Ghost Inventory: What It Actually Costs in Duplicate Spend

Introduction

Every field inventory system eventually produces the same strange result. An item gets marked missing. Finance writes it off. Months later it turns up again, at a different location, under a different custodian, in a system that had already closed the book on it.

In an orthopedic operation it is a component from a returned tray, missing after a case, rediscovered in another rep’s trunk during an unrelated count. In a capital or surgical operation it is a piece of equipment written off as missing after a territory transition, found later sitting at the customer site where it had been all along, invisible because the record and the reality split during the handoff.

Different products. Same ghost.

Most companies treat this as a curiosity. An audit footnote. A number that got corrected. It is worth treating as something else: a recurring, structural cost that shows up long before the write-off does, in places most reporting never looks.

The Problem With Measuring This in Dollars

It is tempting to ask what a reappearing item costs. The honest answer is that a single dollar figure is close to useless, because the cost of ghost inventory rarely shows up as one clean number on one line. It shows up as a pattern spread across several different signals, each of which looks unrelated until you put them next to each other.

The more useful way to see this is through leading and lagging indicators rather than a lookup table of dollar amounts. A few worth watching:

• Shelf stock that reads low in the warehouse while field-facing systems show ample supply at the territory level. The gap is not always a shortage. It is often inventory the system believes exists in the field, whether a tray in a vehicle or a unit placed at an account, that cannot physically be located or returned.
  
  

• Case coverage issues that do not track with any actual product shortage. If reps report scrambling for inventory the system says they have, that is not a communication problem. That is a location problem.
  
  

• Supply chain spend that has drifted out of alignment with demand planning signals. When purchasing patterns stop tracking demand logic, one common cause is that replenishment is responding to phantom gaps rather than real ones.
  
  

• A return marked complete that nobody actually verified. In a kit-based operation that means no one compared what came back against the bill of materials. In an equipment operation it means product came back from a placement or a transition and was logged as returned without anyone confirming it against the configuration it was supposed to hold. This is the earliest and most frequent point where the gap opens, because it happens at the highest-volume touch-point in the whole system: every kit or placement that comes back.
  
  

• A catch-all field location that keeps growing. It usually starts as something temporary, named along the lines of Misc Field, Unassigned, or Pending Review, created because the system could not figure out where certain inventory actually belonged. Two years later nobody owns it, nobody audits it, and it grows every quarter. A holding bucket like this is the clearest single indicator that variances are being absorbed rather than resolved.

None of these signals is proof by itself. Together, they describe the shape of a ghost inventory problem before it ever reaches a write-off line.

Where the Signal Actually Starts: The Word "Complete"

A kit ships to a case marked complete. It comes back marked complete. Two cases later, at a different hospital, a rep opens the same tray and something is missing. The same failure has a non-ortho form: a console or instrument set comes back from an account during a transition marked complete, and weeks later the receiving rep finds the configuration is not what the record promised. The investigation almost always traces back to a return marked complete without anyone comparing what physically came back against what the record says should have.

This matters because “complete” is frequently an assertion, not a verification. A checkbox tells you someone processed the return. It does not tell you anyone counted the tray against its bill of materials, or checked the returned equipment against its expected configuration. Real verification requires comparing actual contents to the expected contents at the moment of return: catching items that are missing, catching items that showed up but belong to a different kit or set entirely, and assigning responsibility before the inventory moves again. A checkbox does none of that. It just moves the kit along.

When that verification does not happen, the gap between what the system believes and what physically exists opens at the exact moment it should be closing, the moment inventory comes back under direct control.

How the Duplicate Spend Actually Happens

The mechanism from there is simple, and it does not require anyone to make a mistake beyond the unverified return itself. An item is missing from where the system expects it. Maybe it was used during the case and never logged. Maybe it ended up in a different tray during a busy turnover. Maybe a rep transferred it to cover a colleague’s case and only the location updated, not the rest of the record. Maybe it was a placed unit that moved during a territory handoff and the record never caught up. Whatever the cause, the system’s picture of the world and the physical world have now diverged, and nothing in the workflow forces them back into agreement.

Supply chain does not know any of this happened. What it sees is a gap: a location that should have inventory and does not, according to the system of record. So it does what it is designed to do. It ships replacement inventory to close the gap. Sometimes the timing compounds the problem further. A restock ships to where a kit was, and by the time it arrives the kit has already moved to cover a different case at a different location. Now the restock and the original are in two different places, and the system has no clean way to reconcile either.

