IT Procurement Services: A Practical Guide for Teams Deciding Whether to Outsource Buying

TL;DR

• IT procurement services means paying somebody else to do some or all of your technology buying: sourcing, negotiating, ordering, delivering and sometimes the record afterwards.
• You do not need this while one person enjoys doing it and the volume fits around their job. That is a real and common situation.
• Four models exist, from a reseller who just sells you things to a managed service that owns the whole function.
• The decision is rarely about unit price. Vendors with buying power rarely beat a direct manufacturer account by enough to matter on its own.
• What you are actually buying is coverage, time, and somebody being accountable for a date.
• The hidden cost is the record. If the service does not leave you with a clean asset register, you have outsourced the work and kept the problem.

The Thing That Was Supposed to Take Twenty Minutes

A company of about ninety people needs eleven laptops for a team starting in six weeks. Three are in Spain, two in Poland, the rest at home in the United Kingdom.

The office manager, who has bought laptops before, starts on the Monday. By Wednesday she has discovered that the business account she usually uses does not ship to Poland, that the Spanish orders need a local tax identifier she does not have, and that the specification she ordered last time has been replaced by two models with confusingly similar names.

She asks IT, who are mid-migration. She asks Finance, who want to know why this was not in the forecast. On the Friday she finds a reseller who can do the United Kingdom machines next week and will come back about the others.

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

The others take three weeks and arrive in two shipments, one of which sits at customs for nine days because the declared value was wrong on the paperwork somebody else filled in.

None of this is incompetence. The job looked like ordering eleven laptops and was actually five different procurement problems in a trench coat, and the person doing it was never given the tools or the time to find that out in advance.

The real issue is not that buying is hard. It is that buying across borders is a specialist job that most companies assign to whoever has capacity.

When You Do Not Need This

While it genuinely fits in somebody's week. A company of forty, in one country, buying perhaps three machines a month, has no procurement problem. It has an ordering task. Introducing a service here adds a party and a margin and solves nothing.

When it is annoying but bounded. The friction stage looks like occasional spec confusion, the odd late delivery, and a growing sense that nobody is quite sure what the standard machine is. Fixable with two documented configurations and a named owner.

When it becomes a liability rather than an annoyance. The tipping point is almost always geography rather than volume. The first time you hire in a country where your existing route does not work, the problem changes shape entirely and does not change back.

The edge case that overrides everything: a hiring spike with a date attached. Doubling a team in a quarter is not the same problem as steady growth, and it is the clearest case for bringing in help temporarily even if you would not outsource permanently.

What Teams Actually Need

"Can we get machines where our people are?"

The honest version is a list of countries and a yes or no against each, with a lead time. Most companies have never written this down and discover the gaps one hire at a time.

It matters because this is the only question where outsourcing reliably beats doing it yourself. Buying power is marginal. Reach is not.

"Who is accountable for the date?"

Somebody has to own the promise that a machine is there on day one. Inside a company that accountability is usually diffuse. A service contract can make it explicit, which is often what you are really buying.

It matters because a late laptop has a cost that is real and never budgeted, and diffuse accountability is why nobody treats it as urgent until the week it happens.

"What are we actually paying?"

Unit price, service fee, shipping, duty, and whatever is charged for the things that are not ordering. Several models hide most of the cost outside the unit price.

It matters because comparing a managed service against your current approach on machine price alone will mislead you in both directions.

"Who keeps the record?"

If a third party buys the machines, who maintains the list of what you own, where it is and what it is worth. The answer is frequently nobody, which is worse than before.

It matters because the register is the asset that compounds. Everything else is a transaction.

"How do we get out?"

If this arrangement ends, what do you keep. Supplier relationships, pricing, the device history, anything held in a warehouse.

It matters because procurement arrangements are sticky by design and the exit terms are negotiable precisely once.

Measure Before You Shop

The questions above tell you to measure your lead time and then leave you to work out how. Here is the exercise. It takes about two hours and it decides whether you need IT procurement services at all.

Take the last twenty machines you bought. For each one find four dates: when the need was confirmed, when the order was actually placed, when it shipped, and when the person had it working. Then look at four intervals rather than one total.

Nearly every team that runs this finds the first interval is bigger than the other three combined. That gap is entirely internal, costs nothing to fix, and is invisible because no system records it. A supplier cannot help with it, and a managed service will inherit it.

The second finding is usually about the fourth interval. A machine that arrived on Friday and was usable the following Wednesday was late, but every system says Friday because tracking stops at delivery.

