- Ben Merva
Global inventory distortion—out-of-stocks plus overstocks—runs about $1.73 trillion a year, roughly 6.5% of global retail sales, per analyst firm IHL Group. Zebra Technologies puts the shopper side at 52%, leaving a store without everything they came for. Yet almost everything written about on-shelf availability stops at the retailer’s front door.
That’s the gap. Search the term, and you get shelf cameras, electronic shelf labels, planogram checklists, night-shift staffing advice—all written for retailers, not you, the consumer brand selling through those retail stores.
For a brand, on-shelf availability isn’t a store-operations metric. It’s a revenue and supply chain problem with a relationship problem attached. Lost sales you never see. Trade spend burned on a promotion that ran dry on day two. A buyer who quietly stops trusting your forecasts.
None of it shows in a shelf photo. It shows in POS and inventory data from your retailers, at SKU and store level. This guide covers what on-shelf availability is, how to calculate and benchmark it, what causes poor on-shelf availability, and how to improve on-shelf availability with real-time data and agentic AI.
Table of Contents
What is on-shelf availability (OSA)?
On-shelf availability (OSA) is the percentage of time a listed product is physically present and available to buy on the shelf when a shopper looks for it. It measures the moment of purchase, when revenue is generated. A product can read as “in stock” and still have 0% shelf availability.
That last sentence is the ballgame. On-shelf availability is a shelf-level execution metric, not a supply chain one. Inventory in a DC, in transit, in the backroom, or in the wrong location doesn’t count.
Most teams treat on-shelf availability as one number reported monthly. It’s a daily signal that behaves differently across a particular category, a specific store, or a promotional window. The four types of retail analytics put OSA in context, and most retail analytics software never models it from the brand’s side.
On-shelf availability vs in stock vs out of stock
These three are used interchangeably in buyer meetings, which causes trouble.
In stock: inventory exists somewhere in the store or the retailer’s inventory system—backroom, overflow rack, the wrong location.
On-shelf availability: the product is on its assigned store shelf, priced, faced, and buyable now.
Out of stock: it’s missing from the shelf entirely.
A store can report 100% in stock and deliver 0% shelf availability on your top-selling SKUs—the most common poor OSA brands never catch. Shelf availability and out-of-stock are mirror opposites: if OSA is 92%, the out-of-stock rate is 8%.
How to calculate on-shelf availability (formula)
OSA (%) = (SKUs available on shelf ÷ total listed SKUs) × 100
A brand has 100 listed SKUs in a chain. On a given day, 92 are physically present on the store shelf. On-shelf availability is 92%; out-of-stock rate is 8%.
Simple math, harder execution. The number only means something when calculated daily at the SKU/store level. Measure monthly at the account level, and you get a comfortable 96%, hiding a chronic 78% in one region.
Why backroom stock doesn’t count
Backroom stock is the most frustrating lost availability: the inventory exists, retailers paid for it, and the shopper still walks out empty-handed. Product sitting on a pallet fifty feet from the aisle contributes nothing. Your retailer’s inventory records say the units are in the building. They are. They’re just not where the money is.
Book a Demo
See Alloy.ai
Synchronize execution to eliminate waste, mitigate risk, and capture every revenue opportunity.
Once you submit a demo request form:
- A sales rep will email you within 24 hours to schedule a brief intro call.
- Our team will provide an overview of Alloy.ai and learn more about your data goals and objectives.
- We’ll create and walk you through a live demo, customized to your specific data needs.
Schedule a Demo
Why on-shelf availability matters for consumer brands
Poor on-shelf availability is a silent revenue leak, and the sale is rarely postponed. In the largest study of shopper reactions to stockouts, only 15% delay the purchase—31% buy at another store, 26% switch to a competing brand. That makes shelf availability a P&L line item, not an operations footnote.
Lost sales and the customer satisfaction hit you never see
A shopper sees a gap where your offering should be, shrugs, and buys from your competitor. You lose that purchase, and possibly the next twelve, because switching costs are near zero in most retail environments. That’s how a customer satisfaction problem becomes a customer loyalty problem, and customer satisfaction is the part nobody puts a number against.
Greg Buzek, president of IHL Group, describes North American retail as stuck in “the whack-a-mole game of solving one crisis only to have another emerge.” Your brand feels it at the shelf first.
Wasted trade spend and shrinking shelf space
You negotiated the discount and paid for the endcap. Then the SKU went out of stock on day two, and seven days of premium shelf space sold air.
