Equipping Your First 50 Hires Without an IT Function

TL;DR

• The point where a spreadsheet and a company card stop working is narrower than people expect, and it is driven by countries and movement rather than headcount.
• Three people abroad is a courier problem. Thirty people across six countries is a platform problem, even though the headcount is similar.
• Build the asset register before you buy anything. It is the most durable thing you own here and it survives changing vendors.
• Decide the device standard early. Two specifications covering most roles beats eleven covering everybody perfectly.
• A device management system is a separate purchase from logistics, and most teams discover that when a sealed laptop arrives and nobody can set it up.
• Nobody in this category publishes a price, so comparison means writing a specification and asking rather than searching.

The Spreadsheet That Was Nearly Right

A company at forty-two people had done better than most. Somebody in operations kept a spreadsheet, machines were ordered from one reseller, and the whole thing worked without anybody calling it a system.

Then three things happened in the same quarter. Two people left, one in another country. A new engineer needed a specification nobody had bought before. And the person who maintained the spreadsheet moved into a different role.

Six weeks later nobody could say with confidence how many laptops the company owned. The spreadsheet existed and had last been updated before the departures. One returned machine was in a cupboard and not recorded. The other had never been collected, because the leaver was abroad and the conversation about shipping it back had drifted into nothing.

Recommended Tools

Best tools for Device Management

B BambooHR HR management system L Lattice Performance management D Deel Global payroll & EOR

View all recommended tools

Nothing dramatic went wrong and nothing was anybody's fault. The arrangement had depended on one person's attention, and attention is the thing that moves when somebody changes job. That is the real transition at this size, and it has nothing to do with headcount.

The First Fifty Are Not One Problem

Equipping fifty people sounds like one job done fifty times. It is four jobs with different failure modes, and they arrive in sequence rather than together.

Choosing what to buy. Early on this is somebody picking whatever they would want themselves, which produces eleven different machines across fifty people and a support situation nobody can hold in their head.

Getting it there. Easy in one country, genuinely hard across borders, and the difficulty is not proportional to distance. Two people in the same continent can be a straightforward courier booking and a customs problem respectively.

Making it usable. The gap nobody plans for. A sealed laptop arriving at a new starter who has nothing to log into is a wasted first week, and it happens because procurement and configuration are separate purchases that look like one.

Knowing where it all is. Invisible until somebody leaves or an auditor asks, and by then the answer has to be reconstructed rather than looked up.

A spreadsheet and a company card handle the first two adequately for a while. They never handle the third, and they handle the fourth only if somebody is disciplined about it from the beginning.

When the Manual Version Stops Working

Headcount is the wrong trigger and it is the one everybody uses. Three things predict the breaking point far better.

Countries, by headcount rather than by flag. Forty people in one country with three abroad is a different operation from forty across six markets. The second generates customs questions, local stock questions and replacement questions continuously. The first generates them a few times a year.

Movement. Joiners, leavers and reassignments annually. A stable team of fifty with four leavers is nothing. Fifty people with twenty-five movements is a part-time job somebody is doing badly between other things.

Whether anybody owns it. If device logistics is genuinely nobody's work, no tool fixes that. A managed service might, and a platform will not.

The second row catches people. Three people in other countries feels international and is operationally a handful of bookings a year, which no subscription is worth.

Build the Register First

Whatever you decide about vendors, this is worth doing this week and costs nothing.

One row per device: serial number, model, who holds it, where, since when, what it cost, and under what arrangement. Include the machines somebody bought on a company card eighteen months ago, because those are the ones that go missing.

Record the gaps explicitly. A device you cannot account for is a device that will not come back, and the list of unknowns is itself a finding worth showing somebody.

Export it monthly to a file you control, even after you adopt a platform. A register living only inside a vendor's product is a register you rent, and it is the one asset here that survives changing suppliers.

Buying a platform in order to get a register is backwards, and it is the commonest sequencing error at this size. The register is what you hand a vendor during evaluation, and it is what lets you answer the existing-fleet question with a number rather than an estimate.

Decide the Standard Before the Fiftieth Hire

The second cheapest thing you can do early, and it gets harder every month you delay.

Two specifications cover most companies. A standard machine for the majority and a heavier one for people who genuinely need it, meaning designers, engineers running local builds, anybody working with video. Resist a third unless somebody can name the role it exists for.

Write it as a specification, not a model. Processor class, memory, storage, screen size, operating system. Models go out of stock constantly and a standard anchored on one specific machine expires within months.

Write down who approves an exception and on what grounds. Not to be restrictive. Because without it, exceptions are granted by whoever asks most persistently, which is both unfair and expensive.

Standardising late is painful. Every non-standard machine already issued stays in the fleet for its full life, so a company that standardises at a hundred people carries the consequences of the first hundred decisions for three or four years.

There is a second benefit that only appears later. A standard fleet is a fleet a vendor can hold stock of, which is what makes next-day replacement possible at all. Eleven different machines across fifty people means no vendor can keep a spare that fits anybody, so every failure becomes a fresh purchase and a shipment rather than something sitting in a warehouse an hour away.

