On-Shelf Availability: OSA Formula, Benchmarks & Root Causes

On-shelf availability (OSA) is the share of listed SKUs that a shopper can find and buy in the planned place on the shelf at the moment of a store visit. Brands and retailers track on shelf availability as a percentage to check whether shoppers can buy the full agreed assortment from the shelf.

Retailer scorecards and brand reports often show availability close to 100%, while shoppers in the same stores find empty spaces, because most reports take the numerator of the formula from inventory or POS data. This data shows that the store has the product somewhere, while the shelf itself may be empty. This guide explains how to calculate OSA from the shelf, what level to aim for, why shelves go empty and how to improve OSA.

What you’ll learn

The OSA formula, with a worked example you can copy into a report
How to set OSA benchmarks by channel and SKU velocity
Three layers of root causes: supply chain, in-store execution and data
Five ways to measure OSA and what each one misses
A six-step improvement program with a 90-day roadmap.

Table of Content:

Key Takeaways

OSA counts only what a shopper can find and buy on the shelf, so the numerator of the formula comes from a shelf check, and inventory data serves as a second source for finding phantom inventory.
Empty shelves have causes in three layers: the supply chain, in-store execution and data; and each layer needs its own owner, report and actions.
An OSA program raises sales when reps fix problems during the visit and the team sets separate targets by SKU velocity, store format and season.

What Is On-Shelf Availability (OSA)?

On-shelf availability (OSA) is a retail metric that shows the share of listed SKUs a shopper can find on the shelf and buy. Field teams measure OSA in each store at every visit, but stock in the backroom does not count.

On-Shelf Availability Definition

A SKU counts as available only if

it belongs to the agreed assortment from the minimum stock list (MSL) or planogram,
it is on the shelf in its planned place
it has at least one sellable unit within the shopper’s reach.

This is a binary SKU availability check. If the team also needs to measure the required number of facings or shelf quantity, record those as separate compliance metrics.

If one condition fails, for example, the product is in the wrong bay, the SKU counts as unavailable for that visit.

On-Shelf Availability Meaning: The Shopper’s View vs. the System’s View

OSA shows what a shopper finds in the aisle, while inventory metrics show what the retailer’s system records for the store. A product can be in stock and unavailable at the same time, for example, when the cases are in the backroom, when staff put the product in another aisle after a reset, or when a promo display blocks the facing. In each of these cases the system shows stock, and the shopper cannot buy the product.

OSA vs. In-Stock vs. Service Level vs. Share of Shelf

Teams sometimes use these four metrics as if they meant the same thing. The table shows what each metric measures and why only OSA tells you whether the shopper can buy the product at the shelf.

Metric	What it measures	Who usually owns it	Data source	What it cannot show
On-shelf availability (OSA)	Share of listed SKUs present and correctly placed on the shelf	Brand trade marketing and field sales	Shelf audits, shelf photos, sensors	Why a SKU is missing
In-stock rate	Share of SKUs with positive on-hand inventory in the store system	Retailer replenishment	Retailer ERP and inventory records	Whether the product is on the shelf
Service level (fill rate)	Share of ordered units actually fulfilled. On-time, in-full delivery (OTIF) is a separate performance measure	Manufacturer or distributor supply chain	Order and shipment data	What happens after delivery to the store
Share of shelf	Brand’s share of facings in the category	Category management, trade marketing	Shelf audits, shelf photos	Whether each listed SKU is present

What Does “On-Shelf Customer Availability” Mean?

On-shelf customer availability is OSA measured at the hours when shoppers are in the store. A shelf that is full at 7 a.m. and empty after 5 p.m. shows good average OSA and poor customer availability. The term is close to customer service level and to “availability at the moment of truth”, the seconds when the shopper stands at the shelf and chooses. To estimate it, weight each check by store traffic at that hour.