So the honest sequence is: measure, fix the internal gap, measure again, and only then decide whether what remains is a supplier problem. Teams who buy first frequently discover their new provider is fast and their process is still slow.

Audit Your Coverage, Market by Market

The second measurement matters more than the first if you are distributed, and almost nobody does it before going to market.

Write down every country you will ship a machine to in the next twelve months. Against each, record what route exists today and what it costs in days. Most companies find the list is shorter than they feared and that two or three entries account for all the pain.

That last row is the one that justifies buying something. The others are mostly tractable internally with a decent supplier relationship and some patience.

Take this table into every vendor conversation and ask them to fill it in for your markets specifically. A provider who answers with a country count rather than a lead time per market has not answered the question, and the difference between the third and fourth rows is where most of the disappointment in this category lives.

Where the Margin Actually Sits

Comparing proposals is difficult because the four models put their money in different places, and a straight unit-price comparison will mislead you in both directions.

A reseller makes its margin on the machine. The unit price looks slightly high and almost everything else is either free or not offered. Easy to compare, easy to budget, limited in what it covers.

A lifecycle platform typically makes less on the hardware and charges for the service around it: storage, shipping, retrieval, the console. The machine price can look excellent while the total is higher, and the costs that are not in the unit price are the ones that recur.

A managed service tends to bundle, which is comfortable and makes it hard to tell what any individual component costs. That is fine until you want to unbundle, at which point you discover you have no idea what the pieces were worth.

The practical response is to ask every provider for the same breakdown: machine, delivery to the person, storage per month, retrieval per device, and anything charged per transaction. Then model it against your real numbers, which means your actual hiring rate and your actual country mix rather than a round figure. Two proposals that look 5 per cent apart on unit price can be 30 per cent apart once retrieval is included, and retrieval is the part you will use most as the team matures.

The Four Models

Transactional reselling

You order, they fulfil. A business account with a reseller, some negotiated pricing, maybe imaging and asset tagging.

When it is right: one country, predictable volume, somebody internal who owns the process and likes it. Cheapest option by a clear margin.

When it fails: the moment you need something outside their footprint. Resellers are built around offices and pallets. Eleven individual machines to eleven home addresses in three countries strains the model, and the price reflects that strain if they will do it at all.

Procurement consultancy

Advice rather than execution. Category strategy, supplier selection, negotiation support, sometimes a tender run on your behalf.

When it is right: a large one-off commitment where a few per cent genuinely matters, or a company with real volume and no commercial expertise.

When it fails: when the problem is operational rather than commercial. Excellent advice about supplier strategy does not get a laptop to Lisbon on Thursday.

Equipment lifecycle platforms

Vendors who buy, hold, ship, retrieve and wipe, with a console over the top.

When it is right: distributed teams, especially where a meaningful share of hires are somewhere you have no office. This is the model that actually solves the coverage problem.

When it fails: two ways. Coverage varies more than country lists suggest, and a market served by a partner behaves nothing like one with local stock. And these are logistics businesses: they will not discover devices nobody told them about, and they do not manage software licences.

Fully managed IT procurement services

A provider owns the function: standards, sourcing, ordering, delivery, often support and refresh planning too.

When it is right: companies with no IT function and no intention of building one, or those where procurement has become a genuine bottleneck on hiring.

When it fails: when you still want control over specification, or when the relationship outlasts its usefulness. These arrangements are the hardest to unwind and the most likely to leave you without an internal view of your own estate.

Standardise Before You Outsource, Not After

The decision table recommends two standard configurations more than once, and it is worth explaining why that has to come first rather than being something a provider sorts out for you.

A provider quoting against an undefined specification quotes for variety, and variety is expensive in ways that do not show up as a line item. Different models mean different stock positions, different warranty terms, different resale values and different support paths. A company ordering nine variants across sixty machines is paying for all of that without ever seeing it itemised.

A configuration is more than a laptop model. It is the memory and storage, the operating system version, the peripherals that ship with it, the warranty length and the enrolment route. Two companies can both claim they standardise on one model and have completely different support burdens because one pinned the rest and the other did not.

The exception path decides whether the standard survives. A standard with no route for exceptions is a standard people work around, and the working around is invisible until an audit. A standard with a documented exception path, where somebody can ask, give a reason and get an answer within a few days, holds up.

