Audience Data Glossary

Plain definitions for the terms used across AMAP Data — how cohorts are built, how geographies are scored, and how a cookieless audience reaches a buying platform.

Cookieless audience data

Audience data that carries no cookie, device identifier or personal record. Instead of tracking a person across sites, a cookieless segment describes a place: the measurable characteristics of a geographic area. Because nothing individual is stored or passed, the segment keeps working when cookies are blocked, and it does not create regulatory exposure for the buyer.

Census block group

The smallest geographic unit for which the US Census Bureau publishes detailed sample data, typically covering a few hundred to a few thousand people. AMAP uses the block group as its base unit of targeting: every segment is a score attached to block groups, which is granular enough to be useful and aggregate enough to contain no personal information.

Cohort

A group of geographies that share a defined characteristic — a category interest, a life stage, a household profile, a health or purchase signal. A cohort is the buyable object in AMAP: rather than a list of users, you buy the set of block groups where the described behaviour or attribute is concentrated.

Over-index

A geography over-indexes when a characteristic appears there more strongly than it does across the comparison population. Over-indexing is what makes a cohort worth buying: it tells you the audience is unusually concentrated in that area, so impressions delivered there reach proportionally more of the people you care about than a flat geographic buy would.

Audience index

The numeric score AMAP assigns to each geography for a given segment, combining its input signals into a single comparable value. The index lets you rank markets, compare block groups against each other, and set a cut-off that decides how much of the country enters the audience. Higher index means stronger concentration of the described cohort.

Search listening

The practice of turning live search behaviour into audience signal. AMAP groups queries into clusters that describe a real intent — researching a model, comparing treatments, planning a trip — and resolves the volume of those queries to geography. The result is a segment that reflects what people are actively looking for right now, without touching any searcher's identity.

Social listening

Interest and affinity signals derived from public social activity and aggregated to geographic cohorts. Where search listening captures active intent, social listening captures sustained interest and cultural affinity. Both are modelled at cohort level, so an interest-based audience can be bought and activated with no cookies, hashed lists or device identifiers involved.

The Geograph

AMAP's underlying model of the data signature of places. The Geograph combines census and public datasets with proprietary search and social signals and resolves them all to the same geographic grid, so any segment can be scored against any geography. It is the layer that makes cohorts comparable, composable and consistent across channels.

Activation

Getting a built audience into the platform where media is actually bought. AMAP segments activate through The Trade Desk and every major DSP, and the same cohort can run across CTV, display and DOOH. Because the audience is geographic, activation does not depend on matching users between systems.

First-party onboarding

Turning a client's own data into a modelled audience without an identity graph. Any dataset with a geographic field — latitude and longitude, postal code, address, county or city — can be aggregated to geography, scored against the Geograph and expanded into a lookalike cohort. No individual records are retained, matched or passed on to a platform.

Curation

Assembling cohorts, inventory and deal terms into a single ready-to-buy package rather than leaving the buyer to reconstruct targeting in each platform. Curation moves the audience decision upstream: the segments are composed once, wrapped into a deal, and delivered so the buying team activates a finished audience instead of rebuilding it channel by channel.

Deal ID

The identifier that carries a curated package into a buying platform. A deal ID bundles the audience and the agreed terms so a trader can activate everything with one line item. It is the practical delivery mechanism for curated cookieless audiences, because it moves the targeting logic into the deal rather than into user-level data.

CTV addressability

The ability to reach a specific audience on connected TV. Traditional CTV addressability leans on device graphs and household matching. Geographic addressability reaches households through the areas where a cohort over-indexes, which means the same audience definition can run on CTV, display and DOOH without any device identifier being involved.

POI targeting

Buying media around points of interest — stores, venues, clinics, competitor locations. On its own it treats every location as equally valuable. Combined with cohort data, POI targeting narrows to the locations whose surrounding geography actually over-indexes for the audience you want, so budget concentrates on the places where the buyer is present.

Threshold methodology

The rule that decides when a geography qualifies for a segment. Setting the threshold high produces a small, sharply concentrated audience; setting it lower produces broader reach at a weaker index. Threshold methodology is how AMAP makes that trade-off explicit, so a buyer can choose precision or scale deliberately rather than inheriting a default.