The Configuration Gap

This is the failure that surprises people most, and it is structural rather than careless.

A logistics vendor buys, ships and tracks. It will enrol a device into a management system you already run. Almost none will provide that system. So a company with no IT function buys device logistics, a sealed laptop arrives correctly and on time, and there is nobody to perform setup.

The consequences are not only inconvenience. A machine that is never enrolled has no policy-enforced encryption, no remote wipe if it is lost, and no visibility into what is installed. The new starter works on it anyway, because they have a job to do, and the company's work sits outside every control it has.

Check what you already have before assuming this is new spend. Apple's own platform now carries management at no cost for Apple fleets. Most companies running a Microsoft estate already hold a licence covering it through something they pay for anyway. The gap is frequently a configuration job rather than a purchase.

And name the person who performs setup on day one. If the honest answer is the new starter following a written document, that is legitimate at this scale and only works if somebody has actually written the document. Write it from a clean machine rather than from memory, because the steps you take automatically are exactly the ones that get left out and then block somebody on their first morning.

What Day One Should Actually Look Like

Working backwards from the outcome makes the whole thing easier to plan, because most of the failures are timing rather than capability.

The machine arrives before the start date, not on it. Two or three working days of buffer absorbs a customs delay, a failed delivery and a wrong address, all of which happen regularly. Shipping to arrive on the first morning means any single problem becomes a lost week.

It is enrolled before it ships. The person opening the box signs in rather than configures. This is the difference between a good first day and somebody spending it on a support call, and it is the thing to check a vendor can genuinely do rather than assume.

Somebody is reachable that morning. Not a ticket queue. A name, because the failures on day one are small and urgent, and a next-business-day response means the problem spans the whole of the first day.

And the asset is recorded at dispatch rather than later. Serial, holder, address, date. Recording at dispatch rather than at some point afterwards is the single habit that keeps a register accurate, because later never arrives reliably.

Work backwards from that and the lead times become obvious. Order placed when the offer is accepted rather than when the start date approaches, which also means somebody in People has to tell somebody about the hire, which is the process gap underneath most day-one failures.

The Offer-to-Start Window

The practical constraint nobody models is how long you actually have, and it is usually shorter than the lead time.

A typical notice period gives several weeks between offer and start, which sounds comfortable. In practice the order is placed days before the start date, because nobody triggers it earlier, and a machine shipping internationally with customs clearance can take a week or more on its own.

So the trigger matters more than the lead time. If the device order fires on offer acceptance, almost any lead time works. If it fires when somebody remembers the week before, no vendor can save you.

Make it a step in whatever your hiring process already is. Not a new system: a line on the existing checklist, owned by whoever processes the offer, with the device specification and the shipping address captured at that moment while the candidate is responsive and before they go quiet in their notice period.

And capture the address properly. A surprising share of day-one failures are a wrong or incomplete delivery address, because somebody typed it from memory or the new starter moved between accepting and starting. Confirm it the week before dispatch rather than at offer.

The Options

RemoAsset, named below, is owned by the same people who publish HROpsLab. Its limitations are stated as plainly as everybody else's.

RemoAsset

Procurement, deployment, tracking and retrieval for distributed teams, with enrolment into your management system before dispatch.

At this size: pre-dispatch enrolment directly addresses the configuration gap above, which is the failure most likely to hit a company with no IT function. Its limits stand: it publishes no price and requires a demo, it is weaker for hardware it did not supply, which matters because your first fifty machines were bought before any of this, and it is not a device management system.

Workwize

Hardware lifecycle with owned warehouses. Domain is now goworkwize.com; the older address no longer resolves, verified 14 October 2026.

At this size: warehousing earns its cost through storage and redeployment, which a small stable fleet may simply not generate. Strong if people are already churning. No published price.

Firstbase

Equipment on subscription with retrieval and replacement included by design.

At this size: the per-person model maps neatly onto a small headcount and puts more of the lifecycle on the vendor by default, which is worth real money when nobody internally owns it. You hold no asset at the end. No published price.

GroWrk and allwhere

Both scope per customer, GroWrk across a wide country list and allwhere closer to a managed service.

At this size: allwhere's managed character is a genuine answer to having no internal owner. Both take evaluation time and both can accommodate a small arrangement, which is worth asking about directly rather than assuming. Neither publishes a figure.

Deel IT

Devices attached to an employment platform.

At this size: straightforward if every person holding a device is employed through that provider, which is common early and stops being true the moment you hire directly at home. No published price.

The Comparison

Ask every one of them whether there is a minimum commitment in devices or spend. Several arrangements carry one and discovering it in week four wastes three weeks.

The Decision Table

Getting Devices Back, Which Nobody Plans For

Offboarding is the half of this that gets no attention at fifty people, because almost nobody has left yet. It is also where the money goes once they start to.

Nothing triggers automatically. At a company with no IT function, a resignation lives in an HR process that knows nothing about hardware. So the collection starts when somebody remembers, which is usually after the last working day, which is precisely when the person stops answering email.

