Your ‘Out of Stock’ Is Probably a Lie: The RFID + Camera Stack Quietly Rewriting What Fashion Stores Call “Inventory” (and Why It’s About to Change Prices)
A store’s system can say “in stock” while the rack is empty—and that gap is called phantom inventory. RFID fixes the record; cameras verify shelf truth, and the stack is redefining what “available” even means.
By TheMurrow Editorial
May 14, 2026
Key Points
- 1Recognize phantom inventory: many “out of stock” moments are record drift, misplacement, shrink timing, or backroom execution—not true scarcity.
- 2Understand what RFID really changes: it makes item-level counting fast and frequent, pushing barcode-era accuracy (~65%) toward ~95–99% in mature programs.
- 3Watch the next layer: cameras verify on-rack reality in real time, pairing with RFID to redefine “in stock” as present, located, and sellable.
A shopper stands under bright lights, staring at an empty rail where a popular size should be. The associate checks a handheld device, frowns, and delivers the familiar verdict: “Out of stock.”
Only it often isn’t. The missing shirt may be sitting in the backroom under unworked freight. It may be mis-slotted on the wrong rack. It may be “available” in the system because theft, damage, or late adjustments haven’t caught up. The shopper hears scarcity; the store is suffering something more mundane: a bookkeeping failure.
Retail researchers and vendors have a name for the mismatch—phantom inventory, or inventory record inaccuracy (IRI)—and it helps explain why “out of stock” has become one of the most misleading phrases in modern shopping. The real problem isn’t always supply. It’s visibility: knowing what you have, where it is, and whether it’s actually sellable right now.
“Many stockouts aren’t a lack-of-goods problem. They’re a ‘we lost it in our own building’ problem.”
— — TheMurrow Editorial
The modern “out of stock” problem is often a system failure
A traditional stockout sounds simple: demand exceeded supply. In apparel retail, many stockouts are more embarrassing than that. Store systems may show units on hand, yet the shopper can’t find the item on the floor. Research and industry literature increasingly describe these incidents as system stockouts—the store’s records say the product exists, but the shelf or rack says otherwise. Edge-AI shelf monitoring vendors explicitly frame this gap as a driver of missed sales and customer frustration, noting that “out of stock” often reflects a detection failure more than true depletion. (CamThink, on out-of-stock detection and phantom inventory)
What “out of stock” can really mean in apparel
The label can conceal several operational realities:
- The item is in the backroom, buried in unworked freight or placed in the wrong location.
- The item is somewhere on the sales floor, but misplaced on an incorrect fixture, mixed into the wrong size run, or stranded in fitting-room go-backs.
- The item is gone—stolen or damaged—yet still counted as available because shrink and adjustments lag behind reality.
- The item is in transit or “received” in a system but not physically staged, processed, or accessible to sell.
None of these cases feel like “inventory management” to the shopper. They feel like indifference. For retailers, they represent the quieter kind of loss: sales that vanish not because the goods aren’t there, but because the store can’t produce them on demand.
- The item is in the backroom, buried in unworked freight or placed in the wrong location.
- The item is somewhere on the sales floor, but misplaced on an incorrect fixture, mixed into the wrong size run, or stranded in fitting-room go-backs.
- The item is gone—stolen or damaged—yet still counted as available because shrink and adjustments lag behind reality.
- The item is in transit or “received” in a system but not physically staged, processed, or accessible to sell.
None of these cases feel like “inventory management” to the shopper. They feel like indifference. For retailers, they represent the quieter kind of loss: sales that vanish not because the goods aren’t there, but because the store can’t produce them on demand.
Why this is worse than a simple empty shelf
A true stockout is an upstream planning problem. A system stockout is a store execution problem that can persist even when replenishment is adequate. It also undermines ecommerce promises like “pick up in store,” since those services depend on the idea that records match reality.
The larger implication is uncomfortable: in many stores, the “truth” lives not in a database but in the physical space—and the two drift apart more often than most shoppers assume.