Exceptions divide into four kinds and they deserve different handling. Accessibility or health requirements, which should never wait behind anything else. Genuine role requirements where the work does not fit the standard machine. Location constraints, where the standard is not sold or supported locally. And preference, which it is entirely reasonable to decline. The most damaging design error in this whole area is routing the first category through the same queue as the fourth, and it is also the most common.

Do this work before you go to market and the proposals you get back will be comparable. Do it afterwards and you will spend the first year of the arrangement renegotiating.

There is a second benefit that shows up later. Standard configurations make a fleet reissuable, and a reissuable fleet is worth considerably more than an equivalent pile of mismatched machines. A returning laptop that matches the standard goes straight back out to the next starter. One that does not sits in storage while somebody decides what to do with it, which is how cupboards fill up. So the standardisation you do to make IT procurement services comparable also quietly improves what happens at the other end of the device life.

Running a Selection That Produces Comparable Answers

Most companies in this position run an informal process: three conversations, a spreadsheet, a decision. That is reasonable at small scale and it has one reliable failure, which is that the three vendors answer different questions and the comparison becomes a judgement about who presented best.

The fix is to send the same brief to everybody and insist the answers come back in the same shape.

A brief that works is short. Your configurations, exactly as standardised. Your country list with required lead times, using the market table above. Your hiring rate, honestly, including the seasonality if there is any. What you want included and what you are happy to pay for separately. And the date you intend to decide, which does more to concentrate attention than anything else in the document.

Then ask for four things in writing rather than in a presentation. A lead time per named market, in days. A breakdown of costs by component. The retrieval success rate they actually achieve, not the process they follow. And the exit terms.

Ask all four in the same document, and give a deadline for written answers rather than offering a call. A call lets a vendor answer the easy three and talk around the fourth.

That last request is the most informative and the most frequently dodged. A provider who will commit in writing to returning your device history and your stored machines without a fee has told you something real. A provider who wants to discuss it later has also told you something.

One practical note on timing. These selections take longer than teams expect, because every vendor here quotes rather than publishes and a quote needs your data before it can exist. Four to six weeks from first conversation to signed agreement is normal, and compressing it tends to mean signing a standard agreement rather than a negotiated one.

What Belongs in the Agreement

Five clauses carry most of the risk, and all five are negotiable at the start and close to immovable afterwards.

Your data on exit. You will accumulate a serial-level history inside their system: what was bought, who held it, when it shipped, what was wiped. Establish that you can export it, in a usable format, at any time, without a fee and without a notice period. A register you can only read while you are paying is a register you are renting.

Your machines on exit. If they hold your spare pool or your returns, those are your assets in somebody else's building. The agreement should say how they come back, who pays and how long it takes. This conversation goes badly when you are already leaving.

What the service level actually commits to. Many equipment service levels in this market describe an aspiration and provide no remedy when missed. That can be acceptable. It is only acceptable if you knew before signing.

Which markets are operated and which are partnered. Put your critical markets in the agreement with lead times attached, rather than relying on a country list that changes when the website does.

What a price rise looks like. An annual cap, or a notice period, or nothing at all. A multi-year commitment with no uplift clause is a bet on the provider's restraint, and it is a bet you do not need to make.

None of these requires legal expertise to raise, and raising them in the first commercial conversation rather than the last tells you a great deal about who you are dealing with. Terms in this category vary far more between providers than headline rates do, and two proposals within a few per cent on price can be completely different on what happens when you want to leave.

How to Choose: Five Questions

What is your real lead time today, measured?

Take the last twenty machines and find the four dates: offer accepted, order placed, shipped, working. Most teams find the biggest gap is internal and free to fix, which changes whether you need a service at all.

How many countries will you actually ship to this year?

Not where you have people. Where a machine will go. That shorter list is the brief.

What is the real cost of a late machine to you?

A number, even a rough one: lost productive days plus the time of everybody who handles the escalation. Without it every proposal looks expensive.

Who will own the register afterwards?

Name a person. If the honest answer is the vendor, make sure you can export it.

Would you sign three years today?

If not, do not. The most common regret in this category is a long agreement signed on a good demo with an exit clause nobody read.

The Options

Manufacturer direct accounts. Apple, Dell, Lenovo business programmes. Best margin, shortest warranty path, and for Apple fleets the route that enables automated enrolment. Where they struggle: no stock holding, no retrieval, uneven reach.

Traditional resellers and VARs. Consolidated buying, one invoice, imaging and tagging. Best for office-based companies with mixed manufacturers. Where they struggle: individual home delivery at scale and anywhere outside their footprint.

