Household Identification

Household identification is the IDENTITY-RESOLUTION layer that groups the many devices and people of one home — phones, laptops, the shared smart TV, the family tablet — into a single household unit. It sits one level up from person-level identity: where the GOLDEN-RECORD resolves one human across systems, household ID resolves one address across humans and screens, which is exactly the unit that connected TV, frequency management, and offer-suppression actually need.

The signals that build the graph, by reliability: shared IP address (the workhorse — devices on one home network share an IP, the basis of most CTV household graphs, but confounded by dynamic IPs, VPNs, and the apartment building behind one NAT), login and account graphs (the streaming account, the retailer's family profiles, the deterministic spine where it exists), data co-ops and graph vendors (probabilistic stitching across the ecosystem, accuracy varying by vendor and verifiable mostly by the buyer's own panel checks), and offline anchors (the postal address behind a loyalty account — the CRM's truest household key). Where it matters most and why: connected-TV is a household device by nature (the living-room screen has no single user), so CTV FREQUENCY-CAPPING, reach measurement, and cross-screen attribution are household problems wearing device clothes — manage frequency per device and one family absorbs every cap simultaneously (the 40-impressions-a-week complaint). The privacy fault line this entry must name: household-level targeting can become a proxy for protected attributes and shared-device exposure can leak sensitive intent across a family (the health-or-gift purchase surfacing on the shared tablet — the RETARGETING-on-a-shared-device problem), so 'household' is a sensitive unit under modern privacy law, and the data-co-op supply chains behind it carry the same broker-diligence burden as any third-party graph. The honest accuracy caveat: household graphs are probabilistic infrastructure - precise enough for frequency and reach, too blunt for anything that punishes a wrong join.

Household identification matters wherever the home is the real buying unit — CTV and cross-screen video (its headline use), retail and CPG where one cart serves a family, and frequency or suppression that only makes sense at the address. It matters as a measurement honesty check (per-device reach and frequency numbers quietly lie when the household is the audience) and as a privacy-design constraint (shared-device exposure and proxy-targeting are the named risks). The discipline is deterministic anchors where available, probabilistic graphs priced for their accuracy, frequency and reach read at the household level for household media, and the sensitive-unit treatment applied before the targeting gets clever.

Household identification grew from television's panel-measurement era (Nielsen's household as the unit of TV ratings) into digital's cross-device problem as homes filled with screens; connected TV made it urgent again - the living-room device forcing buyers to plan, cap, and measure at the address the way broadcast always had.

Etymology: source.

Search interest for this term over the last five years:

View interest-over-time on Google Trends →

Areas of marketing where household identification is a core concern: