The IT Ops Brief

The four months nobody plans for | The IT Ops Brief

Written by Carlos N. Escutia | Sep 16, 2026, 3:02:59 PM

I went into our conversation data expecting to find procurement at the top. Or delivery timelines, the things we talk about most. What I found instead was that the single most-raised subject across nearly four thousand customer conversations was the gap between one employee and the next. Not the beginning of a device’s life, and not the end. The middle.

The most expensive stretch of the device lifecycle is the months between one employee and the next — and it’s a matching problem, not a warehousing one. Across 3,977 customer conversations, the three most buyer-raised topics were where devices sit between users (22.8%), offboarding and retrieval (22.6%), and getting devices enrolled before they ship (22.0%) — all of them in the middle of the lifecycle, the part this category competes on least. Country coverage, the thing every vendor leads with, came up 3.1% of the time. Fixing it starts with IT inventory accuracy you can act on: invisible inventory is functionally identical to no inventory.

The signal: what buyers raise before we do

We analysed 3,977 customer conversations from April 2024 to August 2026, counting only the topics a customer raised before anyone on our side brought them up. That distinction matters: raw mention counts mostly measure your own demo script.

The top three came out effectively tied.

  • Where devices sit between users — 22.8%
  • Offboarding and retrieval — 22.6%
  • Getting devices enrolled before they ship — 22.0%
  • Speed and lead time — 17.9%; certified wipe and compliance — 12.1%; HRIS integration — 11.9%

Two findings stand out, and the second one surprised me more than the first. All three of the top themes live in the middle of the lifecycle. Not buying. Not delivering. The parts that happen between one person and the next — which is precisely the stretch this category competes on least. Vendors sell procurement and shipping because those are legible, demonstrable, easy to put on a slide.

And country coverage — the thing every vendor in this market leads with, including us — came up in 3.1% of conversations. It’s how a company gets onto a shortlist. It stops being interesting almost immediately afterward.

The problem: five ways the middle fails

Here’s the shape of it. Someone leaves in March. Someone joins in May. Those are two separate events, months apart, and in between, a physical object has to exist somewhere.

Most companies have solved both ends. The four months in the middle belong to nobody — and that seam is where the money quietly leaves.

I want to describe five ways I’ve watched this fail, because they’re remarkably consistent across companies of every size.

The device that exists in two places

The record says the machine is in a particular warehouse. The warehouse says it isn’t. It was marked ready to deploy over a year ago. Somewhere between those two facts is a real laptop that somebody paid for — and critically, nothing fired when the record and the object drifted apart. No alert, no exception, no reconciliation. The divergence is silent by design.

Invisible inventory

Devices sit in custody, aging, never assigned — not lost, exactly, just never surfaced as available to whoever was doing the buying. So the company keeps purchasing new machines while owning perfectly good idle ones. This is the most expensive pattern of the five, and the least visible, because nothing appears to be wrong. Nobody files a ticket about a laptop they don’t know exists.

The laptop with no charger

A device ships. The power adapter is ordered separately, because almost nobody stocks accessories against inventory. Two arrivals, one employee, one first day that doesn’t work. The uncomfortable lesson: “a device” isn’t a device — it’s a device, a charger, and the accessories that make it usable. An inventory system tracking only the laptop is tracking a fragment.

“Kept”

Sometimes the employee just keeps it. Negotiated, or simply not worth pursuing across a border. Most asset systems have no state for this outcome, so the record either stays open forever or gets quietly deleted. Neither is true. A lifecycle model without a state for devices that never come back produces a count that is wrong by construction — which is why device retrieval and the record have to be the same workflow.

Available isn’t the same as actually there

A device gets quoted from stock and committed to an employee. At fulfilment it isn’t physically in the building. The availability was a record, not a shelf — and nobody had checked one against the other recently enough.

Put those together and the reframe I keep arriving at is this: storage isn’t a warehousing problem, it’s a matching problem. A device sitting somewhere is only worth something if you can match it to the next need — right specification, right country, right condition, complete with its accessories, at the moment it’s needed. Miss any single one of those and you buy new anyway, and the machine you already own carries on depreciating in the dark.

Invisible inventory is functionally identical to no inventory. Existence isn’t the test. Findability is.

The operator takeaway: five moves, in order of return

  1. Assess on return, not on request. Condition, completeness, repairability — recorded when the device comes back, not when someone finally asks for it. An unassessed device is unmatched by definition, because nobody will pull from a box whose contents are unknown.
  2. Verify stock against the shelf, not the record. Any availability figure that hasn’t been physically confirmed recently is a forecast. Reconcile on a cycle, and treat a divergence as an exception that needs an owner rather than a data-entry correction.
  3. Store regionally, near demand. Not where the device came from. A machine that can’t economically reach the next person isn’t inventory; it’s a souvenir. This is the constraint most spreadsheets ignore entirely, because a spreadsheet has no concept of geography.
  4. Model the states you actually have. Including “kept,” “unrecoverable,” and “awaiting repair.” Most systems offer deployed, in stock, or retired — and reality has more states than that. Every missing state becomes a lie in your count, and every one of them should be visible alongside your device management posture.
  5. Track redeployment rate. The percentage of new device needs filled from existing inventory rather than a new purchase. This is the number that makes the whole problem visible to finance, and almost nobody calculates it.

If you want a fast diagnostic before doing any of that: pick five devices from your asset record that show as available and unassigned, and physically confirm where each one is and what condition it’s in. How long that takes, and how many you can actually confirm, tells you most of what you need to know.

One GroWrk lens

We built GroWrk to sit in the middle of the lifecycle, which in retrospect was the hard place to start: recovery, inspection, regional storage, and matched redeployment, running as one system with one record generated by the work itself.

I’ll be straight about why I find this problem interesting rather than claiming it’s finished. Getting inventory accuracy right at scale — across many countries, thousands of devices, and every one of those messy states — is genuinely the hardest engineering and operational problem in this category. It’s where most of our attention goes, and I’d be sceptical of anyone who described it as a solved thing.

What I’m confident about is the shape of the answer: the record has to be produced by the physical work rather than maintained alongside it. Every time a human being is asked to keep a separate system honest about objects moving through the real world, the record and the world drift apart. Not through negligence — just entropy.

One stat

22.8%. The share of 3,977 customer conversations in which the buyer raised storage before we did — the most-discussed topic in the corpus, and the part of the lifecycle this category competes on least.

(GroWrk Call Intelligence — analysis of 3,977 customer calls, April 2024 – August 2026. Buyer-led mentions only.)

The middle of the lifecycle is unglamorous, invisible when it works, and expensive when it doesn’t. It’s also, based on what our customers actually talk about, the thing they most want solved.

If you want to see what the middle looks like when one system owns it, we would like to show you.

Book a demo →