The larger implication is uncomfortable: in many stores, the “truth” lives not in a database but in the physical space—and the two drift apart more often than most shoppers assume.
Inventory record inaccuracy is structural, especially in apparel
Inventory record inaccuracy is not a new scandal uncovered by tech vendors. It is a long-studied feature of retail operations. Academic work in this area—including research streams associated with scholars such as DeHoratius and Raman, and field research tied to the Auburn University RFID Lab—treats IRI as predictable, category-dependent, and shaped by everyday store conditions: complexity, velocity, and human handling.
Apparel has all the ingredients that make records hard to keep accurate. A single “product” can mean dozens of sizes and colors. Items move constantly: from stockroom to floor, to fitting room, back to a return bin, to a different table, and sometimes out the door unpaid. The operational reality is that a store is not a warehouse; it’s a busy, customer-accessible space where inventory gets touched, tried on, and relocated all day.
Apparel has all the ingredients that make records hard to keep accurate. A single “product” can mean dozens of sizes and colors. Items move constantly: from stockroom to floor, to fitting room, back to a return bin, to a different table, and sometimes out the door unpaid. The operational reality is that a store is not a warehouse; it’s a busy, customer-accessible space where inventory gets touched, tried on, and relocated all day.
The “65% accuracy” baseline—and why it matters
Industry-facing summaries often cite a sobering shorthand: average inventory accuracy around ~65% in the barcode era, frequently attributed to findings associated with Auburn University RFID Lab discussions and widely repeated in RFID commentary. (Cybra, on average retailer inventory accuracy)
That number deserves careful handling. It’s a general baseline, not a universal constant. Accuracy varies by retailer, category, and measurement method. Still, the figure persists because it captures a truth many operators recognize: barcode processes often deliver something like “good enough for accounting,” not “good enough for a shopper hunting a size Medium at 6 p.m.”
That number deserves careful handling. It’s a general baseline, not a universal constant. Accuracy varies by retailer, category, and measurement method. Still, the figure persists because it captures a truth many operators recognize: barcode processes often deliver something like “good enough for accounting,” not “good enough for a shopper hunting a size Medium at 6 p.m.”
“A system can be ‘accurate enough’ for finance and still fail at the fitting room.”
— — TheMurrow Editorial
Why apparel stores drift out of sync
Operations research points to practical drivers:
- Infrequent counting: manual cycle counts take time and disrupt the floor.
- High handling: garments are moved, refolded, and rehung constantly.
- Complexity: sizes, styles, and presentations multiply the chances of misplacement.
- Shrink and timing gaps: losses or damages may be recognized later, leaving records stale.
The story here isn’t malice or incompetence. It’s that the old tools—barcodes and periodic manual counts—were never designed to produce near-real-time shelf truth in a dynamic store.
- Infrequent counting: manual cycle counts take time and disrupt the floor.
- High handling: garments are moved, refolded, and rehung constantly.
- Complexity: sizes, styles, and presentations multiply the chances of misplacement.
- Shrink and timing gaps: losses or damages may be recognized later, leaving records stale.
The story here isn’t malice or incompetence. It’s that the old tools—barcodes and periodic manual counts—were never designed to produce near-real-time shelf truth in a dynamic store.
~65%
A commonly cited barcode-era baseline for average inventory accuracy—useful as a shorthand for how far records can drift from physical reality. (Cybra; Auburn RFID Lab commentary)
What RFID actually fixes: item-level visibility, faster counting
Radio-frequency identification is often discussed as if it “solves inventory.” A more accurate claim is narrower and more persuasive: RFID makes it practical to count items frequently and reconcile records to reality. The major advantage is straightforward: RFID can identify tagged items without line-of-sight scanning, unlike barcodes. That changes the labor math of counting.