The OSA Metric: Who Owns It and Why Manufacturers Can’t Delegate It

The retailer manages the shelf, so many manufacturers rely on the store’s own availability figures. When an SKU is missing and the shopper buys another brand in the same store, the retailer keeps the sale and the manufacturer loses it. If the shopper switches retailers or abandons the purchase, both can lose revenue. In a 2026 survey of 1,000 US adults, DOSS found that 62% had switched brands at least once because of a stockout. The survey covered stockout experiences across shopping contexts, so this figure should not be treated as a direct measure of in-store shelf behavior. Trade marketing and field sales therefore need their own OSA data from the shelf.

The On-Shelf Availability Formula (With Worked Examples)

The on shelf availability formula divides the number of listed SKUs found on the shelf by the number of SKUs that should be there. This section explains each element, shows the calculation for one store and explains how to combine results for a chain.

The Core On-Shelf Availability Formula

OSA (%) = (Number of SKUs present and correctly placed on the shelf ÷ Number of SKUs that should be on the shelf) × 100

Each element needs a written rule, because without one two auditors can count the same store differently.

Present. At least one sellable unit is in the SKU’s own facing. A facing filled with another product counts as absent.
Should be on the shelf. This list comes from the current planogram, the MSL or the retailer’s listing. Use one source for all stores in a report.
Level of record. Record each result as store × SKU × visit and combine the results later, so you can see which store and SKU caused a low figure.

Example: One Store, 12 SKUs

A beverage brand has 12 SKUs listed in a mid-size grocery store, and a field rep records the status of each one. The figures are illustrative.

Status of the SKU	SKUs	Available in OSA	In stock in the system
On the shelf, in the right place	9	Yes	Yes
In the backroom	1	No	Yes
On a neighboring shelf	1	No	Yes
Not delivered, zero on hand	1	No	No
Total	12	9	11

OSA = 9 ÷ 12 × 100 = 75%

System in-stock rate = 11 ÷ 12 × 100 ≈ 92%

The figures differ by about 17 percentage points. The system counts the backroom case and the misplaced SKU as in stock, and the shopper can buy neither, so two of the three missing SKUs need a store task and only one needs a delivery.

Aggregating OSA: Store, Chain, Category and Weighted OSA

A simple average gives every store the same weight, and a weighted average weights each store by its sales.

Weighted OSA (%) = Σ (OSA of store i × Sales of store i) ÷ Σ (Sales of store i)

In an illustrative chain, 10 large stores generate 50% of a brand’s sales and 500 small stores generate the rest. OSA falls to 80% in the large stores and stays at 96% in the small ones. The simple average is about 95.7%, while the sales-weighted OSA is 88%, and 10 of the 12 missing points come from the large stores.

Home-Shelf OSA vs. Total-Store OSA

Home-shelf OSA counts only the SKU’s main location in its aisle. Total-store OSA also counts endcaps, floor pallets and checkout displays. Use home-shelf OSA for everyday replenishment and planogram compliance, and total-store OSA for promotions, when the brand pays for secondary placements. Report the two figures separately. If a team combines them, a full endcap and an empty home shelf give an average that looks acceptable.

Three Metrics to Calculate Alongside OSA

Out-of-stock rate (%) = 100% − OSA

Phantom inventory rate (%) = SKUs in stock in the system but absent on the shelf ÷ SKUs in stock in the system × 100

Estimated lost sales ($) = Missing SKU-days × Average daily unit sales × Price

In the 12-SKU example, the out-of-stock rate is 25% and the phantom inventory rate is 2 ÷ 11, or about 18%, including backroom and misplaced stock. If the three missing SKUs stay off the shelf for three days, sell four units a day and cost $2.50, the store loses about $90. The guide on how to calculate the cost of out-of-stocks shows how to convert this figure into lost revenue and profit.