Workwize. Lifecycle platform with strong warehousing, which is the thing that removes the customs step. Publishes no rate. Best for distributed teams needing in-region stock.

allwhere. Lifecycle platform covering procurement, deployment and retrieval. Publishes no rate, quote and demo only. Best for distributed teams wanting one vendor across the device life.

GroWrk. Lifecycle platform with a wide country list. Publishes no rate of its own. Worth knowing that its pricing page carries estimated costs for competitors, so figures attributed to this category elsewhere should be traced before use.

Firstbase. Provisioning and retrieval for remote teams. Publishes nothing. Use the firstbase.com domain; the older address returns a 200 and redirects to a homepage, which a link check alone would pass.

RemoAsset. Buying, renting and leasing with shipping, retrieval, storage and wiping, and a register of what it handled. Disclosure: RemoAsset is our own product. We build it, which is why its limits appear here with the same prominence as its strengths. Best for distributed teams whose difficulty is physical rather than commercial. Its Subscription tier is the only tier that publishes a figure, from $1,499 per year, and everything else is quoted like the rest of this list. Where it struggles: no network discovery, so it cannot find machines nobody enrolled through it, and no software licence management, so your application spend stays outside it. If you were hoping one vendor would cover buying and software estate together, that vendor does not exist in this category.

Keeping the Record When Somebody Else Does the Buying

The cost section below argues that the record is the thing most likely to be lost, so this is the practical version of avoiding that.

When a third party buys your machines, three records exist and they drift apart. Theirs, which is accurate about what they handled. Yours, which may or may not be updated. And reality, which is whatever is actually in people's hands. The arrangement works until somebody asks a question that only reality can answer.

Four habits keep this honest and none of them is expensive.

Export on a schedule, not on demand. Pull the provider's data monthly into somewhere you control. Monthly is arbitrary and the point is that it is automatic, because an export you have to remember is an export that stops happening around month four.

Reconcile against something independent. Comparing the provider's list against your own purchase records proves very little, because both derive from the same orders. The check that works is against the current employee list, which finds machines assigned to people who have left, and against a physical sample, which finds everything else.

Record what the provider cannot see. Machines bought before the arrangement, equipment bought locally in an emergency, anything a team expensed. These are invisible to a provider by definition and they are usually the devices that cause problems later.

Assign ownership internally by name. Not to a function. One person who would notice if the monthly export stopped arriving. This sounds trivial and it is the control that actually works, because everything else degrades quietly without somebody who would spot it.

The underlying point is that outsourcing the buying does not outsource the accountability. An auditor, an insurer or an acquirer asks you, not your supplier, and reconstructing a register two years late means asking every employee what they have, which is a project rather than a query.

The Decision Table

The Hiring Spike, Which Is a Different Problem

The edge case noted earlier deserves developing, because a spike is the situation where buying help makes sense even for companies that would never outsource permanently.

Steady hiring is a flow problem. You need a repeatable route and a lead time you can promise against. A spike is a capacity problem with a fixed date, and the two need different answers.

What breaks during a spike is rarely the supplier. It is everything around them. The person who normally places orders now has forty to place alongside their actual job. The approval route designed for occasional purchases now has forty items queued behind one person. Stock that was always available at threes and fours is not available at forty of the same model. And the spare pool, if one exists, empties in the first week and nobody notices until it is needed.

Three responses work, and they are worth considering in this order. Pre-order against forecast rather than against confirmed starts, accepting that you will over-order slightly and that this is cheaper than being late. Raise the approval threshold temporarily so forty orders do not queue behind one approver. And bring in temporary help for the shipping and tracking, which is the labour-intensive part, rather than committing to a long arrangement under time pressure.

A spike is also the moment when the case for IT procurement services is easiest to make internally and hardest to make well. Everybody can see the problem, budget tends to be available, and the pressure to decide quickly is high. Those are good conditions for getting approval and poor conditions for negotiating terms, which is a combination worth recognising while it is happening rather than afterwards.

The thing to avoid is signing a three-year agreement to solve a three-month problem. Vendors are understandably keen to convert urgency into commitment, and the terms you get while panicking are rarely the terms you would accept otherwise. If a spike is the trigger for this conversation, say so openly and ask for a short initial term.

The First Ninety Days With a Provider

If you do appoint somebody, the first quarter decides whether this works, and most of the value is won or lost before anybody ships anything.