The Auburn University RFID Lab has published work on RFID-enabled visibility and retail inventory record inaccuracy, focusing on field experiments and the practical impact of RFID on store accuracy. (Auburn RFID Lab)
The Auburn University RFID Lab has published work on RFID-enabled visibility and retail inventory record inaccuracy, focusing on field experiments and the practical impact of RFID on store accuracy. (Auburn RFID Lab)
The numbers retailers chase: from ~65% to the high 90s
In industry guidance and implementation literature, mature apparel RFID programs are commonly associated with moving from barcode-era accuracy—often described in the ~65–75% range—to ~95–99% item-level accuracy, assuming disciplined processes and solid tag compliance. (CPCON Group, RFID clothing tracking guide)
Those ranges are broad for a reason. “Accuracy” depends on definitions and methods. Some retailers measure accuracy as whether the system correctly reflects “in store” presence. Others require correct location, correct size, and correct sellable status. RFID improves the first problem dramatically; the rest still require work.
Those ranges are broad for a reason. “Accuracy” depends on definitions and methods. Some retailers measure accuracy as whether the system correctly reflects “in store” presence. Others require correct location, correct size, and correct sellable status. RFID improves the first problem dramatically; the rest still require work.
“RFID doesn’t make stores perfect. It makes counting cheap enough to tell the truth more often.”
— — TheMurrow Editorial
Why frequency matters more than perfection
RFID’s practical superpower is not that it never makes mistakes. It’s that it enables frequent cycle counts, which reduce the time inventory has to “go stale.” A store that can validate counts weekly—or even more often—will correct record drift sooner than a store that counts quarterly or only during annual inventory.
For shoppers, the downstream effect is fewer maddening moments when an associate’s handheld insists the item exists, but the rack tells a different story.
For shoppers, the downstream effect is fewer maddening moments when an associate’s handheld insists the item exists, but the rack tells a different story.
~95–99%
A commonly reported range for item-level inventory accuracy in mature apparel RFID programs with disciplined processes and strong tag compliance. (CPCON Group)
RFID isn’t magic: the unglamorous limits retailers run into
RFID is often marketed as a clean technological leap. On the ground, it behaves like every other store system: it succeeds when the process is respected and fails when the details are ignored.
Tagging compliance and encoding quality
Item-level RFID depends on tags being applied correctly and encoded consistently, often across a complex vendor ecosystem. That is why many retailers rely on mandates and compliance programs. A single weak link—incorrect tags, missing tags, or inconsistent data—creates blind spots that look like accuracy problems downstream.
Read rates and physical environments
Stores aren’t friendly laboratories. Dense racks, metal fixtures, and layout quirks can degrade read performance. Even when the system improves dramatically over barcodes, the final few percentage points of accuracy can become expensive, not because the technology is flawed, but because the environment is.
“In the building” is not “available to buy”
RFID can tell you an item is present somewhere, but shoppers need it:
- on the correct rack,
- in the correct size run,
- with the correct ticketing and presentation,
- accessible without an associate scavenger hunt.
RFID narrows the search radius. It does not automatically guarantee shelf discipline, prompt replenishment, or tidy go-backs. Shrink still happens, too—RFID can improve detection and reconciliation, but it cannot prevent every loss.
The credibility test for RFID is whether retailers speak honestly about these limits. The best programs treat RFID as infrastructure that exposes operational truth—and then build the habits to act on that truth.
- on the correct rack,
- in the correct size run,
- with the correct ticketing and presentation,
- accessible without an associate scavenger hunt.
RFID narrows the search radius. It does not automatically guarantee shelf discipline, prompt replenishment, or tidy go-backs. Shrink still happens, too—RFID can improve detection and reconciliation, but it cannot prevent every loss.
The credibility test for RFID is whether retailers speak honestly about these limits. The best programs treat RFID as infrastructure that exposes operational truth—and then build the habits to act on that truth.
Key Insight
RFID strengthens the inventory record (“it exists”), but retail execution determines sellability (“it’s on the rack, in the right place, right now”).