Worse, the event reads as a failure in the post-mortem. Buyers reallocate shelf space on performance, rarely adjusting for the fact you couldn’t sell what wasn’t there. Chronic availability issues lose facings your brand spent years earning.
Distorted demand signals and unreliable inventory data
Quietly, the most expensive one. When a product is unavailable during a key window, sell-through reads as soft demand. Your planner lowers the forecast. Smaller order, more out of stock, tighter loop. Breaking it means separating “didn’t sell” from “couldn’t sell”—what unconstrained demand modeling in CPG analytics is for.
Eroded customer trust and buyer confidence
Retail brands live and die on retailer relationships. A buyer who hears about your out-of-stock from a store manager, not from you, has learned something unflattering about your visibility. Long-term success at any account starts with knowing the gaps before the meeting.
See how Alloy.ai’s AI agents sense demand shifts and surface revenue opportunities across your retail network. Book a demo.
What causes poor on-shelf availability?
Poor on-shelf availability rarely comes from one supply failure. It comes from systemic issues that compound: inaccurate inventory records, stock stranded in the wrong place, forecasts built on the wrong signal, promotions that outrun replenishment, and data spread across too many retailer portals to act on.
Phantom inventory: when the inventory system says yes and the shelf says no
Phantom inventory is the retailer’s system showing units on hand while the shelf is empty. Because the system thinks stock exists, replenishment never triggers, and the gap persists for weeks.
Inventory management that stops at the DC
Plenty of brands run excellent inventory management to the retailer’s DC, then nothing. Product ships, the OTIF scorecard looks fine, and visibility ends.
But DC stock levels are not shelf availability. Inventory can sit at a DC while stores in the same region run dry. A demand-driven supply chain reads the shelf, not the dock—inventory management that stops at the dock door can’t protect availability.
Replenishment gaps and stranded stock in the backroom
The product made it into the store and never onto the shelf. Understaffed night crews, planogram resets, slow replenishment cadences, a case buried behind seasonal displays—all of it leaves stock physically inside the building and unavailable.
Corsten and Gruen’s research consistently shows most out-of-stocks originate in store-level ordering rather than upstream supply. Uncomfortable, because that’s the part you control least.
Forecasts built on shipments instead of sell-through
Plan on shipment data, and your supply chain reads a lagging signal. Orders tell you what retailers decided to buy weeks ago, not what shoppers do today, and nothing about current shelf stock levels.
Real-time data through daily POS forecasting at the SKU-location level closes that gap. Demand sensing is table stakes now.
Promotion-driven demand spikes
Demand outruns replenishment exactly when visibility matters most. Gruen and Corsten found out-of-stock rates on fast-moving and promoted products regularly push past 10%. The moment you’ve spent most to create demand is the moment on-shelf availability is likeliest to fail.
Siloed data across dozens of retailer portals
Sales and inventory levels arrive in different formats from Target, Kroger, Costco, Amazon Vendor Central, distributors, and 3PLs—different cadences, different definitions of “on hand.”
Most teams spend more time cleaning that data than acting on it. By then, the store visit that would have fixed it is two weeks late. Automated data collection is the unglamorous foundation everything sits on.
Book a Demo
See Alloy.ai
Synchronize execution to eliminate waste, mitigate risk, and capture every revenue opportunity.
Once you submit a demo request form:
- A sales rep will email you within 24 hours to schedule a brief intro call.
- Our team will provide an overview of Alloy.ai and learn more about your data goals and objectives.
- We’ll create and walk you through a live demo, customized to your specific data needs.
Schedule a Demo
How to measure and improve on-shelf availability
To improve on-shelf availability, brands track OSA daily at the SKU and store level, detect phantom inventory automatically, predict out-of-stocks before they happen, quantify lost sales in dollars, and automate the recovery order. Each step moves you from reporting a gap afterward to closing it before the sales window shuts.
You don’t improve on-shelf availability by measuring harder, but by shortening the distance between signal and action.
Track OSA daily across every store
Measure consistently, not occasionally. A monthly average will conceal a store that’s been out for eleven straight days.
Watch high-velocity SKUs and promoted items closely; monitor daily sales against on-hand stock levels—the divergence is your earliest warning. Brands that improve on-shelf availability fastest start narrow: the SKUs carrying 80% of volume.
Detect phantom inventory automatically
The fastest way to improve on-shelf availability is to stop trusting reported on-hand numbers. Compare them against actual sales velocity. When a specific store shows inventory but near-zero sales for an abnormal stretch, that’s likely phantom inventory, not soft demand.