The Most Frequent Mistake: Using Inventory Data as the Numerator

Many teams estimate shelf availability using ERP or POS data because it is already in their systems, but these estimates can differ from what a shelf audit finds. Inventory-based measures often overstate shelf availability because they count stock that shoppers cannot find. Backroom stock, misplaced units and phantom inventory, that is, units the system shows on hand while the store has none, all count as available. IHL Group’s 2026 Inventory Distortion Study counts an out-of-stock every time a shopper leaves without the item, including cases when the system confirms inventory that nobody can find.

Key rule. When the numerator comes from inventory data, the result shows how much stock the store has on record. OSA counts only the products that a person or a device has seen on the shelf.

On-Shelf Availability Benchmarks: What Level to Aim For

No public study from the last five years gives an audited OSA benchmark by channel. Many articles still quote store audits from the early 2000s, so this section uses recent loss estimates and shows how to set your own target.

The Industry Target and Why Few Teams Reach It

In our experience with FMCG field teams, brands usually set OSA targets in the high nineties for priority SKUs and lower targets for the long tail. Few companies publish their actual OSA, but recent loss estimates show the size of the problem. In its 2022 OSA Barometer release, NielsenIQ calculated that out-of-stock and out-of-shelf items cost US CPG retailers 7.4% of sales in 2021, or $82 billion. IHL Group’s 2026 study puts the global cost of out-of-stocks and overstocks at $1.7 trillion, or 6.2% of retail sales, down from 10.4% in 2021.

OSA Benchmarks by Channel and Category

Each channel has its own delivery frequency, backroom size and staff time per aisle, so OSA risk differs. The table gives qualitative guidelines, and each brand should check them against its own shelf data.

Channel / format	Typical OSA	Target OSA	Main risk factor
Grocery	90–93%	97–98%	Fresh and perishable items, plus demand spikes from promotions
Mass merchandise	89–92%	96–97%	Large assortment and stock that sits in the backroom instead of on the shelf (phantom inventory)
Convenience	87–91%	95–96%	Little backroom space, small shelf capacity and infrequent DSD deliveries
Drug / pharmacy	90–93%	96–97%	Many slow-moving SKUs and locked-up items with high theft risk
E-commerce pickup (click & collect)	85–92% (first-pick fill rate)	97%+	Inaccurate inventory records and pickers competing with in-store shoppers for the same stock

Ranges are indicative and vary by market, retailer and category. Measurement method strongly affects the numbers.

Promotions, Launches and Resets: When OSA Drops Below Target

OSA usually falls lowest during promotions, resets and launches, and these are the periods when the brand spends the most money. During a promotion, demand grows faster than regular order rules expect, and staff move stock from the home shelf to the display. After a reset, some SKUs stay in the old location. In our experience, the first two to four weeks after a reset or launch need the most frequent shelf checks.

What One Point of On-Shelf Availability Is Worth

Value of 1 OSA point ($ per year) = Annual sales of the SKU group in the chain × 1% × Share of shoppers who do not buy your brand when the SKU is missing

The last factor covers brand switching and lost purchases. Take it from loyalty card data, or run the calculation with two or three values. In an illustrative case, a brand sells $5 million a year in one chain, and with a factor of 40% one OSA point is worth $20,000 a year. This is a simplified planning estimate, not a guaranteed revenue gain. It assumes that the affected SKUs account for a proportional share of sales and that the improvement reaches the stores and hours when demand occurs. Use SKU-level sales and stockout history when available to build a more accurate estimate.

How to Set Your Own OSA Benchmark

With one near-100% target for the whole assortment, field teams spend time on SKUs that sell a few units a month. Set targets by segment with this checklist:

Split SKUs by velocity into hero, core and long tail.
Split stores by format and by the brand’s sales in them.
Set separate targets for everyday, promo and post-reset weeks.
Measure a baseline for four to six weeks, then fix the targets and review them quarterly.

