How much does a refresh actually cost — and where the budget always leaks

A 500-device refresh shows up in the IT budget as a $750,000 hardware line. That number is wrong. Not because the hardware costs less — laptops do cost about $1,500 a piece — but because the hardware is rarely more than 60% of what the refresh actually consumes. The rest hides in places that are hard to budget for, easy to underestimate, and almost never accounted for honestly when the CFO signs the capital request.

This piece walks through where that other 40% lives. It's written for the people who have to defend a refresh budget to a finance committee, and for the people who have to live with the consequences when the budget turns out to be wrong. Both readers have probably been through this before.

The hardware number is the easy one

Start with what's visible. A modern business-class laptop costs $1,200 to $2,200 depending on configuration, vendor, and the negotiating leverage of the buyer. Multiply by fleet size; that's the hardware line. For a 500-device fleet at an average of $1,500 per device, that's $750,000.

Three things are already wrong with this number, but they're wrong in directions the IT team often doesn't catch.

First, the average is rarely the right number. Persona-driven device strategy — the discipline of matching each user to the right device class rather than defaulting laptops to everyone — typically reveals that 20-40% of users are being over-equipped. A clinical workstation user doesn't need a $2,200 laptop; a small form factor desktop at $700 does the same job better, with a longer service life and no thermal throttling. A call-center agent might be better served by a thin client. Run the persona model honestly and the hardware line shrinks meaningfully. Skip the persona work and the hardware budget is sized for a fleet you don't actually need.

Second, the vendor list price isn't what sophisticated buyers actually pay. Direct-purchase pricing from Dell, HP, or Lenovo varies by deal size, customer relationship, and the specific configurations on the cart. Cross-shopping the same fleet through value-added resellers like SHI or Insight produces yet another set of numbers. Most IT teams quote one vendor for one configuration and stop there; the result is a hardware bill that's typically 8-15% higher than it could be with real procurement discipline.

Third, the warranty add-on is usually invisible. Most refresh quotes default to extended onsite warranty coverage — three to five years of vendor-provided service. The premium is real, often 5-15% of per-device cost depending on vendor, model, and coverage tier, and it's often not necessary. Operators who maintain onsite spares, or who work with logistics providers that maintain spare pools, can run standard mail-in depot warranty and capture the premium as savings. Most refresh budgets line-item the extended warranty without examining whether the response time it buys is actually faster than the operational alternative.

So the $750,000 hardware line is usually a ceiling, not a target. A disciplined refresh project — one that runs the persona model, brokers competitive pricing across vendors, and optimizes the warranty strategy — can land 20-30% lower on hardware alone.

That's the easy 40% of the story. The hard part is what isn't on the quote.

The labor cost is bigger than anyone admits

A refresh isn't just shipping new laptops; it's a structured operational program. Image building. Asset tagging. Migration planning. End-user coordination. Hypercare. Retrieval. Disposal. Each phase consumes hours, and the hours add up.

The honest accounting goes something like this. For a 500-device refresh, allocate roughly 4-8 hours of internal IT time per device when the program is run in-house — covering imaging, deployment, user coordination, support during the swap, and retrieval. That's 2,000 to 4,000 hours of labor. At a fully-loaded internal IT cost of around $75-$125 per hour, that's $150,000 to $500,000 of labor cost on top of the hardware. Most refresh budgets don't model this because the labor is "already paid for" — the IT team is on salary, so the project doesn't show up as an additional expense.

But the labor is paid for somewhere, and it's paid for twice. Once in the project itself, where the IT team is consumed by the refresh and isn't available for normal operations. And once in the long tail of degraded service quality during the project — the ticket queue grows, response times slip, normal infrastructure work gets deferred. Customers notice. Employees notice. The CFO doesn't see the line item, but the COO sees the consequences.

