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

1,148,913 category households · Meridian loyalty base

Build the audience, and see the size before you spend

Rules run against household behaviour inside the enclave. You see a count, never a list. The threshold is checked before activation, not after — so a segment too small to be anonymous never reaches a media plan in the first place.

Rules

Size updates as you choose
Bought a competitor — last 52 weeks
Any selected brand counts. Leave empty to include all category households.
Relationship with your brand
Category weight
Store cluster

Segments other brands in this category start from

Select one to load its rules
SegmentHouseholdsCategory $ / hhYour share of their basket
Counts are rounded to the nearest 100 households. Minimum activatable segment: 2,000 households. Reachable means addressable in Meridian media: 73.4% of category households. Email reaches the opted-in 48.1%; a DSP reaches 68% after identity match. Plan against the destination column, not the headline count. Protection here is a counted allowance on access plus a suppression floor — not differential privacy. Neither adds noise to these counts.

What to do with this

  1. Check the deliverable size, not the segment size. The headline count is households that match. What actually reaches a campaign is the destination figure, and for email it is roughly half. Plan against that column.
  2. Narrow until it nearly breaks, then step back. The tightest rules give the strongest audience and the thinnest reach. Push until the threshold warning appears, then loosen one rule — that is usually the efficient point.
  3. Hand the same definition to the model. Rules are transparent and blunt. On the activation screen the same intent, stated as a goal, usually produces a smaller audience with a higher predicted response.
  4. Save the rule, not the audience. Household membership shifts weekly. A stored rule re-evaluates; a stored list goes stale and quietly gets worse.
Proof of this run Template, datasets, policy, enclave quote
runcr-4e18c2a9 · 2026-10-09 15:41:02 UTC
templatesegment-build v1.8 · sha256 6fa033…91bd
retailer datameridian.loyalty.tx-52w · sha256 4a81d0…bb02 · 1,841,206 households
policymin-segment 2,000 hh · counts rounded to 100 · signed by Meridian Markets, 2026-08-14
outputaudience id only — no household identifiers leave the enclave
query budget14 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 and figures are invented. What is real is the method: rules evaluated inside the enclave, counts rounded, the minimum checked before activation, and no identifiers in the output.