What Causes Poor On-Shelf Availability? The Real Root Causes

Empty shelves have causes in three layers: the supply chain, in-store execution and data. Each layer has its own owner and needs its own actions, and teams that mix them often fix the wrong one. These layers are diagnostic categories, not mutually exclusive causes. For example, an inventory record error may prevent replenishment and later appear as an empty facing. Start with the shelf observation, check the physical and recorded stock, and then assign the primary cause to the responsible team.

The IHL Group 2025 inventory distortion report attributes $742.5 billion of global inventory distortion, which covers out-of-stocks and overstocks, to store operations, $626.5 billion to the supply chain and $359.5 billion to manufacturer issues.The IHL figures cover total inventory distortion, including both out-of-stocks and overstocks. They do not directly measure the distribution of shelf-level OSA failures across these three layers.

Layer 1: Upstream Forecasting, Allocation and Replenishment

This layer covers everything before the product reaches the store, so the backroom and the shelf are empty at the same time. For example, the forecast does not include a promo uplift, or the shelf empties on Saturday while the next delivery comes on Tuesday. Min/max levels in automatic ordering may not match current sales, and a distributor may ship fewer cases than the store ordered. Supply chain planners and distributors are responsible for this layer, and service level and in-stock rate show most of its problems.

Layer 2: In-Store Execution, the Least Measured Layer

Retailer systems record almost nothing about this layer. In our experience with FMCG field teams, it accounts for a large share of the empty facings that shelf checks find.

Backroom stock. Cases stay in the backroom because staff did not know the shelf was empty.
Planogram non-compliance. The product is in the wrong bay or on the wrong level. The article on why planograms and actual execution don’t match explains why this happens.
Resets and blocked facings. Staff leave new facings empty after a reset, or a display, pallet or competitor’s product covers the facing.
Missing price tag. A shopper who cannot see the price may skip the product. Track this as a separate shelf execution issue unless the missing price information makes the product unavailable for purchase under the audit’s rules.
Damage, expiry and shrink. The store cannot sell damaged or expired units, and theft removes units while the record still counts them.

Layer 3: Data, Inventory Record Inaccuracy and Phantom Inventory

Inventory record inaccuracy means that the on-hand quantity in the system differs from the actual quantity in the store. It comes from receiving errors, unrecorded write-offs, wrong pack conversions and sweethearting, when a cashier passes an item without scanning it. When the record shows more stock than the store has, the automatic order does not trigger, sales fall to zero, and the forecasting model treats the zero sales as low demand and cuts the next order. The SKU stays off the shelf until a field rep or store employee checks it and corrects the record.

Root-Cause Diagnostic: Problem → Likely Cause → Owner → Fix

Use this matrix to find the likely cause of an OSA problem and the team responsible for it.

Problem	Likely cause	Owner	Fix
Empty facing, system on-hand above zero	Phantom inventory or backroom stock	Field team, store operations	Photo record, backroom check, record correction
Empty facing, system on-hand at zero	Late delivery, wrong order parameters	Supply chain, distributor	Review min/max and safety stock
OSA falls every Monday	Weekend peak without a weekend refill	Supply chain, store operations	Move the delivery day, add a Sunday refill
OSA is high, sales do not grow	Product off its planogram position	Trade marketing	Planogram compliance check
OSA drops only in promo weeks	Uplift missing from the forecast	Trade marketing, demand planning	Separate promo forecast, display refills
New SKUs missing after a reset	Facings not set up	Field team	Reset checklist, extra visits
Same SKU missing in many stores	Stockout at the distributor or DC	Supply chain	Allocation review, distributor escalation

Why Shipping More Product Doesn’t Fix an Execution Problem

If the cause is in layer 2, extra shipments only add cases to the backroom, write-offs grow and the next order shrinks. For this reason teams should find the layer before they act. A supply chain problem needs a new order or delivery plan, and an execution problem needs a store task and a follow-up shelf check.

How to Measure On-Shelf Availability in Retail

Teams measure on-shelf availability in retail with five groups of methods.