The original item was never actually lost. It was somewhere else, doing exactly what field inventory does. But by the time anyone finds it, a second unit has already been ordered, shipped, and deployed to cover a shortage that did not physically exist. That is the duplicate spend. It does not appear on any invoice as a reorder of an item that was never missing. It appears as a normal replenishment order, indistinguishable from a legitimate one.

The Cascade Once the Gap Opens

The duplicate spend is rarely the only cost, and often not even the most disruptive one. Phantom replenishment ships product a territory does not actually need. That territory then shows excess in the next reporting period. When a rep in that same territory later submits a legitimate replenishment request, ops looks at the numbers, sees apparent excess, and flags the request instead of approving it, because the system says there should be plenty. A rep who genuinely needs product is now fighting a data problem that has nothing to do with their actual position.

And the unverified return does not stay an accounting problem. A rep opens a tray at a hospital and finds something missing that the return process should have caught. In a non-ortho case, a specialist reaches for a placed item the record showed as present and in-date and finds it missing or expired at the point of use. Sometimes the case proceeds with a substitute. Sometimes it does not. Either way, that is not an internal operational variance anyone can quietly resolve later. That is a customer experience event, in front of a surgeon and a hospital, and it traces back to the same unverified return that eventually shows up as a write-off months afterward.

Not Every Variance Is Real Loss

There is a mirror-image version of this problem that is worth naming, because it wastes the same investigation labor while pointing in the opposite direction. Product management changes a tray configuration, or an equipment set’s expected contents. Items get added or removed for good operational reasons. The change never propagates to the configuration of record in the system. Every audit afterward flags variances against a configuration that has not been current for months. Someone spends two days investigating, and the conclusion is that the change happened a long time ago and nobody updated the record.

That is a phantom variance in the other direction: the physical inventory is correct and the system’s expectation is wrong. It matters here because it means a variance count on its own does not distinguish between real loss, recovered inventory, and a stale configuration. All three land in the same bucket unless the system is built to tell them apart.

Why More Visibility Can Make This Worse

The instinctive response to all of this is more visibility: better dashboards, more real-time reporting, cleaner interfaces. That instinct is understandable and, when the underlying data is unreliable, actively counterproductive.

A dashboard renders whatever data it receives with the same confidence. Nothing in the presentation signals that a reconciliation event papered over a discrepancy instead of resolving it, or that a found item was quietly written back on after being written off. A cleaner interface makes a recovered variance and a true match look identical, because both read as in stock once the dust settles. Worse, clean-looking numbers discourage the investigation that would separate real accuracy from self-correcting churn, so the write-off/write-on cycle runs underneath a dashboard that looks perfectly healthy.

Why the Overall Accuracy Number Hides This

There is a version of this problem that is easy to miss because it is hiding inside a number that looks good. A field audit reports 96 percent accuracy. That sounds like a strong result, and in a narrow sense it is. But a single blended accuracy figure can conceal two very different things happening underneath it. Some portion of that 96 percent reflects inventory that was exactly where the system said it would be, in the expected quantity and condition. That is real accuracy. Another portion may reflect items that were missing, then found again during the same audit, or written back on after being written off. Once they are back on the books, they count toward the accuracy figure too. But being found is not the same as never having been lost, and a variance that self-corrects during the audit is not the same as a location that was never wrong in the first place.

This distinction is not theoretical. It maps directly to how audit data should be structured in the first place: expected quantity and counted quantity as separate tracked values, missing items and excess items logged as distinct outcomes rather than netted against each other, each with its own resolution status. A single blended percentage throws away exactly the information that structure is built to preserve. This is an illustrative breakdown, not a benchmark figure. The exact split will vary by company. The point is structural: a single blended accuracy score treats recovered variances and true matches as the same thing, when they carry very different risk profiles.

Why This Compounds Instead of Resolving

The reason this problem persists rather than self-correcting comes down to timing. Most companies rely on scheduled full field audits, typically annual, where an internal audit team physically counts inventory at each location. Between those audits, the system’s record and physical reality can drift for a long time without anyone noticing, because nothing in daily operations forces a reconciliation. That means a ghost inventory event discovered in this year’s audit did not start this year. It has likely been sitting in that gap since the last one, quietly generating downstream effects the whole time: a phantom replenishment order here, a flagged legitimate request there, a rep telling ops they simply do not have something because nobody can verify either way until someone shows up and counts. This is not a one-time correction. It is a cycle that resets every audit period.