Doing this manually across 20,000 stores isn’t a job for a spreadsheet. Automated, it surfaces availability issues weeks earlier, with lead time to follow up while the fix matters. One Alloy.ai customer: phantom inventory work and more intentional placement drove $4 million in sales.
Predict availability issues before they happen
Use forward-looking weeks of supply, replenishment cadence, and projected stock levels to sense which SKU/store combinations will go out of stock in the coming days. Rank those availability issues by dollar impact, not unit volume.
Fixing 200 low-velocity gaps moves nothing. Fixing the twelve carrying real revenue impact changes the quarter. Predictive analytics and forecasting automate that triage.
Quantify lost sales in dollars
Unconstrained demand estimates what POS data could have shown without the gap. Subtract actual, and you have a dollar figure for the empty shelf.
This is the most useful thing to bring a buyer. “We’re seeing low OSA in the Southeast” gets a nod. “This gap cost us both $187,000 over eleven days; here are the 340 stores, here’s the order” gets action. That turns a complaint into a revenue optimization conversation—and a path to recover lost sales.
Automate replenishment with agentic AI
Detection isn’t the finish. The retail replenishment AI agent monitors retailers 24/7, detects stock risks, calculates the order quantity, and auto-drafts the buyer email, charts included.
That’s the shift: fire drills become a review-and-approve workflow, where the team makes informed decisions instead of assembling evidence. For build-versus-buy, the complete retail analytics guide covers what to demand from any platform, and the Crisp data platform alternatives for CPG brands comparison is the short version.
See it on your own data. Watch agentic AI sense out-of-stocks and phantom inventory across your retailers, quantify what it takes to recover lost sales, and draft the order for approval. Book a demo.
On-shelf availability benchmarks: what good looks like
Strong on-shelf availability generally runs 95–98% across most FMCG and retail categories. Below 95%, lost sales become measurable, especially on high-velocity SKUs and promoted items where every hour of absence will impact sales. Below 90%, you have a buyer-trust problem as well as a revenue problem.
| OSA level | What it signals |
|---|---|
| 98-100% | Best-in-class execution on priority SKUs |
| 95–97% | Healthy; the typical strong FMCG/retail benchmark |
| 90–94% | Warning; measurable revenue leakage, especially on hero SKUs |
| Below 90% | Critical; significant lost sales and buyer-trust risk |
Realistic benchmarks for CPG retailers and brands
Nobody hits 100%. Not Coca-Cola, not anyone with a national footprint and physical supply chain.
The practical goal is high product availability on priority SKUs, measured daily, with fast follow-up when something slips. The worldwide out-of-stock average has sat near 8.3% for decades, putting typical shelf availability nearer 92% than the 98% most brands assume. ECR Retail Loss found that improving inventory accuracy can lift sales by up to 8%—inventory records and shelf availability are closely related, which most CPG retailers and brands underestimate.
Manual shelf audits, electronic shelf labels, and AI-powered OSA monitoring
Manual and photo-based shelf audits sample a few retail stores periodically. Shelf monitoring hardware watches specific fixtures. AI-powered OSA monitoring watches every single SKU/store combination continuously from POS and inventory signals, detects phantom inventory, and quantifies lost sales in dollars.
A field rep on a store visit gives ground truth for one store, one day. Electronic shelf labels and RFID tags improve shelf compliance and inventory accuracy in-store, but that’s an investment your retailers control, not you. Neither scales nationally.
| Capability | Manual/periodic shelf audits | AI-powered OSA monitoring |
|---|---|---|
| Coverage | Sampled stores, checked periodically | Every SKU-store combination, daily |
| Phantom inventory | Difficult to detect | Flagged from POS vs on-hand signals |
| Lost revenue | Not quantified | Quantified in dollars (unconstrained demand) |
| Prioritization | Manual judgment | Ranked automatically by $ impact |
| Action taken | Field visit or static report | AI agent drafts recovery orders |
| Scalability | Limited by labor | Scales across 450+ retail connections |
The honest take: audits are how you verify; continuous monitoring is how you find. Neither improves on-shelf availability alone—improvement comes from what you do in the 48 hours after a gap surfaces. Cover every store daily, then send field reps to the stores your data flagged. Corrective actions stop being guesswork.
How Alloy.ai improves on-shelf availability for CPG brands
Alloy.ai is a purpose-built retail intelligence platform that unifies POS, inventory, and supply chain data across every one of your retailers, then layers agentic AI on top to protect on-shelf availability at the SKU and store level, from sensing the gap to drafting the recovery order.