Why retailers are adding cameras: inventory accuracy isn’t the same as shelf truth
If RFID makes counting easier, why do some retailers look to cameras and computer vision? Because customers don’t shop the stockroom. They shop the floor.
A recurring complaint in store operations is the gap between inventory accuracy (“the item exists in the store”) and on-shelf or on-rack availability (“the item is actually in the selling location, right now”). Vendors promoting shelf-monitoring and edge-AI cameras position their systems as a way to detect out-of-stock conditions and phantom inventory scenarios that traditional systems miss. (CamThink)
A recurring complaint in store operations is the gap between inventory accuracy (“the item exists in the store”) and on-shelf or on-rack availability (“the item is actually in the selling location, right now”). Vendors promoting shelf-monitoring and edge-AI cameras position their systems as a way to detect out-of-stock conditions and phantom inventory scenarios that traditional systems miss. (CamThink)
Periodic counts vs. continuous observation
Even frequent RFID cycle counts are still periodic. A store can count in the morning and be messy again by noon. Returns arrive. Fitting rooms overflow. Popular sizes migrate. Human beings rearrange tables with good intentions and poor follow-through.
Computer vision systems, by contrast, are pitched as continuous. Instead of asking, “Did we count this item this week?” the camera asks, “Is the shelf empty right now?” That’s a different question, closer to the shopper’s reality.
Computer vision systems, by contrast, are pitched as continuous. Instead of asking, “Did we count this item this week?” the camera asks, “Is the shelf empty right now?” That’s a different question, closer to the shopper’s reality.
What cameras can see that RFID often can’t
RFID is excellent at presence detection. Cameras can be better at:
- confirming whether a specific fixture or shelf position is empty,
- identifying presentation issues (depending on the system’s sophistication),
- alerting staff to real-time gaps on the floor.
Cameras, of course, raise their own issues—privacy sensitivity, installation complexity, and the risk of false positives in messy environments. Still, the motivation is clear: retailers want a system that reflects not just “inventory,” but sellability.
- confirming whether a specific fixture or shelf position is empty,
- identifying presentation issues (depending on the system’s sophistication),
- alerting staff to real-time gaps on the floor.
Cameras, of course, raise their own issues—privacy sensitivity, installation complexity, and the risk of false positives in messy environments. Still, the motivation is clear: retailers want a system that reflects not just “inventory,” but sellability.
“Continuous”
The core promise of shelf-monitoring cameras: not just whether the store owns the item, but whether the shopper-facing fixture is empty right now.
The “RFID + camera” stack: toward a single version of store reality
The most compelling argument for combining RFID and cameras is that each solves a different layer of the same problem. RFID strengthens the inventory record. Cameras verify the customer-facing truth.
In vendor narratives, the promise is a store that can answer three questions with confidence:
1. Do we have it? (RFID-backed inventory record)
2. Where is it? (RFID location workflows, backroom vs. floor)
3. Can a shopper buy it right now? (camera-confirmed shelf/rack availability)
In vendor narratives, the promise is a store that can answer three questions with confidence:
1. Do we have it? (RFID-backed inventory record)
2. Where is it? (RFID location workflows, backroom vs. floor)
3. Can a shopper buy it right now? (camera-confirmed shelf/rack availability)
How the stack addresses “phantom inventory”
Phantom inventory thrives in the gaps—when systems say “yes” and the floor says “no.” RFID reduces the size of those gaps by improving record accuracy. Cameras reduce them by monitoring whether the selling location is actually stocked.
That combination targets a familiar failure mode: the associate who trusts the device, the shopper who distrusts the store, and the item that exists somewhere but not where commerce happens.
That combination targets a familiar failure mode: the associate who trusts the device, the shopper who distrusts the store, and the item that exists somewhere but not where commerce happens.
A practical example: the missing size on a busy weekend
Consider a common scenario: a popular denim style in a fast-selling size. The system shows three units on hand.
- If RFID counts are current, the store knows those three items are physically present.
