that each band has a real distribution, and the useful output changes from an average to an average plus a spread. The spread is what tells you whether your standards are actually being followed.
Exceptions become a line of their own. At small scale an off-standard request is handled individually. At larger scale the exception rate is a number, and it is frequently the most interesting number in the whole exercise, because an exception rate above about one in six usually means the standard is wrong rather than that people are being unreasonable.
Geography starts to dominate. A company hiring across 15 markets has a shipping and recovery picture that varies more than its device picture, and at that point the location adjustment is doing more work than the band. That is the point where the recovery line stops being a rounding item and becomes the thing worth negotiating with a vendor about.
What to do: at scale, report the distribution rather than the average, and report the exception rate next to it. Two numbers that together tell you whether your equipment standard reflects what people actually need, which a single average cannot.
Where the Data Lives and Why It Is Hard to Assemble
The reason this calculation is rarely done is not analytical difficulty. It is that the six lines sit in six places.
The device and peripheral costs are in procurement records or the asset register, which is the easy part and is why the laptop-only figure exists.
Shipping is in a logistics or facilities budget, usually as an aggregate the finance team cannot split by person without effort.
Setup labour is in nobody's ledger. It has to be estimated from how long the work takes multiplied by a cost rate, and the estimate will be challenged, which is why people leave it out rather than defend it.
Software seats are in an IT software budget organised by application rather than by person, so attributing day-one seats to a hire takes a reconciliation.
Recovery is in whichever budget paid the courier, months after the hire, with no link back to the person who was equipped.
What to do: accept banded estimates for the lines you cannot measure precisely, and label them as estimates. A figure with four measured lines and two labelled estimates is defensible and useful. A figure with four measured lines and two zeros is neither, and it is the one most companies are planning with. We cover the register-side half of this in asset register HRIS integration, because the attribution problem is largely a data-linkage problem.
When You Don't Need This Number
Under about 30 hires a year. The total equipment spend is small enough that a laptop-price-times-headcount estimate is close enough and the error is not material to any decision.
Everybody is in one location and on one specification. The bands collapse to one, the shipping is a local delivery, and the recovery is somebody walking a laptop to a desk. The calculation adds precision nobody will use.
You are on a per-device monthly arrangement for everything. If a vendor charges you a single monthly fee per person that covers the device, the shipping and the recovery, your cost per hire is that fee times the term and the work is already done. That is one of the genuine advantages of those arrangements and it is rarely stated.
Nobody will act on it. If the headcount plan is fixed and the equipment standard is fixed, a more accurate cost figure changes nothing this year. Worth building before the next planning cycle rather than during this one.
The Tools, and What They Can Tell You
No product produces this number for you, and it is worth being clear about why before you evaluate anything.
Asset registers hold the device line and part of the peripheral line. AssetTiger publishes a straightforward ladder by asset count, with tiers at $15, $20, $40, $75, $140 and $275 per month, and it will tell you what you own and what it cost. Snipe-IT is free and open source self-hosted, with managed hosting published at $399.99 per year for the small tier, and its documented API makes it the easiest of these to query for a cost report.
Quote-based platforms hold more of the picture. Lansweeper, Device42 and Asset Panda each publish no readable price and route to a quote, and each covers discovery and lifecycle to a degree a spreadsheet does not.
Equipment lifecycle platforms hold the lines a register cannot. RemoAsset is one of these, aimed at distributed teams, and it is owned by the same people who publish this site, which you should weigh against everything said about it here. Because it handles the procurement, the shipping and the retrieval in one place, it holds the outbound logistics and the recovery line that are otherwise scattered across budgets, which is the part of this calculation that is hardest to assemble. Its limitations are real: pricing is quote-based with no published figure, so it cannot be compared on price from a web page, and it is a newer platform with a smaller footprint than the long-established ITAM names.
| What you need | Where it realistically comes from |
|---|---|
| Device and peripheral cost | Asset register or procurement records |
| Outbound shipping | Logistics budget, or a lifecycle platform if you use one |
| Setup labour | An estimate from measured time; no system holds it |
| Day-one software seats | IT software budget, reconciled by person |
| Recovery and disposition | Whoever paid the courier, or a lifecycle platform |
| The total | A spreadsheet, every time |
What to do: do not buy a tool to produce this figure. Build it once in a spreadsheet from whatever sources you have, find out which two lines are hardest to get, and let that inform the tooling question rather than the other way round.
The Mistakes That Produce a Wrong Number
Five specific errors, each of which has produced a materially wrong figure somewhere.
Using the list price rather than what you paid. If you buy through a reseller or on a volume agreement, the list price is not your cost, and the gap can be substantial. This error always overstates, which sounds harmless and is not, because an overstated figure gets challenged, loses credibility, and takes the correct lines down with it.
Counting the software stack rather than the day-one seats. The whole software budget divided by headcount is a per-employee software cost, which is a legitimate number and a different one. What belongs here is only the seats that exist because this person arrived. Using the former inflates cost per hire by an amount that has nothing to do with equipment.
Treating the second screen as an exception when it is the standard. Equipment standards drift upward quietly. A monitor that was a request two years ago is issued by default now, and a cost model built on the old standard understates every single hire. Check what is actually being issued rather than what the policy document says.
Averaging across years when prices moved. Device prices and courier rates both move, and a three-year average produces a figure that was right in the middle year and is wrong now. Use the most recent complete period and state which period it is.