Weeks one to three. Give them the data properly: configurations, country list, hiring forecast, and your existing estate if you want it tracked. Agree what the monthly export looks like and confirm you can actually read it. Set the reconciliation habit now, while there is attention on the relationship.

Weeks four to eight. Run the hard markets first, deliberately. The instinct is to start with easy domestic deliveries because they will go well, which proves nothing. Send the first shipments to the two or three markets your coverage audit flagged as difficult, while the relationship is new and the provider is motivated to solve problems.

Weeks nine to twelve. Test retrieval. This is the step everybody defers and it is the one most likely to disappoint, because it is genuinely harder than delivery. If you have a leaver in this window, use it. If you do not, ask the provider to walk you through a recent real example including one that did not succeed.

Then measure the same four intervals you measured at the start. If the internal gap is still the largest, the provider is not your constraint and you should stop blaming them. If the delivery intervals have improved and retrieval has not, you know exactly what to renegotiate at the first review.

The Cost of Getting This Wrong

There are three layers, and the first is the one that gets measured.

The visible cost is unit price and service fees, and it is the smallest of the three. Realistically, the difference in machine cost between a competent reseller and a lifecycle platform is a few per cent, and it is dwarfed by everything else on this page.

It is also the only layer that appears in a proposal, which is why proposals are a poor basis for this decision on their own.

The second layer is time, spread thinly enough that nobody records it. An office manager spending two days on a shipment that should have taken twenty minutes. A hiring manager chasing. A new starter idle. None of this appears anywhere, and all of it recurs.

The third layer is the one that outlasts the contract: the record. If you outsource buying and nobody maintains an internal view of what you own, you have not solved the asset problem, you have moved it somewhere you cannot see. Two years later you need that list for an audit, an insurance renewal or a disposal decision, and reconstructing it means asking every employee what they have.

So the question is not whether IT procurement services are cheaper than doing it yourself. It is this: after a year of this arrangement, would you still be able to produce a complete, accurate list of every machine you own and who holds it, without asking the vendor for it?

What Crossing a Border Actually Involves

Several sections mention customs as though it were weather. It is not. It is a set of specific obligations, and knowing them changes what you ask IT procurement services to take on.

A machine crossing a border is a commercial import rather than a parcel. It carries a commodity code, a declared value and a duty treatment, and somebody is the importer of record. Ship to an employee's home and nominate them as recipient and you may have made a private individual the importer of a company asset, which creates a liability for them and sometimes a question for you.

Duty and import tax are charged on declared value. Understating it to reduce the charge is fraud, so the landed cost of a machine in some markets runs meaningfully above its purchase price, and that difference belongs in a budget rather than arriving as a surprise invoice months later.

Lithium batteries carry handling restrictions, which is why a laptop sometimes cannot travel by the fastest available route. That is not a provider being slow and no amount of escalation changes it.

And automated enrolment usually depends on where the machine was bought. For Apple fleets the device generally has to be purchased through Apple or an enrolled reseller to appear automatically in device management. Buying locally in-market because it is quicker can therefore cost you the zero-touch setup, so the person spends an hour in a setup assistant instead of ten minutes. That trade is sometimes worth making and it should be a decision rather than a discovery.

The reason this matters for a buying decision is that it tells you which questions separate providers. Holding stock inside a market removes all of the above, because the shipment never crosses a border. That is the single largest operational difference between vendors in this category and it is invisible on a feature comparison.

When You Are Ready to Move Beyond Ordering It Yourself

Doing your own buying is not a sign of immaturity, and plenty of well-run companies at two hundred people still order directly because somebody owns it properly and the geography is simple.

What changes the answer is reach and accountability rather than volume. The first time you hire somewhere your existing route cannot serve, you have a coverage problem that no amount of internal process fixes. And when the person doing it can no longer promise a date with confidence, the accountability has already broken even if nothing has gone visibly wrong yet.

There is a second signal worth watching for, and it is softer. When the people doing this work start describing it as firefighting rather than as a task, the process has already exceeded what it was designed for. That happens well before any metric moves, because capable people absorb a surprising amount of disorder before it becomes visible to anybody above them. By the time a late laptop reaches a leadership conversation, the person handling it has usually been improvising for months.

At that point the useful purchase is narrow: coverage in the specific markets you cannot serve, and somebody contractually accountable for a delivery date. You do not have to outsource the whole function to get both, and the companies who get the most from this category are usually the ones who bought the smallest version of it that solved their actual problem.

Frequently Asked Questions