How "Lost" Becomes "Found" Becomes Reordered

Introduction

A write-off looks like a single event. A line item, a date, a dollar amount, a signature. In reality it is the visible endpoint of two timelines that have been running in parallel and mostly out of sync with each other: what is physically happening to the product, and what the system believes is happening to it. The write-off is where those two timelines finally get forced to agree, months after they first split apart.

Where "Lost" Actually Starts: A Transfer Is Not One Update

A field transfer looks simple from the outside. Product moves from one rep to another, or one location to another. Underneath, that single event requires several things to update at once: the physical location, the custodian of record, the ownership and billing assignment if the inventory is on consignment, the linkage to any order the item was tied to, the audit trail entry documenting who moved what and why, and the replenishment routing so future restocks go to the right place.

When a transfer happens informally, a phone call, a parking lot handoff, a favor between two reps covering each other's cases, none of those updates happen automatically. The most common outcome is partial: the location gets updated eventually, because that is the piece someone notices when they go looking for the product. Custody, ownership, order linkage, and replenishment routing lag behind or never update at all.

That partial correctness is worse than an obvious error. A record with the right location and the wrong custody chain does not look broken. Nothing flags. Nobody investigates. The error only surfaces later, when a commission calculation runs against the wrong rep, or a restock routes to a location the product left months ago, or an auditor finds product at a site the system never associated with it.

Where "Lost" Also Starts: The Checkbox That Isn't Verification

The second common origin point is the loaner return. A kit ships to a case marked complete. It comes back marked complete. Real verification would mean comparing what physically came back against the bill of materials, catching anything missing, catching anything that showed up but belongs to a different kit, and assigning responsibility before the kit goes out again. A checkbox does not do any of that. It records that someone processed the return, not that anyone confirmed the contents.

This is a harder engineering problem than it looks, which is exactly why so many platforms skip it. Real bill-of-materials validation requires bidirectional tracking between kits and items, version control on the bill of materials itself so legitimate configuration changes do not generate false variances, and exception handling for items that turn up but belong somewhere else entirely.

Most platforms get one piece of this right and skip the rest. Some compare against the bill of materials but have no path for excess items, so a component that comes back belonging to a different tray has nowhere to go. Some flag missing items but never assign responsibility, which turns the variance into someone's spreadsheet problem rather than a resolvable exception. Some validate against a configuration that was set at product launch and never updated, so legitimate tray modifications generate phantom variances on every audit afterward.

Shipping a checkbox is faster to build than solving all of that in one workflow that does not slow the warehouse team down. The cost of skipping the harder version does not disappear. It moves downstream to whoever eventually has to explain the variance.

Where This Connects to Regulatory Scale

Stock rotation and condition assessment failures are the same category of problem, and they are not hypothetical. In July 2023, the FDA cited Integra's Boston facility for failing to facilitate proper stock rotation and for not assessing product condition, tied to sterility findings connected to a five-year recall. The result was a $22 million inventory write-off and an estimated $110 million in lost revenue over the following two years. That is what happens when unverified movement and un-rotated stock play out at regulatory scale instead of a single territory: the same undocumented-inventory problem, with consequences several orders of magnitude larger.

The Reorder Trigger

Once the location record is wrong, or the return was never really verified, the system behaves exactly as designed. A par level check, a low-shelf-stock alert, or a rep request sees a gap and supply chain closes it. Nobody at this point is checking whether the missing item might actually be sitting somewhere else under a bad record. The system does not have a mechanism to ask that question. It has a mechanism to fill gaps, and it fills the one it can see.

Sometimes the timing makes it worse. A restock ships to where a kit was located. By the time it arrives, the kit has already moved to cover a different case at a different location. Now there are two things in circulation that the system cannot reconcile against each other, and both look like normal, unremarkable inventory positions.

This is the moment the duplicate spend actually happens, and it happens well before anyone knows there is a problem. The reorder is not a mistake in the sense of anyone doing something wrong. It is the correct response to incorrect information.

Where the Variance Goes When Nothing Can Resolve It

Between the divergence and the eventual audit, there is a stage most descriptions of this cycle skip: what the organization does with variances it cannot explain in the moment.

The common outcome is absorption rather than resolution. A catch-all location gets created in the system, initially as a temporary holding bucket for inventory that does not match any known location. Excess items with no clear home go into a physical drawer, and the drawer eventually becomes a side spreadsheet of items with no assigned location. Extra buffer stock gets pushed to the field to cover variance nobody can account for.

None of these are unreasonable responses. Each one is a rational attempt to keep operations moving when the system offers no structured path to resolve the exception. But each one also converts a visible, resolvable discrepancy into a quiet, permanent one. The variance stops generating friction, which means it stops generating attention, which means it survives until an audit forces the issue.

Why the Reversal Doesn't Undo the Cost

Eventually, often at the next full audit, the original item turns up. Finance reverses the write-off. On paper, this looks like the story resolving itself: the item was never really gone, the books get corrected, no harm done.

That reading misses what already happened in between. The replacement unit was already ordered, shipped, and deployed. Finding the original does not un-ship the replacement. It also does not undo the ripple effect on legitimate replenishment: because the territory now shows excess in the reporting period following the phantom order, the next real replenishment request from that same territory is more likely to get questioned, not approved.

Finding the original closes the accounting loop. It does not close the operational one, and it does nothing at all about the workflow that produced the gap, which is why the same sequence is available to run again next cycle.

Where the Fix Has to Happen

Every point in this sequence after the movement happens is already too late to prevent the cost. Audits catch it, accounting reverses it, but neither one prevents the reorder, because by the time either function is involved, the replacement has usually already shipped.

The only point early enough to matter is the movement itself: verifying at put-away, return, and transfer that every field the system depends on, not just location, matches physical reality before that record gets relied on for a replenishment decision. That is a workflow question, not an audit-frequency question, and it is the subject of the framework piece in this series.