Units Per Store Per Week: The Velocity Number Retailers Use to Delist Your SKU
A brand's internal reporting shows 3.2 units per store per week. The buyer's line review shows 1.8. Both numbers are correct.
The brand was measuring shipments. The buyer was measuring scans.
Units per store per week, UPSPW, is the single metric buyers use to determine whether a SKU earns its shelf space. It is also the metric most brands measure incorrectly. The gap between what a brand reports to itself and what a buyer sees in POS data is rarely a data entry error. It is a structural disagreement between two different definitions of "store" and two different definitions of "velocity."
Understanding which number governs the decision, and why the two diverge, is the work that must happen before a delist conversation, not during one.
What units per store per week means and how it is calculated
Units per store per week is the average number of units of a single SKU sold per retail location per week. The formula is straightforward: total units scanned at the register during a given period, divided by the number of stores carrying the product, divided by the number of weeks in that period.
UPSPW = Total units sold / Number of stores / Number of weeks
The number comes from scanner data, the UPC reads captured at checkout, not from shipment records, warehouse pulls, or distributor invoices. SPINS, IRI, and NielsenIQ all calculate UPSPW from point-of-sale scans collected at the store level. Retail Link provides it directly from Walmart register data. The metric answers a specific question: how fast does this product leave the shelf and get replaced by consumer purchases?
What constitutes a "good" UPSPW depends entirely on category, channel, and store format. Retailers and syndicated data providers do not publish official cut lines, so the working ranges below are drawn from broker guidance and category-review practice in natural and specialty grocery. Treat them as orientation, not gospel; the number that governs is the one your buyer's category report shows:
- Shelf-stable snacks and bars: 2-5 UPSPW. High-velocity items in this space clear 7+.
- Sauces, condiments, and shelf-stable spreads: 1-3 UPSPW. These are slower-turning categories where a 2.0 can be perfectly adequate.
- Refrigerated beverages: 4-8 UPSPW. The cold case demands fast turns because holding costs and shrink are higher.
- Frozen meals and entrees: 1.5-3 UPSPW. Lower traffic but also lower spoilage risk than fresh refrigerated.
- Fresh dips, dressings, and dairy alternatives: 3-6 UPSPW. Short shelf life raises the floor; anything below 2.5 often triggers a review.
These ranges shift by retailer. A SKU performing at 2.0 UPSPW in a Whole Foods region may be adequate; the same 2.0 at a high-volume conventional chain may be below the category cut line.
The number the brand sees and the number the buyer sees often diverge for structural reasons beyond simple measurement error. The brand typically pulls velocity from its ERP or distributor reports, which measure units shipped, not units scanned. The buyer pulls from POS data, which measures units sold at stores where product actually moved. If 60 of 200 authorized stores have not received inventory in the current period, the brand divides by 200 and the buyer divides by 140. The arithmetic is correct on both sides. The denominator is different, and so is the conclusion about whether the SKU is performing.
That denominator gap is why a brand can walk into a line review confident in its velocity and walk out having lost shelf space. Understanding that UPSPW is a scan-based metric, and that the brand's internal version may be measuring something else entirely, is the first step in closing the gap. The second step is the portfolio-level audit that identifies which SKUs are actually earning their shelf cost and which are averaging their way past a problem.
Two velocity numbers, only one of which governs
A brand measuring velocity from its own shipment data counts units shipped to a warehouse and divides by authorized distribution, the number of stores the brand has agreements to stock. That number can look healthy even when sell-through is weak, because authorized stores and scanning stores are not the same list.
When product sits in distribution centers, or when authorized locations have not received inventory in weeks, the brand's denominator is larger than the buyer's. Retail Link reports scan data at stores where product actually moved. SPINS measures sell-through at retail, not shipment out of a warehouse. What a brand pulls from its ERP is shipment-based. These are genuinely different measurements of genuinely different things.
A brand with 200 authorized locations and 140 actively scanning stores divides by 200 on its own scorecard. The buyer divides by 140. A 3.2 and a 1.8 can both be arithmetically sound. The difference is not a rounding error, it is a 30% gap in distribution coverage that the brand has been averaging away. Treating the two figures as equivalent is how a brand enters a line review believing its velocity is adequate and exits having lost facings.