Packaging decides the return rate more than anything else. A former employee asked to find a box and arrange a courier mostly does not. A prepaid, pre-labelled container arriving at their door gets materially better returns, and the difference is larger than any other single factor including how politely you ask.

Condition needs recording at both ends. A machine arriving with a cracked screen is an argument you cannot win without a dispatch condition note and a photograph. Decide now that you will record it, because retrofitting that to a fleet already issued is impossible.

And some devices are not worth recovering. A four-year-old laptop with somebody in another country frequently costs more to collect than it is worth. Write the rule in advance based on age and location, because the alternative is three months of chasing followed by the same conclusion and a worse mood.

At fifty people with normal attrition that is a handful of devices a year. The habits are cheap to establish now and expensive to introduce at two hundred, when the volume makes them urgent and the existing fleet has no condition records at all.

The Four Costs Nobody Budgets

Hardware is the number everybody plans for. These four are the ones that appear afterwards.

Shipping and duty, which varies enormously by destination and can be a meaningful fraction of the device cost on a single international delivery. The question of whether duty sits inside a vendor's quoted figure or arrives as a bill at your new starter's door is worth settling explicitly.

Replacement and repair during the life of the machine. Something always fails, and at fifty devices the question is not whether but how many per year. A spare machine or two removes most of the pain and is cheaper than expedited shipping every time.

Storage between holders. A machine returned in March and reissued in June sits somewhere for three months. Either you pay somebody, or it occupies a cupboard and the cost is hidden rather than absent.

And the hours. Whatever the vendor does not cover stays with you. Put a rough figure on it, even an approximate one, because that number is what justifies paying more for a vendor who covers more, and it is the only part of the comparison that never appears on any quote.

A fifth, if you ship internationally with any regularity: the cost of a delivery that fails and has to be repeated. Wrong address, nobody home, customs holding the parcel. It happens often enough at fifty people to be worth a line rather than a surprise, and it is the item most likely to be billed as an extra rather than absorbed.

Add those to the hardware cost and divide by headcount to get a real annual figure per person. It will be higher than the device price suggests, and it is the number to take to whoever approves spending rather than the machine cost alone.

What Getting This Wrong Costs

The visible cost is a subscription that does less than expected, and that is the smallest part.

The real cost is hours from people who have other jobs. A company with no IT function pays for device logistics in founder time and People time, and those hours never appear on an invoice. A platform covering three of the four jobs leaves the other one exactly where it was, so a team expecting to get a day a week back gets half of it and concludes the product was oversold. It was not. The specification was.

Then there is the compliance exposure nobody counts. Work sitting on an unconfigured machine because nobody enrolled it is a real risk, and it surfaces only when somebody asks where the files are. That question tends to arrive at an inconvenient moment, usually during a customer security review or a funding diligence process.

And there is the standardisation debt. Decisions made in the first fifty hires stay in the fleet for the full life of each machine, so a company that never set a standard carries eleven device types into its second hundred people.

The Week Somebody Finally Owns It

Every failure described in this article traces back to device logistics being somebody's fourth priority, so the highest-return change is organisational rather than technical.

It does not need to be a full role. At fifty people it is a few hours a month, and the work is mostly noticing rather than doing: that a hire is coming, that a machine has not arrived, that a leaver's laptop is still out, that the register has drifted from reality.

It does need to be one person rather than a function. Shared ownership of an unglamorous recurring task means nobody owns it, and the task reverts to whoever is most conscientious, which is how it ended up with People or with a founder in the first place.

And it needs the authority to spend within a limit. An owner who has to escalate every purchase is a bottleneck rather than an owner, and the delay shows up as somebody waiting for a laptop.

Give them four things to do each month and nothing more. Check the register against reality for anyone who joined or left. Confirm nothing is sitting with a departed employee. Note anything that broke and whether a spare is needed. And look at what is coming, so orders fire on offer acceptance rather than on panic.

That is the whole job at this size, and it is the difference between a fleet you can account for at a hundred and fifty people and one you cannot.

When You're Ready to Buy Something

Two things before any vendor call, and both are free.

The register, with its gaps. Every device you can account for and an honest list of the ones you cannot. This is what you hand a vendor during evaluation, and the import question is the one that most distinguishes them.

A named owner with actual time. Not a function. One person, because every failure described here traces back to device logistics being somebody's fourth priority. Vendors can tell within one call whether there is an owner on your side, and arrangements bought to avoid naming one tend to fail by month six regardless of which vendor was chosen.

Then write one page covering headcount by country including single-person markets, the device standard as a specification, last year's real joiners and leavers, and a list of jobs each vendor marks as included, extra or not offered. Send the identical page to three or four of them.

The whole exercise is a day of work. It is also the difference between choosing a vendor and choosing a number.

One last thing worth doing before you commit to a term. Run a single hire through the whole chain, in whichever country is hardest for you, and deliberately raise a support question partway through. A pilot that only tests the happy path shows you what the demo already showed you. How a vendor handles a real problem while they are still selling is the best available prediction of how they will handle one afterwards, and at this size you are likely to be on the receiving end of that answer personally.

Frequently Asked Questions

HROpsLab takes no vendor money and publishes no paid placements.