- If the shelf is empty, a camera alert can push replenishment: the units are probably in the backroom, in returns, or mis-racked.
- If a replenishment search fails, the discrepancy becomes actionable: investigate shrink, damages, or tagging errors.
None of that guarantees a perfect outcome. The point is to turn a vague “out of stock” into a specific operational diagnosis with a next step.
- If RFID counts are current, the store knows those three items are physically present.
- If the shelf is empty, a camera alert can push replenishment: the units are probably in the backroom, in returns, or mis-racked.
- If a replenishment search fails, the discrepancy becomes actionable: investigate shrink, damages, or tagging errors.
None of that guarantees a perfect outcome. The point is to turn a vague “out of stock” into a specific operational diagnosis with a next step.
Turning “out of stock” into an operational diagnosis
- 1.Confirm item presence with current RFID counts
- 2.Verify whether the selling fixture is empty via camera/shelf monitoring
- 3.Trigger replenishment search (backroom, returns, mis-racks)
- 4.Escalate discrepancies to shrink/damage/tagging investigation
What this means for shoppers—and for retailers that want trust
For shoppers, the immediate takeaway is oddly reassuring: many “out of stock” moments aren’t proof that a product is rare or that demand forecasting failed. They are evidence that physical retail is hard, and that the store may not have mastered the last 50 feet between backroom and rack.
For retailers, the lesson is sharper. Inventory visibility has become part of brand credibility. When systems lie, customers assume the store is disorganized or dismissive. When pickup orders get canceled because the item “can’t be found,” customers blame the retailer, not the complexity of store operations.
For retailers, the lesson is sharper. Inventory visibility has become part of brand credibility. When systems lie, customers assume the store is disorganized or dismissive. When pickup orders get canceled because the item “can’t be found,” customers blame the retailer, not the complexity of store operations.
Practical implications for shoppers
A few strategies follow from how system stockouts work:
- Ask whether the item is in the backroom or tied up in returns/go-backs, not just “in stock.”
- If shopping across channels, treat “available for pickup” as a probability, not a promise, unless the retailer has strong item-level processes.
- When a store offers to check other locations, remember that another store’s records can suffer the same drift.
- Ask whether the item is in the backroom or tied up in returns/go-backs, not just “in stock.”
- If shopping across channels, treat “available for pickup” as a probability, not a promise, unless the retailer has strong item-level processes.
- When a store offers to check other locations, remember that another store’s records can suffer the same drift.
What to ask when you hear “out of stock”
- ✓Is it in the backroom under unworked freight?
- ✓Could it be in returns/go-backs or fitting-room racks?
- ✓Was it received in the system but not processed to the floor?
- ✓Can you check other locations—and how current are their counts?
Practical implications for retail operators
The research-backed operational message is not “buy tech.” It’s “choose what you want to know, then build the discipline to act.”
- If the goal is record accuracy, RFID-supported frequent counting has a strong track record in apparel research and industry reporting (Auburn RFID Lab; CPCON Group).
- If the goal is on-shelf availability, you need workflows that convert alerts into action—whether those alerts come from cameras, staff audits, or both.
- If the goal is trust, publish more honest availability signals and reduce the gap between what the website says and what the floor can deliver.
The future of “in stock” may be less about a single database number and more about a hierarchy of truths: present, located, sellable.
- If the goal is record accuracy, RFID-supported frequent counting has a strong track record in apparel research and industry reporting (Auburn RFID Lab; CPCON Group).
- If the goal is on-shelf availability, you need workflows that convert alerts into action—whether those alerts come from cameras, staff audits, or both.
- If the goal is trust, publish more honest availability signals and reduce the gap between what the website says and what the floor can deliver.
The future of “in stock” may be less about a single database number and more about a hierarchy of truths: present, located, sellable.
The hierarchy of “in stock” shoppers actually need
Present: it exists in the building.
Located: staff can find it fast.
Sellable: it’s on the right rack/fixture, right size run, accessible now.