Manual Shelf Audits

A field rep walks the aisle with a checklist, so the numerator comes from the shelf. A rep covers only some stores, accuracy drops toward the end of a long route, and results arrive in MS Excel days later.

POS and Inventory Data Algorithms

Gap detection algorithms flag SKUs whose sales fall below the expected level, and NielsenIQ’s OSA Barometer applies this approach to daily store-level POS data. The method covers every store with data, shows the result without the cause and misses slow sellers and planogram violations.

RFID, Shelf Sensors and Electronic Shelf Labels

Tags and shelf sensors report status continuously, and IHL Group’s 2025 report names RFID as the technology with the largest single effect on inventory distortion. The retailer owns the equipment, so the brand gets data only where the retailer installs and shares it.

AI Image Recognition at the Shelf

A rep or a fixed camera takes a shelf photo, and software identifies each SKU, counts facings, finds empty facings, reads price tags and compares the shelf with the planogram. The brand does not need access to retailer systems for this. Blurred photos reduce accuracy, new packaging needs training images, and a photo shows only the front row.

Retailer Scorecards and Syndicated Data

Retailers share availability through supplier portals, such as Walmart’s reporting tools. The data helps in joint business planning, arrives days or weeks later and often counts system in-stock as availability.

Method Comparison: Latency, Coverage, Cost and What Each Method Misses

Method	What it measures	Data latency	Relative cost	Coverage	What it misses
Manual audits	Direct SKU presence and placement at the time of the visit	Days	Medium to high	Sample of stores	Days between visits, auditor errors
POS and inventory algorithms	Estimated availability or stockout signals from sales and stock records	One to several days	Low to medium	Stores with data access	Causes, slow sellers, planogram violations
RFID and shelf sensors	Tagged inventory movement or shelf-level signals, depending on deployment	Minutes	High	Equipped stores	Stores without equipment
AI image recognition	Visible SKUs, facings, placement and shelf gaps	Minutes after the photo	Low to medium	Every visited store or camera	Units behind the front row
Retailer scorecards	Retailer-defined availability and inventory KPIs	Days to weeks	Low	Selected chains	Shelf position, method differences

On-Shelf Availability Solutions: A Six-Step Program to Raise OSA

On shelf availability solutions give results when the team applies them in order, because each step uses the data from the previous one.

1. Define the Standard Before You Measure

An outdated planogram or MSL makes the denominator wrong. Update planograms per store format, agree the MSL with each retailer and name an owner who updates it after every reset.

2. Take the Numerator From the Shelf

Apply the key rule from the formula section in every store. Write down how reps should count present, absent and misplaced SKUs, train every rep on these rules, and use inventory data only to find phantom inventory.

3. Separate Supply-Chain OSA From Execution OSA

Mark each missing SKU as a supply chain case or an execution case, and send a separate weekly report to the team responsible for each layer.

4. Prioritize With Pareto: Top Stores × Hero SKUs

Following the Pareto principle, start with the roughly 20% of stores and SKUs that produce most of the loss, and check this split on your data. Rank store × SKU pairs and recalculate monthly.

Priority score = Probability of OOS × Unit sales velocity × Unit margin

5. Fix the Shelf During the Visit

A dashboard that shows an empty facing a week later reports a problem the rep could have fixed in the store. Send each check result to the rep within minutes with a task, such as moving the case from the backroom, and ask for a second photo to confirm the fix. Track the share of issues resolved during the visit and the average time from detection to correction. These operational KPIs show whether better measurement leads to faster shelf recovery.

6. Escalate to the Retailer With Evidence

Only the retailer can fix wrong order parameters, phantom records and facings that competitors have taken. Bring the category manager photos with a timestamp and geotag, the OSA history and an estimate of lost sales, and agree on actions, owners and dates.