Externalizing the labor — engaging a vendor that runs the refresh as a structured program — moves this cost from "hidden" to "visible." That's uncomfortable for budget purposes but honest. A managed refresh program, run by a vendor that's done it many times, typically delivers the same work in half the hours, because the program already exists. The customer isn't paying for the vendor to learn how to run a refresh.

The productivity cost during the swap

Here's a number that almost never makes it into the budget: every laptop swap costs the user a half-day of productivity minimum, often more, depending on how well the migration is handled.

The half-day is the floor. It assumes everything goes right — files are in OneDrive, applications reinstall cleanly, the network works, the user knows how to log in. In practice, the actual time loss per user is closer to a full day, sometimes two, because something always doesn't go right. The home printer that won't reconnect because the user doesn't have admin rights to install the driver. The VPN configuration that didn't come over. The application that needs to be re-licensed. The file the user thought was on their desktop that turns out to have been local-only.

For a 500-device refresh, even at a conservative half-day per user, that's 250 days of lost productivity — 2,000 hours. At a typical mid-market employee fully-loaded cost of $50-$75 an hour, that's $100,000 to $150,000 of productivity loss that doesn't show up on any line item but absolutely shows up in operational reality.

The way to compress this number isn't to swap faster. It's to engineer the migration so the half-day stays the half-day. OneDrive sync monitoring before the swap. Known Folder Move verification. Pre-imaging that lays down all the standard applications. A hypercare team available to solve the home-printer problem in the moment it appears, rather than leaving the user to fight with it for three days. This is exactly the operational discipline that separates engineered refresh programs from improvised ones, and the productivity savings it generates often exceed the labor savings it generates.

The long-tail costs that show up after the project is "done"

A refresh ends, on the project plan, when the last new device is deployed. On the ledger, it ends six months later, when the trailing costs finally taper off.

What does the long tail look like? Help-desk tickets in months three and four about applications that didn't survive the migration. Lost-file recovery requests in month five from users who realize they had data only on the old laptop. Compliance evidence requests in month six when the auditor wants to know what happened to specific devices and the documentation is incomplete. Unaccounted-for retired devices in month seven when someone notices the asset register doesn't match physical inventory.

Each of these is small individually. Each takes IT time. Each represents a process that wasn't engineered upstream. A typical refresh leaves behind 100-200 of these long-tail items for a 500-device fleet, consuming perhaps 200-400 hours of IT time over the six months following deployment. That's another $15,000 to $50,000 of cost that won't appear in the budget, won't be attributed to the refresh, and will simply be absorbed as "normal IT work."

A well-engineered refresh closes most of these gaps upstream. 30-day secure hold of retired devices before wipe accounts for the "I forgot to copy a file" requests. Self-service documentation accounts for the application-reinstall questions. Comprehensive asset tagging and chain-of-custody documentation accounts for the audit questions. The work happens in the project so it doesn't have to happen in the long tail.

The honest total

For a 500-device refresh, the budget number on the capital request will be the hardware: $750,000.

The real cost, including labor, productivity loss, and long-tail expense, is closer to $1.1M to $1.5M. The difference — the $350K to $750K that the budget didn't capture — is what makes refresh projects feel painful in retrospect even when they "came in on budget."

The point of this isn't to scare anyone away from refresh. Fleets age out. The work has to happen. The point is that engineering the refresh as a structured program, rather than improvising it project-by-project, captures most of the hidden costs as visible savings. The hardware line might not move dramatically. The labor, productivity, and long-tail lines do.

How Surya thinks about this

We run computer refresh as a productized program — seven phases, structured operations, fixed scope per engagement. We don't make money on hardware markup; the customer buys direct from the vendor and we bring the procurement discipline. The savings on hardware, labor, productivity, and long-tail accrue to the customer, and they're typically larger than the cost of running the program with us.

If you're sizing a refresh and trying to model the real cost, we offer a fixed-scope refresh assessment that produces a written fleet picture and cost model your team can act on. Fleet size and complexity drive the scope; we quote within two business days of an initial conversation.

Request a refresh assessment →