Nobody sets out to keep a bad inventory. On the day a hardware register is reconciled, it's usually close to right. Then a technician swaps two docking stations between floors. A laptop goes out for repair and comes back to a different desk. A handheld scanner sits offline for a week and its scans never upload. A cart of monitors rolls out a side door the RFID portal doesn't cover.
None of these events is a failure on its own. Each one just fails to reach the record. Give it six months and the register that was right on audit day now points to the wrong room for a noticeable share of devices, and nobody can say exactly when it went wrong.
That gradual separation between the record and the room is inventory drift, and it's the main reason asset inventory accuracy declines between audits even when nobody makes an obvious mistake.
Inventory drift is the growing gap between what your system of record says and what physically exists: which assets you have, where they are, what state they're in, and who holds them. It builds from many small, unrecorded physical events rather than from one bad import or one careless technician.
The difference matters for how you fix it. A data error is a single wrong entry, and you can correct it. Drift is a rate. You can clean every record today and the inventory will start separating from reality again tomorrow, because the process that caused the gap is still running.
Drift shows up in four dimensions, and each can drift independently:
Drift compounds, because each unrecorded change makes the next one more likely to go wrong.
Take one laptop whose location record is stale. At the next spot check, it isn't where the system expects, so it gets flagged as missing. Someone marks it lost. Two months later it turns up in a conference room drawer, and a well-meaning technician creates a new record for it. Now one device has two records, one "lost" and one "active," with its repair history and warranty data attached to the wrong one. One missed move has produced a location error, a false loss, and a duplicate.
The arithmetic is simple. Drift is roughly the volume of physical events multiplied by the share your processes fail to capture. An organization with 1,000 moves, adds, repairs, and returns a month that captures 95% of them still leaves 50 records stale every month. Without continuous correction, that's around 600 inaccurate records a year from a process most teams would call good.
This is why accuracy peaks right after an audit and slides from there. The slope depends on how much physical change your environment sees and how much of it gets recorded at the moment it happens.
Planned moves come with a ticket, and tickets get closed. The drift comes from everything else: desk swaps, borrowed monitors, loaner laptops handed off person to person, servers shuffled between racks during an incident at 2 a.m. These moves are fast and informal on purpose, which is exactly why nobody stops to update the record.
Every asset tracking method has blind spots. Barcode tracking only works when someone scans the label, and people skip scans when they're rushed. RFID removes the manual step but depends on reader placement. A portal at the loading dock does nothing for the side exit, and a handheld that can't sync keeps its reads to itself.
The dangerous part is that a missed scan doesn't look like missing data. The system still shows a location. It's just the last location anyone confirmed, which may be weeks old. If your inventory management platform doesn't show a "last verified" timestamp next to every location, stale data and fresh data look identical.
Duplicates come from re-entry when a "missing" asset reappears, from a procurement import and a discovery tool each creating their own record, or from one system keying on serial number while another keys on asset tag. The fuller mechanics of duplicate records are well documented. For drift, the point is that duplicates inflate counts and split history, so the record that looks current may be the one with no repair or custody trail.
Hardware changes state constantly: staged, deployed, in repair, sent for RMA, returned to stock, retired, disposed of. Each transition tends to happen in a different system. The service desk logs the repair. The vendor portal tracks the RMA. The disposal partner issues a certificate of destruction as a PDF that lands in someone's inbox.
If those events don't write back to the asset record, the status quietly goes out of date. The most expensive version is the ghost asset: a device that was wiped and recycled but still counts as active, still carries a depreciation schedule, and still shows up in license true-ups.
Most enterprises spread hardware data across procurement, the service desk, the CMDB, HR offboarding, and the finance fixed-asset register. Each system updates on its own schedule and owns a different slice of the truth.
Offboarding shows the problem clearly. HR's checklist marks "laptop returned." The device goes onto a shelf in the IT room. The ITAM record still lists the departed employee as custodian, because nobody connected the HR step to the asset system. Multiply that by every handoff between teams and you have drift built into the org chart.
Inaccurate inventory data costs money in a few predictable places. Teams buy hardware they already own because spares sitting in a closet don't show up as available. Finance depreciates and insures equipment that's gone. Audits take longer because every discrepancy needs a manual investigation, and grant-funded or regulated equipment carries compliance exposure when it can't be located.
Security takes a hit too. The Center for Internet Security puts inventory and control of enterprise assets as the first of its 18 CIS Controls, since a device you don't know about is a device you can't patch, monitor, or wipe. Its guidance calls for reviewing the enterprise asset inventory at least twice a year. Drift decides how far off that inventory is by the time the review comes around.
Then there's the time cost. When technicians don't trust the location data, they walk the floor instead of querying the system, and the inventory stops being a tool anyone relies on.
You can't manage a rate you don't track. These measures show whether drift is growing or shrinking:
Watch the trend more than any single month's number. A match rate that holds steady at 96% is healthier than one that hits 99% after an audit and falls to 90% by the next.
Since drift comes from physical events that don't reach the record, the fix is to capture data at the point where the physical change happens instead of reconstructing it later. That's the core idea behind hardware-enabled asset management: let readers, tags, and scan points record movement and custody as a side effect of the work, so accuracy doesn't depend on someone remembering to update a field.
In practice, that means a few changes:
Put capture points where assets change hands. Fixed RFID readers at doorways and docks, smart cabinets for loaner pools and spares, rack-level sensors in the data center, and barcode check-in and check-out at the IT counter each close a specific gap. AssetVue's hardware asset tracking platform combines passive and active RFID, barcode labels, fixed and handheld readers, smart cabinets, and rack sensors for this reason. Different locations need different capture methods.
Choosing between them comes down to volume and environment. The trade-offs between RFID and barcode tracking usually favor RFID where many assets move through a chokepoint without line of sight, and barcode where handoffs are deliberate and one at a time.
Make the physical tag the primary identifier. When every system references the same tag ID, the serial-versus-asset-tag mismatch that creates duplicates goes away.
Route status changes through one event. A repair ticket, an RMA shipment, or a disposal certificate should update the asset record directly through integration, rather than waiting for someone to re-key it.
Show verification age and work an exception queue. Surface "last seen" timestamps, and route assets that miss an expected scan to a queue someone owns. A missed read should become a task, not a silent gap.
Physical audits still have a job, but a smaller one. With continuous capture in place, an audit confirms the inventory instead of rebuilding it from scratch. When the audit finds little to fix, your capture points are working.
If you want a fast read on where you stand, answer these five questions:
Each "no" marks a place where drift is already accumulating.
Inventory drift starts the day after your last audit. AssetVue captures asset movement and custody with RFID and barcode tracking at the moment it happens, so your records keep up with your hardware. Schedule a call to see where drift is building in your inventory.