Your First 90 Days: An OSA Improvement Roadmap

Period	What you do	Who owns it	What counts as success
Days 1 to 30	Update planograms and MSL, write counting rules, measure a baseline in priority stores	Trade marketing, retail execution	Baseline OSA established; agreed counting rules and verified assortment lists
Days 31 to 60	Split losses by layer, rank store × SKU pairs, start in-visit tasks	Field sales, supply chain	Root causes classified; in-visit resolution rate and time-to-fix tracked
Days 61 to 90	Escalate recurring causes, adjust order parameters, set segmented targets	Key account managers, supply chain	Hero SKU OSA above baseline; repeat gaps reduced; retailer action plans in place

How Goods Checker Measures and Improves On-Shelf Availability

Goods Checker is a computer vision platform for FMCG retail execution. Its modules cover the steps of the program above.

Plano Creator: Planograms and MSL for the Denominator

Plano Creator builds planograms and MSL for each store format, and reps see them in the mobile app.

Shelf Eye: SKU Recognition From Shelf Photos

Shelf Eye recognizes SKUs, facings, empty facings, price tags and POSM in a photo and compares the shelf with the planogram.

Check & Go: Tasks During the Visit

Check & Go sends the rep a task while the rep is in the store. Geolocation confirms the store, and the app blocks uploads from the phone gallery.

Analytics: OSA as a Managed KPI

Dashboards show OSA by store, SKU, category and merchandiser, with the reasons reps record for each missing SKU. The API sends the data to BI and CRM systems.

What Changes for Field Teams

In the Lex Marketing case study, a merchandising agency that uses Goods Checker in more than 4,500 outlets cut audit time per store by about 60% and reporting time by about 70%, with recognition accuracy above 95%. The results show how faster shelf audits can give field teams more time to address execution problems during store visits.

On-Shelf Availability FAQ

What is on-shelf availability in simple terms?

It is the share of listed products a shopper can find and buy on the shelf right now. If 18 of your 20 listed SKUs are in their places, OSA is 90%.

What is the difference between OSA and in-stock rate?

In-stock rate shows whether the store’s system records inventory. OSA shows whether the SKU is on the shelf, so a product in the backroom counts as in stock in the system and as missing in OSA.

What is the formula for on-shelf availability?

OSA (%) = SKUs present and correctly placed on the shelf ÷ SKUs that should be on the shelf × 100. Take the denominator from the planogram or MSL and the numerator from a shelf check.

What is a good OSA benchmark?

There is no single OSA benchmark that applies to every retail channel and SKU. Some FMCG teams use targets of 98% or higher for priority products, but these are operational goals rather than a verified industry average. Measure your own baseline first, then set separate targets for hero SKUs, core SKUs and the long tail.

What is the average out-of-stock rate in retail?

A widely cited global retail study by Gruen, Corsten and Bharadwaj reported an average out-of-stock rate of 8.3% in 2002. This is a historical benchmark, not a verified current average. NielsenIQ estimated that out-of-stock and out-of-shelf items accounted for 7.4% of unrealized US CPG sales in 2021, but that figure measures lost sales, not the out-of-stock rate.

What is phantom inventory and how does it distort OSA?

It is stock that the system records while the store has none. OSA calculated from inventory data counts it as available, and automatic ordering stops because the system sees stock.

How often should on-shelf availability be measured?

Check hero SKUs in priority stores at every visit and the long tail less often. Check more often during promotions and after resets or launches.

Does on-shelf availability include secondary displays and endcaps?

Home-shelf OSA covers only the main location, and total-store OSA adds endcaps and displays. Report them separately.

Can a manufacturer measure OSA without access to retailer data?

Yes. Shelf audits and image recognition take data from the shelf, so they need no retailer systems. Retailer data still helps to find phantom inventory.

The Bottom Line: Measure OSA on the Shelf

Before your next OSA report, check where the numerator comes from and which of the three layers causes each missing SKU. To measure OSA from the shelf in your stores, book a call with a Goods Checker team.