The data platform handles ingestion and normalization through 450+ pre-built connectors, giving daily SKU/store granularity across retail and ecommerce partners, distributors, 3PLs, and your ERP. Data export into SAP IBP and similar systems keeps supply chain performance and inventory management up to date with actual sell-through, instead of a quarter behind.
On top sit the capabilities for maintaining on-shelf availability: phantom inventory detection, unconstrained demand and lost-sales dollars, forward-looking weeks of supply, and cross-retailer scorecards you drill from account to store to SKU in seconds, kept up to date daily. The performance reporting AI agent explains the “why” behind shifts, so inventory management decisions are informed decisions.
Crayola, BIC, Valvoline, and Melissa & Doug are among the customers using the artificial intelligence suite to keep product on the shelf. Kayla Harriss, supply chain manager at Valvoline Global Operations, described the change after gaining end-to-end visibility: avoiding costly out-of-stocks with retailers and making sure “consumers will always find our products when they need them.”
Turning shelf availability into increased sales and a competitive advantage
Availability is one of the few areas where operational discipline converts directly into results. Alloy.ai customers routinely see 35%+ fewer out-of-stocks and up to 35% better forecast accuracy—showing up as increased sales, stronger customer loyalty, and better line reviews with their retailers.
That’s the competitive advantage. Not that your brand never has a gap, but that you find yours first, quantify it, and arrive at the buyer meeting with corrective actions drafted. Every recovered purchase is one your competitor didn’t get. The guide to building strong partnerships with retailers covers it.
Bringing it together
On-shelf availability is where your manufacturing, logistics, and trade spend either convert into revenue or quietly leak away. Every dollar creating demand assumes shelf availability you may not have.
For consumer brands selling through retailers, the availability that matters isn’t audited monthly from a sample of stores. It’s sensed daily across every store, quantified in dollars, prioritized by revenue impact, and acted on before the sales window closes. Product availability isn’t a KPI you report—it’s a revenue stream you defend.
Shelf cameras and store-execution apps solve problems for retailers. They don’t tell you which SKU sits in a backroom in 340 stores while your buyer wonders why the promotion underperformed.
See what AI-powered on-shelf availability looks like for your brand. Book a demo with Alloy.ai.
Turning shelf availability into increased sales and a competitive advantage
Availability is one of the few areas where operational discipline converts directly into results. Alloy.ai customers routinely see 35%+ fewer out-of-stocks and up to 35% better forecast accuracy—showing up as increased sales, stronger customer loyalty, and better line reviews with their retailers.
That’s the competitive advantage. Not that your brand never has a gap, but that you find yours first, quantify it, and arrive at the buyer meeting with corrective actions drafted. Every recovered purchase is one your competitor didn’t get. The guide to building strong partnerships with retailers covers it.
Frequently Asked Questions
What is on-shelf availability (OSA)?
On-shelf availability (OSA) is the percentage of time a listed product is physically present and available to buy on the shelf when a shopper looks for it. It measures the moment of purchase, when revenue is generated. A product can sit in the backroom or read as “in stock,” but if it isn’t on its assigned shelf, OSA is zero.
How is on-shelf availability calculated?
On-shelf availability is calculated by dividing the listed SKUs available on the shelf by the total listed SKUs, then multiplying by 100. If a category carries 100 listed SKUs and 92 are on the shelf, OSA is 92%, and the out-of-stock rate is 8%. Brands measure OSA daily at the SKU and store level.
What is the difference between on-shelf availability and in stock?
In-stock and on-shelf availability are different steps in the supply chain. In stock means inventory exists somewhere in the store or system: the backroom, an overflow area, the wrong shelf. Shelf availability means it’s on the assigned shelf and buyable now. A store can report an item in stock while its OSA is zero.
What is the difference between OSA and out of stock (OOS)?
On-shelf availability and out-of-stock are two sides of the same coin. OSA measures the share of listed products present and buyable on the shelf; OOS measures the share missing. If OSA is 92%, the out-of-stock rate is 8%. OSA is the better executive metric, tracking availability over time rather than isolated stockouts.
Book a Demo
See Alloy.ai
Synchronize execution to eliminate waste, mitigate risk, and capture every revenue opportunity.
Once you submit a demo request form:
- A sales rep will email you within 24 hours to schedule a brief intro call.
- Our team will provide an overview of Alloy.ai and learn more about your data goals and objectives.
- We’ll create and walk you through a live demo, customized to your specific data needs.