Lift Mosaic
/Meridian Markets /Tortilla chips /Northfield
Synthetic demo data

52 weeks ending Oct 4, 2026

Northfield grew +6.3% at Meridian Markets

Brand sales
$6.98M
from $6.57M
Buying households
209,400
11.4% of loyalty base
Loyalty households
1,841,206
1,148,913 buy the category
Stores
212
6 states, all banners

Decomposition of the change

Select a bar for detail
Vertical axis starts at $6.40M, not zero — the bars show change, not scale. Cells below 100 households are suppressed.

Who the volume came from, and who took it

Household-level switching, same category
BrandHouseholdsNet $ moved$ per household

What to do with this

  1. Build the Salt & Stone win-back audience. They took $98K from you and the 11,100 households involved are identifiable as a group. That is the single clearest action on this screen, and it is one click away on the segment builder.
  2. Stop counting new category buyers as growth. Only 21% of the gain came from them. If your plan assumes category expansion, the plan is working from the wrong mechanism and the media mix should shift toward switching and frequency.
  3. Leave the 6,900 lapsed-category households alone. They stopped buying tortilla chips entirely. Brand media does not recover them and the budget is better spent on the 18,700 who switched rather than left.
  4. Re-run after the next campaign. The value of this screen is the delta between quarters, not the snapshot. One run tells you what happened; two tell you whether what you did worked.
Proof of this run Template, datasets, policy, enclave quote
runcr-7f3a91c4 · 2026-10-09 14:22:07 UTC
templategrowth-decomposition v2.1 · sha256 9c4b2f…e17a
retailer datameridian.loyalty.tx-52w · sha256 4a81d0…bb02 · 1,841,206 households
brand datanorthfield.exposure.q3 · sha256 d0f2a7…31c9 · 2,104,880 rows
policymin-cell 100 hh · min-segment 2,000 hh · signed by Meridian Markets, 2026-08-14
query budget12 of 40 used this quarter · requester: j.alvarez@northfield
enclaveIntel TDX · quote verified · measurement 1f8c44…90de
outputsha256 7e55bb…a0d4

Neither side's raw records leave its own environment, and this file is checkable without us — drop it into the open verifier and it re-derives the hashes and the enclave quote on its own.

Demonstration screen built on synthetic data with the structure and distributions of a regional grocery loyalty file. Retailer, brands and figures are invented. What is real is the method: the decomposition, the suppression rule, and the manifest format.