Share of Shelf Is a Velocity Number, Not a Space Number
The buyer took back three of Cinderhaven Provisions' eight facings, and every number on the line-review slide was correct. Cinderhaven held 16.7% of the hot-sauce shelf at a 220-store grocery chain. It sold 11.2% of the category's jars. Losing the weakest SKU in that reset costs $94,354 over the following year: $25,626 in contribution that stops scanning, plus $68,728 to buy the slot back. Cinderhaven Provisions is a fictional company, and this set (48 facings, a category scanning 50 jars per store per week) is a synthetic example built to show a line-review mechanism that is entirely real.
The brand calculated its share of shelf. The buyer calculated what that shelf was earning. Only one of those numbers decides the next planogram.
The tape measure gives you half the formula
Share of shelf has two standard formulas. Linear share divides the shelf length a brand occupies by the category's total shelf length. Facing share counts product fronts instead. Both multiply by 100.
Linear share of shelf = brand shelf length ÷ category shelf length × 100
Facing share of shelf = brand facings ÷ category facings × 100
Cinderhaven runs four hot sauces in the set at two facings each. Eight of 48 facings is 16.7%. The jars in this synthetic set are close enough in width that linear share lands in the same place, which is the condition under which facing share is safe to use. Where pack sizes vary, a narrow jar wins facings it has not won in inches.
Most guides stop here. The number is accurate and incomplete, because it measures what the brand holds and says nothing about whether the brand has earned it. Category work has compared the two for decades: Desmet and Renaudin's 1998 study modeled share of space against share of sales across a retail chain's categories and found that sales respond to space more strongly in impulse-driven categories. The buyer's working version of that model needs one division.
Share of shelf is what the brand holds. Share of sales is what it earns.
The buyer divides by velocity
Share of sales comes from register scans, not shipments. The input is units per store per week: Cinderhaven's four SKUs scan 2.0, 1.6, 1.2 and 0.8 jars per store per week, 5.6 in total. The category scans 50. Divide 5.6 by 50 and Cinderhaven's share of sales is 11.2%.
The buyer reads that from the retailer's portal or a syndicated pull. Which data source answers which question matters less here than the fact that every one of them counts scans. Shipment data never enters the line review.
| SKU | Facings | Units per store per week | Share of shelf | Share of sales | Sales ÷ shelf |
|---|---|---|---|---|---|
| A | 2 | 2.0 | 4.2% | 4.0% | 0.96 |
| B | 2 | 1.6 | 4.2% | 3.2% | 0.77 |
| C | 2 | 1.2 | 4.2% | 2.4% | 0.58 |
| D | 2 | 0.8 | 4.2% | 1.6% | 0.38 |
| Line total | 8 | 5.6 | 16.7% | 11.2% | 0.67 |
A ratio of 1.0 means the space matches the sales. Cinderhaven's line runs at 0.67, and SKU D runs at 0.38. Space that matches sales would be 11.2% of 48 facings: 5.4, rounded to five. The buyer takes back three. SKU D leaves the set, and SKU C drops to one facing.
Seventeen percent of the shelf, bought with eleven percent of the scans.
A cut SKU costs more than its facings
The visible cost is the contribution SKU D stops earning. It scans 0.8 jars × 220 stores × 52 weeks = 9,152 jars a year. At Cinderhaven's $2.80 contribution per jar, that is $25,626.
The larger cost arrives if the brand wants the slot back. A returning item is a new item to the chain, and new items pay to get on the shelf. The slotting math for this same 220-store chain runs $250 per item per store: $55,000. Free fill adds one case per store, 220 cases at $62.40 to produce: $13,728.
| Layer | Arithmetic | Cost |
|---|---|---|
| Lost contribution, year one | 9,152 jars × $2.80 | $25,626 |
| Slotting to return | $250 × 220 stores | $55,000 |
| Free fill | 220 cases × $62.40 | $13,728 |
| Total | $94,354 |
Re-entry alone costs $68,728, which is 2.7 years of what SKU D contributes at its current pace.
Winning the slot back costs more than the slot earned.
Check the denominator before you concede
Every velocity in that table divides by 220 stores. That assumes every store set every SKU. Suppose 33 stores never cut the line in after the last reset, the kind of void that never shows up as an out-of-stock because the item was never there to run out. The jars Cinderhaven sold came from 187 stores, not 220.
Rerun SKU D on the stores that carry it: 0.8 × 220 ÷ 187 = 0.94 jars per store per week. Rerun the whole line the same way and 5.6 becomes 6.59. The category grows with it, from 50 to 50.99, so Cinderhaven's share of sales is 6.59 ÷ 50.99 = 12.9%. Fair share moves from 5.4 facings to 6.2. On the data alone, the buyer's case for taking three facings becomes a case for taking two.
SKU D still runs below its space at 0.44. The void does not rescue the weakest item. It changes which facings are in dispute and what the brand should ask for first: set the 33 stores before conceding a facing that the scans never had a chance to earn. Voidfinder maps those stores from the same scan feed the buyer uses.
A void is not a demand verdict.
Send the line-review slide
Send the share-of-shelf page from your last line review and the scan data behind it. I will rebuild the ratio SKU by SKU and tell you which facings your numbers can defend, and which losses are voids rather than demand. Send it here. I'll write back. No call.
Next step — Authorized, on the shelf, and not moving
Get the number for your own data. Seven business days.
The offers below run this on your own data — the Snapshot credits in full toward the audit.
Private, expiring upload — never email. Mutual NDA before anything moves. Files destroyed within 30 days of delivery, with a certificate. Methods published, tools open source.