Located: staff can find it fast.
Sellable: it’s on the right rack/fixture, right size run, accessible now.
The new meaning of “in stock”: a promise retailers must earn
Retail has always run on imperfect information. What’s changed is customer tolerance. Shoppers now arrive with screenshots, online size checks, and pickup expectations. An empty rack no longer feels like bad luck; it feels like a broken contract.
The most useful way to think about RFID and camera systems is not as flashy retail tech, but as tools for aligning the digital and physical versions of a store. RFID narrows the gap between records and reality. Cameras narrow the gap between reality and the shelf.
If the work is done well, “out of stock” becomes rarer—and more truthful when it appears. That is a modest ambition with outsized consequences: fewer wasted trips, fewer canceled pickups, fewer awkward associate interactions, and a store that feels competent again.
The quiet victory isn’t perfect prediction. It’s operational honesty—knowing what you have, where it is, and whether a shopper can actually buy it.
The most useful way to think about RFID and camera systems is not as flashy retail tech, but as tools for aligning the digital and physical versions of a store. RFID narrows the gap between records and reality. Cameras narrow the gap between reality and the shelf.
If the work is done well, “out of stock” becomes rarer—and more truthful when it appears. That is a modest ambition with outsized consequences: fewer wasted trips, fewer canceled pickups, fewer awkward associate interactions, and a store that feels competent again.
The quiet victory isn’t perfect prediction. It’s operational honesty—knowing what you have, where it is, and whether a shopper can actually buy it.
“The quiet victory isn’t perfect prediction. It’s operational honesty—knowing what you have, where it is, and whether a shopper can actually buy it.”
— — TheMurrow Editorial
T
About the Author
TheMurrow Editorial is a writer for TheMurrow covering style & fashion.
Frequently Asked Questions
What is “phantom inventory” in retail?
Phantom inventory describes inventory that a retailer’s system shows as available even though the item can’t be found where it should be. The product may be misplaced, stuck in the backroom, lost to theft or damage, or tied up in timing gaps between receiving and floor placement. The result is a “system stockout” that feels like a real stockout to shoppers.
How inaccurate are retail inventory records, really?
Industry summaries commonly cite around ~65% average inventory accuracy in barcode-driven retail environments, often linked in commentary to Auburn University RFID Lab-related baselines (Cybra). The exact figure varies by retailer, category, and how “accuracy” is defined. The larger point holds: many stores operate with records that drift significantly from physical reality.
How does RFID improve inventory accuracy?
RFID allows item identification without line-of-sight scanning, making cycle counts faster and more frequent than barcode counting. Research and industry reporting associated with the Auburn University RFID Lab emphasize RFID’s impact on reducing inventory record inaccuracy through better visibility and reconciliation. Many apparel deployments report ~95–99% item-level accuracy under mature, disciplined processes (CPCON Group).
If RFID is so good, why do stores still have empty shelves?
RFID can confirm that an item is in the store, but shoppers need it on the rack in the right place, correctly sized and accessible. Items can be trapped in backrooms, mixed into wrong fixtures, or stuck in returns. Store execution—replenishment, go-backs, and merchandising discipline—still determines what customers experience on the floor.
What do cameras add that RFID doesn’t?
Computer vision systems are often pitched as monitoring on-shelf/on-rack availability continuously—whether the selling location is actually stocked right now. RFID typically supports periodic counting and presence detection, while cameras can flag real-time empty fixtures or presentation problems. Vendors frame this combination as a way to reduce “phantom inventory” and misleading out-of-stock signals (CamThink).
Are RFID and camera systems a privacy concern?
Any in-store camera deployment raises legitimate privacy questions, especially if shoppers or staff could be identifiable. Many retail camera use cases are described as shelf/fixture monitoring rather than facial identification, but governance matters: what is captured, how long it’s stored, and who can access it. Retailers that deploy cameras responsibly need clear policies and transparent communication.
