Your biggest number is not your best case

The largest gap in dollars is the 13 oz multipack in metro stores — but demand there runs 14% below the chain average, so a buyer will push back and be right. The same SKU in suburban stores is worth less and argues far better: demand 31% above average, 37 stores not carrying it.

Modelled opportunity

$761K

annual, across all gaps

Against current sales

10.9%

of $6.98M at this retailer

Store-SKU gaps

469

of 1,060 possible combinations

Biggest single gap

$170K

Classic 13 oz multipack · Metro

Realization factor What share of chain-average velocity a store that does not stock the SKU today would actually achieve once it does. This is the assumption the whole screen rests on — move it and watch.

0.35

Distribution against demand

Select a cell for detail

SKU	Metro87 stores	Suburban66 stores	Rural59 stores

shading = modelled opportunity top line = stores carrying it below = household demand index, 100 = chain average

By SKU

Opportunity is modelled, not measured — see the limits below

SKU	Distribution	Stores missing	Demand index	Modelled opportunity

What this does not prove

Opportunity is a model, not an observation. It assumes a store without the SKU would sell it at the chain's velocity for that SKU, adjusted by the demand its own shoppers show, multiplied by a realization factor. The slider above that factor is not decoration: at 0.20 the total is under half what it is at 0.45, and nothing in the data tells you which is right. It is a judgement about shelf position, competing facings and how hard the category manager will work the listing — which is why it belongs in the conversation with the buyer and is recorded in the manifest as an input rather than buried as a constant.
It does not account for shelf space. Every listing displaces something, sometimes one of your own SKUs — the growth decomposition screen is where you would check whether that happened afterwards.
Demand index is built from the buying behaviour of each store's households in adjacent segments, not from stated preference. It is a decent signal for pack size and flavour, and a weak one for genuinely new propositions.
Store-level cells below 100 households are suppressed, so the smallest rural stores are absent from the demand estimate entirely and appear here only as distribution counts.

What to do with this

Take the suburban 13 oz case to the buyer, not the metro one. Metro is the bigger number and the weaker argument — demand there runs below the chain average and a category manager will say so. Suburban has demand 31% above average across 37 stores that do not carry it.
Bring the store list, not the total. Listing decisions are made per cluster and per store. The named stores in the right-hand panel are a concrete ask; a single modelled total is not.
Decide the realization factor with the buyer, not for them. Every figure here scales with it. Agreeing 0.35 together before the meeting is a stronger position than defending a number they did not choose.
Do not chase Thin 8 oz on this screen alone. Demand index 142 in metro is interesting; 1 rural store carrying it suggests a supply or shelf constraint the loyalty data cannot see. Ask before you model.

Proof of this run Template, datasets, policy, enclave quote

runcr-1a77d530 · 2026-10-09 16:12:48 UTC

templateassortment-gap v2.4 · sha256 e5b904…2f6c

retailer datameridian.loyalty.tx-52w · sha256 4a81d0…bb02 · 1,841,206 households

retailer datameridian.assortment.store-sku · sha256 77c1ab…04e8 · 212 stores

parametersrealization factor 0.35 · set by requester, recorded in output

policymin-cell 100 hh · signed by Meridian Markets, 2026-08-14

query budget15 of 40 used this quarter · requester: j.alvarez@northfield

enclaveIntel TDX · quote verified · measurement 1f8c44…90de

Demonstration screen built on synthetic data with the structure and distributions of a regional grocery loyalty file. Retailer, brands, stores and figures are invented. What is real is the method: demand estimated from household behaviour inside the enclave, compared with the retailer's own assortment file, with the modelling assumption stated on the face of the result.