Planograms and shelf space to sales

Why the shelf is where the plan gets real

The Sprouts refrigerated salsa set is 48 facings wide and not one facing wider, and that fixed 48 is why space to sales is the argument every category review eventually lands on: which SKUs get which of those slots relative to the $18.4M they actually sell. The number does not grow. The buyer will not add a foot of steel to the salsa fixture because Verde Fresca had a good quarter, so every facing you win comes off some other SKU. You can win the assessment, agree the scorecard, and pick a strategy, but none of it touches a shopper until it becomes a planogram: the shelf diagram that says this SKU sits here, at this many facings, at this height. Everything upstream is a memo. The planogram is the plan made physical.

Facings are the scarcest resource in the whole category, and that scarcity makes the shelf the one tactic where the math has to be airtight. Because a move you want is a move some other SKU loses, the buyer will make you show your work before she shifts a single slot, and a facing argument that cannot survive her pivot never gets made twice. This chapter is how I built the case for a facing move I could defend to the Sprouts category manager, and where the math breaks if you are not careful.

The space-to-sales principle

The governing idea is simple to state: a SKU's share of facings should roughly track its share of category sales. If fermented salsa drives 11% of the $18.4M set, it has some claim to about 11% of the shelf. This is the space to sales rule, and it is the starting point for almost every planogram optimization exercise a category captain runs.

Simple to state, and wrong if you stop there. Two SKUs at identical dollar share can have very different facing needs, and the reason is velocity and days of supply. A SKU that sells $42 per store per week through one facing empties that facing faster than a SKU selling $31 through the same slot. If you give both the same number of facings, the fast mover runs out mid-week and loses sales the report never records, because a shopper who finds a hole on the shelf does not file a complaint, she buys the next brand. So space-to-sales is the anchor, and then you adjust it up for velocity and down for the SKUs that are only holding shelf out of habit.

Here is the full input set I use to allocate facings, not just the two-column version everyone starts with:

The days-of-supply column is the one analysts skip and buyers never do. A facing holds a fixed number of units. If a SKU draws 30 units a week through a facing that holds 12, and the store restocks twice a week, that facing supplies about 24 units of weekly capacity against a 30-unit draw, so it is out of stock before each delivery no matter how many dollars its share entitles it to. The fix is a second facing, not a bigger share argument, and the days-of-supply math is what proves it.

Share of shelf versus share of sales

Share of shelf, on its own, is a vanity metric. Verde Fresca holds some percentage of the 48 facings, and that number by itself tells you nothing, because a lot of shelf holding a slow brand is a problem, not a trophy. The number only means something as a ratio against share of sales, and that ratio is the over/under-index. I go deeper on the metric itself in what is share of shelf; here is the operating version.

The formula is one line:

Over/under-index = share of sales ÷ share of shelf

An index above 1.0 means the SKU or segment converts more sales than its shelf real estate would predict, so it is under-shelved and has a claim to more space. An index below 1.0 means it holds more shelf than it earns, and it is the donor pool when you need facings for something better. This is the single most useful number in a planogram conversation, because it turns "give my brand more shelf" into "this segment is under-indexed at 1.75, so the shelf is leaving sales on the table," which is a sentence the buyer can act on.

One honest complication, the same one I flag on the SPINS metrics decision tree: the share-of-shelf half of that ratio does not come from SPINS. SPINS gives you dollars, units, velocity, distribution, share of sales. It does not know how many facings anything has. Facing counts come off a completely separate vendor surface, a shelf audit, an image-recognition scan, or the retailer's planogram software, and that data lives in its own system. So the over/under-index is never a single query. It is a reconciliation: share of sales from the syndicated data, share of shelf from the shelf-audit export, joined by hand. Anyone who tells you the index popped out of one dashboard is either using a stack you do not have or has not actually built it.

A worked example: the fermented salsa move at Sprouts

Here is the real case I would build for the Sprouts refrigerated salsa reset. The fermented and probiotic segment is 11% of category dollars and growing 28% year over year, the fastest thing in the set. Verde Fresca has two SKUs in that segment, running $42 per store per week against a category median of $31. The fresh segment is 62% of dollars but only growing 3%, and three of Verde Fresca's five SKUs live there, flat.

Now the shelf data, off the audit surface. Suppose the fermented segment holds only about 6% of the 48 facings, just 3 slots, while it drives 11% of dollars. Run the index:

Fermented over/under-index = 11% share of sales ÷ 6.3% share of shelf = 1.75

An index of 1.75 is loud. It says the fermented segment is converting sales at more than one and a half times the rate its shelf would predict, which is exactly the fingerprint of a growth segment that got frozen into a planogram built two resets ago, before it took off. Meanwhile the fresh segment, at 62% of dollars and holding something closer to 68% of facings, indexes at about 0.90: slightly over-shelved, which is where the donor facings come from.