What This Actually Costs

Once ghost inventory surfaces, whether at audit or by accident, the consequences are specific and repeatable:

• Unnecessary replenishment. A second unit gets ordered and shipped to fill a gap that never physically existed, creating excess in the field nobody needed.
  
  

• Wasted audit labor. Every variance triggers an investigation. When the variance is a phantom rather than a genuine loss, the labor spent tracing it produces no operational insight, only the eventual reversal of a write-off that should never have been made.
  
  

• E&O accumulation downstream. The excess inventory shipped to cover a phantom gap does not disappear. It sits in the field, ages, and becomes tomorrow's excess and obsolete exposure, with a root cause that traces back to a data problem rather than a demand problem.
  
  

• Financial misattribution. When an undocumented transfer moves inventory without updating ownership and billing along with location, commission calculations and territory revenue reports get built on a custody record that is already wrong, and someone has to untangle the billing after the fact.
  
  

• Buffer inventory as a coping mechanism. When variance becomes routine and unexplainable, the practical response is to place extra units in the field to absorb it. That buffer is itself deployed capital, sitting in the field for no reason other than that nobody trusts the count.
  
  

• Eroded confidence in the numbers. Once operations and finance have both been burned by chasing phantom variances, some sales leaders stop using system inventory data for planning entirely and start running plans off rep call notes instead. That is not a system anymore. That is a parallel process running alongside the platform, at a real cost, in a company that is already paying for the platform it stopped trusting.

Why This Is Not a People Problem

It is tempting to attribute this to reps not documenting transfers, warehouse staff not double-checking a return, or auditors missing a variance. That framing puts the fix in the wrong place. People are working within whatever the system makes fastest and easiest, and in most field inventory operations today, checking a box is still faster than actually comparing a tray against its bill of materials or an equipment set against its expected configuration.

The fix is not more diligence layered on top of the same workflow. It is closing the gap between when inventory moves and when the system knows about it, so that what the system believes and what physically exists stop drifting apart at the exact moment they should be reconciling, the moment something comes back under control. That is a workflow enforcement problem, not a training problem, and it is the only kind of fix that actually holds between audit cycles instead of resetting the clock until the next one.

Three Questions Worth Asking Your Own Team

If you want to know whether this is happening in your operation, the diagnostic is short:

• Has your team written off inventory in the last 12 months that later reappeared? If yes, you have a documented instance of this cycle, not a hypothetical one.

• Does a catch-all or holding location exist in your system, and has anyone reconciled it this year? The answer to the second half is usually no.

• Of your last 50 returns, whether loaner kits or equipment coming back from a placement, how many included an actual comparison against the expected contents rather than a completion checkbox? Most teams do not know, and that uncertainty is itself the finding.

In the next part of the series, we follow a single item all the way through the loop, and show exactly how “lost” becomes “found” becomes reordered, and where the cycle can be broken.

Frequently Asked Questions

• What is ghost inventory in medical device field operations?
  Ghost inventory is product that is recorded but gone, or present but invisible. It shows as an orthopedic component written off after a case and later found in another rep’s trunk, or a placed unit written off during a territory transition that was at the customer site all along. Either way, the record and reality have split, and the company often pays to replace product it already owns.
  
  

• How does ghost inventory lead to duplicate spend?
  When an item is missing from where the system expects it, supply chain sees a gap and ships a replacement to close it. The original was never truly lost, just somewhere else, but by the time it resurfaces a second unit has already been ordered and deployed. The reorder looks like a normal replenishment on the invoice, which is why the duplicate spend stays hidden.
  
  

• Why do unverified returns cause inventory write-offs?
  A return marked complete without a real check is where the gap usually opens. If no one compares a returned kit against its bill of materials, or returned equipment against its expected configuration, missing and misplaced items go unnoticed at the exact moment inventory comes back under control. The discrepancy then surfaces later as a write-off or a shortage.
  
  

• Is a high inventory accuracy score reliable?
  Not on its own. A single blended figure like 96 percent can mix inventory that was exactly where it should be with items that were missing and then found during the same audit. Being found is not the same as never having been lost, so a blended score can hide recurring variances unless expected and counted quantities, and missing versus excess items, are tracked separately.
  
  

• How do I know if my team has a ghost inventory problem?
  Three quick checks: whether you have written off inventory in the last 12 months that later reappeared; whether a catch-all or holding location exists and has actually been reconciled this year; and how many of your last 50 returns included a real comparison against expected contents rather than a completion checkbox. Not knowing the last one is itself the finding.