UPSPW Means Scans, Not Shipments: The Velocity Number, Defined

One Cinderhaven Provisions SKU sells 4.0 units per store per week, 3.0, or 2.5, depending on who is counting. All three are arithmetic on the same product over the same six weeks. Only one of them is the number a buyer will read at the line review, and it is the smallest.

Cinderhaven is a fictional company and its scan data is a synthetic dataset; the three-way disagreement is what happens on any brand's velocity report the first time someone checks the definitions. Units per store per week, UPSPW, is three decisions disguised as one metric: which units, which stores, which weeks. Brands and buyers make those decisions differently, and neither side usually knows it.

The definition has three parts, and each one is a choice

SPINS, whose reports are the currency of the natural channel, defines velocity as sales, in dollars or units, divided by distribution, represented as stores, ACV, or total distribution points, and gives the store version its name: average weekly units per store selling. A broker's plain-English version does the same math in one sentence: take total units sold, divide by the number of stores where at least one unit scanned, and divide again by the number of weeks. Their example: 3,600 units across 60 stores over 12 weeks is 5.

Each of the three inputs is a choice. Units means scanned at a register or shipped from a warehouse. Stores means stores where at least one unit scanned, "not stores where your product is authorized or listed", or the authorized count on the planogram. Weeks means the whole period, or only the weeks the product was on shelf. A number is UPSPW under any combination. The combinations do not agree.

Same formula, three inputs, and every input has two answers.

One SKU, three correct numbers

Take Cinderhaven's best-selling salsa at a regional chain. It is authorized in 240 stores. Over the six weeks after a shelf reset, the brand shipped 960 units a week to the distributor against that chain's orders, part of a promotional load-in, while the registers scanned 600 units a week from 200 stores, because 40 stores had gone dark in the reset and never restarted replenishment.

| Who is counting | Units | Stores | UPSPW | Where the number lives | |---|---|---|---|---| | The brand, from its ERP | 960 shipped to the distributor | 240 authorized | 4.0 | Sales report, board deck | | The buyer, from POS | 600 scanned | 200 selling | 3.0 | Line review, SPINS or Retail Link | | The category manager, per authorized door | 600 scanned | 240 authorized | 2.5 | The planogram decision |

Every row is a valid calculation. The first is sell-in: the wholesale transaction, the brand ships N units to the retailer and invoices them. The second is sell-through, the sale from the retailer to the end consumer, which is what the buyer's system holds. The third is the buyer's number applied to the buyer's denominator: the shelf space the chain committed, including the 40 doors returning nothing.

The gap between rows one and two is 360 units a week that shipped and did not scan: over the six weeks, 2,160 units, about $17,000 at Cinderhaven's $8 wholesale price. Those units are real. They are sitting in the distributor's warehouse and the retailer's back rooms, and the brand's report counted them as velocity. The gap between rows two and three is the 40 dark stores, which lower the per-door average by half a unit and are invisible in a report that only counts stores selling.

The brand's 4.0 is true. It is also the only one of the three that nobody at the retailer will ever see.

The buyer's number is the one that governs

The reason the definitions matter is that only one of them decides anything. Velocity is a key metric retailers use to judge the performance of a product, and if a brand does not consistently meet the velocity threshold, the product is at risk of discontinuation. The mechanics are not subtle: buyers rank the set by velocity and trim from the bottom, and the velocity is derived from point-of-sale data, so the rate of sale ties straight back to what scanned. Scans, stores selling, the period the buyer chose.

So a brand walking into a review with 4.0 from its own ERP is arguing a number the buyer's screen does not contain. Buyers measure scans, and the delisting conversation runs on their number: the 3.0 or the 2.5. The brand that shows up with 4.0 has brought a different metric with the same name.

The thresholds themselves are category-specific and mostly unpublished. The published rules of thumb are rough: a 2015 founders' guide to Whole Foods put dry grocery in the center store typically between 1 and 3, and a 2025 vendor guide puts refrigerated beverages at 5 to 7 and supplements as low as 1. What every source agrees on is that the threshold is measured in scans. You might be shipping cases to your distributor, but that does not mean stores are selling them, and the buyer's system never counted the cases.

Compute the buyer's number before the buyer does

The brand's number is a different instrument, and a useful one. Sell-in tells you what the channel absorbed and, against sell-through, how much inventory is building in it. Cinderhaven's 360-unit weekly gap is a pipeline-fill signal that will reverse into short orders when the load-in ends, and it is the earliest warning the brand will get. But it is a supply number, and calling it velocity is how a stockout teaches the forecast to order less in one quarter and a load-in teaches it to order more in the next.

The fix is to hold all three on one page, labeled. Scans over stores selling is the buyer's number; compute it from the 852 feed or the retailer portal, weekly, before the review calendar does. Scans over authorized doors is the planogram number, and the gap between it and the buyer's number is your dark-store count, priced. Shipments over authorized doors is the supply number, and the gap between it and scans is channel inventory. Three rows, one SKU, no argument about definitions in the meeting, because the definitions are printed on the report.

The velocity a brand measures for itself should be the one the buyer will measure for it.

Three velocities for your top SKU

Send me one quarter of shipments and one quarter of scans for your best-selling SKU at one chain, plus the authorized store count. I will write back with the three numbers side by side, the channel inventory the gap implies, and the dark doors the third row exposes. Start with your best seller. If the three agree, you are the first brand I have seen where they do.