So the ask writes itself. Move 2 facings from the slowest fresh SKUs to fermented. Here is the before-and-after I would put in front of the buyer:

The two facings move fermented from a 1.75 index down toward 1.06, which is close to balanced, and only nudges fresh from 0.90 up to 0.96, still slightly over-indexed so the buyer is not giving up a segment she needs to protect. Nothing else on the shelf moves. That is the whole pitch: two specific facings, off two specific slow fresh SKUs, onto the two fermented SKUs already outrunning the category median by 35%.

And I would close it with the velocity number, because the index alone is a shelf-efficiency argument and the buyer is also graded on total dollars. Verde Fresca's fermented SKUs run $42 against the $31 median. Two more facings on a $42 velocity item, at a fixed shelf size, buys more turns off the same four feet of steel. That is the version of the argument that survives the meeting.

Fair share by brand count is the anti-pattern

The lazy way to build a planogram is to count the brands and split the shelf evenly, or close to it. Five brands, roughly a fifth of the facings each, maybe a bump for the leader. It feels fair, and it is the worst thing you can do to a category, because it hands slow brands shelf they never earned and starves the fast ones.

Walk it through on the salsa set. If Casa Marisol (29% share), Tolteca (16%), private label (14%), Verde Fresca (7.7%), and a long tail split the shelf by anything close to brand count, the private-label line and the tail end up massively over-indexed, sitting on facings their velocity cannot justify, while Verde Fresca's fermented SKUs stay pinned at a 1.75 index because "everyone gets their fair share." Fair-share-by-brand-count ignores velocity entirely, which means it ignores the one thing the shelf exists to do, which is turn facings into sales. The buyer who runs her category this way is leaving dollars on every foot of the fixture, and a good captain's job is to catch it. Space-to-sales, adjusted for velocity and days of supply, is the discipline that replaces it.

The out-of-stock angle nobody sees in the report

Here is the failure mode that never shows up in a SPINS pull, and it is the strongest quiet argument for the facing move. A high-velocity SKU with too few facings does not lose sales in a way the data records. It goes out of stock mid-week, the shopper finds a gap where the item should be, and she buys the fermented SKU next to it or walks. The register rings the substitute or rings nothing. The report shows the under-faced SKU selling exactly to its facing capacity and calls it a normal week.

The salsa case makes it concrete. Take a fast fresh SKU that moves 30 units a week through a facing that holds 12, restocked twice: that is about 24 units of weekly capacity against a 30-unit draw, so it empties before Friday most weeks. Add a promotional lift or a good weather weekend and the gap widens. The lost sales are invisible: they do not appear as a decline, they appear as a ceiling. This is why days of supply belongs in the facing table and why velocity, not just share, drives the allocation. The over/under-index tells you the fermented segment is under-shelved. The out-of-stock math tells you the fastest-moving SKUs on the set are also losing sales you will never see in the syndicated numbers, and together they are a much harder case for the buyer to wave off.

The planogram does not stand alone, either. It inherits directly from the assortment plan: you cannot face a SKU you decided to cut, and every discontinued item is a facing freed up for something that earns it. The assortment decides what is on the shelf; the planogram decides how much of the shelf each survivor gets. Build them in that order.

Doing this in Scout

Scout's brand and category views sit on the SPINS extracts, so the sales half of the space-to-sales math is a saved view instead of a rebuilt pivot: share of category dollars by segment, velocity per store per week, and the year-over-year growth that tells you which segment is the fermented one, all in one read. When the fermented segment is up 28% at a $42 velocity, that is a two-minute pull, and it is the half of the argument that carries the dollar weight.

The honest limit, the same one I state on the velocity decision tree: the share-of-shelf half is not in Scout. Facing counts live on a separate vendor surface, a shelf audit, an image-recognition scan, or the retailer's planogram software, and that data is not integrated into Scout today. So the over/under-index is a manual reconciliation. You export share of shelf from the shelf-data system, pull share of sales from the Scout view, and join them by hand to get the 1.75. Scout assembles the sales evidence fast and reliably. It does not audit the shelf, draw the planogram, or move the facings, and it will not tell you the index without the shelf export sitting next to it. Bring both.

The short version

• Facings should track share of sales, adjusted up for velocity and days of supply and down for SKUs holding shelf out of habit, never split evenly by brand count.
• The over/under-index (share of sales ÷ share of shelf) is the number that defends a shelf move: Verde Fresca's fermented segment at a 1.75 index is under-shelved and has a real claim to two more facings.
• Share-of-shelf data comes off a separate vendor surface, not SPINS, so the index is always a manual reconciliation, and an under-faced fast mover also loses out-of-stock sales the report never shows.

Related: Assortment planning in category management · What is share of shelf?