Forgetting that recovery lags. The recovery cost you pay this year belongs to hires from previous years. For a cash forecast that is the right population. For a cost-per-hire figure attached to this year's joiners it is not, and conflating the two makes the number drift in a growing company, because a growing company always recovers fewer devices than it issues.
What to do: run the five checks above against whatever figure you currently plan with, before building anything new. In most cases at least two of them apply, and correcting them is faster than a rebuild.
The Decision Table
| Situation | Scale | Setup | Primary Pain | Recommended Starting Point |
|---|---|---|---|---|
| Under 30 hires a year | Under 100 | One location | Not material | Laptop price times headcount |
| Planning a hiring ramp | 100 plus | Mixed roles | Budget set on a wrong figure | Build three bands at full cost |
| Comparing hiring locations | Any | Distributed | Recovery excluded, so geography invisible | Band plus location adjustment |
| Approving individual requests | 150 plus | Standards exist | Average used where band is needed | Publish the band figures |
| Evaluating device-as-a-service | 200 plus | Owning today | Amortised and full cost mixed | Build full, derive amortised |
| Finance and HR disagree | Any | Two calculations | Different uses, same number | State the use on the figure |
| No recovery data at all | Any | International hires | Cost understated most where it matters | Banded regional estimate |
What Getting This Wrong Costs
The direct cost is a budget variance, and that is the least of it.
The real cost is that an understated equipment cost per hire makes distributed hiring look cheaper than it is, specifically in the markets where it is most expensive. A company that excludes recovery is systematically underpricing hires in exactly the places where recovery is hardest, and the decision it gets wrong is not an equipment decision. It is where to put a team.
The second cost is credibility. When the actual spend comes in at double the planning assumption, the conversation afterwards is not about equipment. It is about whether the numbers coming out of people operations can be relied on, and that is expensive to repair over a planning cycle or two.
And the third is that a laptop-only figure makes the wrong thing look like the lever. If equipment cost per hire is presented as the device price, every cost conversation becomes a conversation about buying cheaper laptops, which is the smallest available saving and the one with the most visible downside. The fixed costs and the geography are where the money actually is, and a number that hides both directs attention at the wrong half of the problem.
So the question is not what a laptop costs. It is whether your figure includes the five other lines, whether the two you cannot measure are labelled as estimates rather than carried as zeros, and whether it is built per band so that somebody can actually act on it.
When You're Ready to Build It
Start with the six lines and fill in what you can measure. Two of them will be estimates and that is fine, as long as they are labelled.
Build three or four bands rather than one average, band on equipment need rather than seniority, and keep location as an adjustment on top rather than a multiplier that creates a matrix nobody will maintain.
And state what each figure is for. Cash forecast, approval decision, or location comparison. The same number used for all three without adjustment is why two teams produce different answers from the same data and both think they are right.
Frequently Asked Questions
What should equipment cost per hire include?
Six lines: the device, the peripherals and accessories, outbound shipping, setup labour, the software seats provisioned because this person arrived, and the recovery cost at the other end including return shipping, the wipe and either resale credit or disposal. Most published figures include only the first and sometimes the second, which is why they typically understate the real number by something close to half.
Why does recovery belong in a hire cost?
Because it is caused by the hire and it is the second half of the same transaction. It also behaves differently from every other line: device and peripheral prices are broadly similar across markets, while recovering a laptop from a country where you have no entity or local partner can cost a multiple of the domestic figure. Excluding it makes international hiring look cheaper exactly where it is most expensive.
Should I use the full device price or amortise it?
Full price for a cash forecast, because you pay for the laptops in the year you buy them. Amortised over your refresh cycle for a cost-per-employee-per-year figure or for comparing against a monthly device arrangement. The failure to avoid is mixing them in one table, which produces a figure that is wrong in two directions at once and is nearly impossible to spot by reading the output.
How many role bands do I need?
Three or four. Standard, high specification, field or customer-facing, and executive if that is genuinely a separate category where you work. Band on what the role needs in equipment terms rather than on seniority, because the two correlate without matching, and hold location as an adjustment on the band rather than multiplying bands by countries into a matrix nobody maintains.
What if I cannot measure setup labour or shipping per person?
Use a banded estimate and label it as one. Setup labour is time multiplied by a cost rate and nobody's ledger holds it, so it will always be an estimate, and that is not a reason to carry it as zero. A figure with four measured lines and two labelled estimates is defensible. A figure with four measured lines and two zeros is the one most companies are planning with.
Is there a tool that produces this number?
No, and that is worth knowing before you evaluate anything. Asset registers hold the device line, procurement holds part of the accessories, logistics holds shipping, an IT budget holds the software seats, and nobody holds setup labour. An equipment lifecycle platform covers more of it than a register does because it handles shipping and retrieval, but the total is assembled in a spreadsheet every time.
How often should the figure be rebuilt?
Annually, before the planning cycle rather than during it, and sooner if your geography changes. The lines that move are shipping and recovery, both of which track where you are hiring rather than what you are buying, so opening a market or losing a logistics partner changes the number more than a new laptop model does.
Does this change if we rent or lease devices instead of buying?
Yes, and usually it simplifies. If a single monthly fee per person covers the device, the shipping and the recovery, then your cost per hire is that fee multiplied by the expected term and most of this calculation is already done for you. That is a genuine advantage of those arrangements and it is rarely stated, though you should still confirm what the fee excludes before relying on it.