Category average, not an absolute floor
The UPSPW threshold a buyer applies is not fixed. It is a category-relative benchmark, and misreading it is nearly as common as misreading the velocity figure itself.
For dry grocery center-store at natural and specialty retailers, a threshold of 1-3 UPSPW is commonly cited. Refrigerated categories (beverages, dairy alternatives, fresh dips and dressings) carry substantially higher expected velocity because the shelf space costs more to hold and product turnover directly affects shrink. The relevant benchmark for any SKU is the category average in the specific buyer's stores, not an industry-wide floor.
SPINS, the primary data source for natural and specialty channel buyers, calculates velocity several ways. The version buyers rely on most is ACV-weighted velocity, scan velocity weighted by store size and annual sales volume, rather than a simple average across all stores in the universe. A SKU performing at 70% of the ACV-weighted category velocity is in rationalization range regardless of how it compares to a flat published threshold.
This distinction matters for brands reading industry benchmarks. A figure of "2.5 UPSPW for better-for-you snacks" is an average across a broad store universe. A buyer is measuring how a specific SKU compares to every other item competing for the same shelf position in their stores. The two numbers are not interchangeable, and a brand that tracks only one may be dangerously wrong about the other.
How delist actually happens
Delist rarely arrives as a single event. It arrives through a sequence most brands interpret too late.
Major grocery chains conduct category reviews multiple times per year, quarterly for high-velocity categories, twice annually for slower-moving ones. A SKU underperforming against category velocity enters the review as a rationalization candidate. The first visible signal is usually a facing reduction: three facings become two, two become one.
A facing reduction is both a penalty and a test. A SKU that holds scan velocity with fewer facings has demonstrated that demand exists and the product was over-slotted. A SKU whose velocity falls further confirms the buyer's original thesis. A brand that notices a facing reduction has roughly one review cycle, often 60 to 90 days, to move the scan velocity number before the next rationalization decision.
The lever in that window is promotional activity that drives consumer pull-through at store level, not shipments into the distribution center. Warehouse inventory does not register in the scan data the buyer is monitoring. A brand that responds to a facing reduction by pushing more product into the supply chain is solving precisely the wrong problem.
The data that closes the gap
Brands that catch velocity problems before reviews do so through regular scan-versus-shipment reconciliation, comparing ERP shipment data against SPINS or Retail Link scan data, store by store.
The reconciliation surfaces phantom inventory: product that has shipped but is not scanning, sitting in back rooms or held at a distribution center, inflating authorized-store distribution without contributing to velocity. The same compliance failures that generate OTIF penalties also create phantom inventory, late ASNs, delivery variances, and receiving documentation mismatches that cause product to sit rather than flow to the shelf. Fixing the shipment compliance problems improves scan velocity in tandem, because product reaches the shelf where it can be sold.
The reconciliation also identifies distribution gaps: authorized stores that have not received product in the current review window and therefore contribute no scans. Those stores count against the brand in any velocity calculation that uses authorized distribution as the denominator. Closing distribution gaps is typically more cost-effective than equivalent promotional spend, and unlike promotions, the improvement is structural rather than temporary.
The deduction patterns that surface compliance failures live in the same data as the velocity discrepancies. A shortage deduction from a retailer and a store that is not scanning are often symptoms of the same upstream data problem. Brands that build reconciliation infrastructure identify both at once. The velocity data also ties directly to the revenue lifecycle, when scan velocity drops, promotional spend increases to compensate, and the gap between invoiced revenue and cash received widens.
Find where your velocity data disagrees
Lailara runs a scan-versus-shipment velocity reconciliation, comparing ERP and distributor shipment data against SPINS scan data, to identify distribution gaps, phantom inventory, and the stores where authorized presence is not translating into measured velocity. The deliverable is a store-level map of where the brand's internal velocity number and the buyer's velocity number diverge, and what is causing each gap. If you are preparing for a category review or have lost facings in the last six months, book a 30-minute scoping call.
See the methodology behind this post. The worked example (50 SKUs scored across five dimensions, 19 kill candidates, 22 fix-or-kill, a methodology to re-run quarterly) is a live demo you can open and explore. SKU Portfolio Audit →
The Ten Decisions is the map behind this post. Every data problem a $25M specialty food brand runs into (chargebacks, deductions, launch economics, OTIF gaps) maps to one of ten decisions being made without adequate information. See the full picture →