Customer data glossary
The vocabulary of customer data platforms, in plain language. One definition per term, with what it means in practice underneath.
Customer Data Platform (CDP)
A CDP is a system that builds one persistent, unified profile per customer from first-party data, and makes that profile available to other systems.
It ties events to a known person using identifiers such as an email address, phone number or account ID, keeps that history for years, and is used for personalization, retention and omnichannel orchestration. The data comes from your own website, app, CRM and point-of-sale systems — never bought from a third party.
Data Management Platform (DMP)
A DMP collects and organizes anonymous audience data for advertising, then pushes it to ad platforms for targeting.
It works with cookies, device IDs and IP addresses rather than personally identifiable information, and traditionally leaned on third-party data bought from data providers. Storage is short-term — typically around 90 days — because its job is capturing immediate intent for acquisition, not remembering a customer for years. Full comparison of DMP and CDP.
Identity resolution
Identity resolution is the process of deciding that activity seen on different devices or sessions belongs to the same person, and merging it into one profile.
Deterministic resolution links records that share a confirmed identifier, such as the same logged-in email address. Done in real time, it is what lets a system know that the visitor browsing on a phone now is the one who abandoned a cart on a desktop yesterday.
First-party data
First-party data is information you collect directly from your own audience, with their consent.
Website behaviour, purchase history, CRM interactions and survey responses all qualify. It is the foundation of privacy-safe targeting, because it does not depend on cross-site tracking or on data bought from anyone else.
Anonymous visitor tracking
Anonymous visitor tracking profiles the visitors who have not identified themselves, so they can be segmented before they ever log in or share an email address.
Typically 97–98% of website visitors are anonymous; only 2–3% end up identifying themselves. Profiling the rest, and connecting their devices to profiles as evidence accumulates, reduces the share of traffic you know nothing about.
Server-side tracking
Server-side tracking moves data collection off the visitor's browser and onto your own server, which then forwards what you choose to third parties.
It is more reliable than browser pixels, which ad blockers and browser privacy features increasingly block; it removes JavaScript from the page, which helps Core Web Vitals; and it puts you in control of exactly what leaves for each destination.
Customer match
Customer match is the mechanism ad networks use to find their own users from a list of hashed identifiers you upload.
You send SHA-256 hashes of normalized email addresses rather than raw ones, so the network can compare without either side exchanging plain contact details. Every network sets a minimum number of matched users before the resulting audience can be targeted, and matching is asynchronous — the matched size appears hours or days later. See the thresholds per network.
Match rate
Match rate is the share of the identifiers you upload that an ad network recognises as its own users.
It is capped by how many of your customers use that network with the same email address. Sending identifiers a network cannot match — phone numbers to a destination that only accepts hashed email, for instance — only depresses the number.
Real-time segmentation
Real-time segmentation evaluates a visitor's membership of an audience as they act, rather than in a nightly batch.
The distinction matters when the response has to happen inside the same session: a banner, a reordered search result or a personalized offer is useful now and worthless tomorrow morning.
Computed traits
Computed traits are profile attributes derived from raw events rather than collected directly.
Lifetime order value, most-viewed category or days since last purchase are all computed: nobody submits them in a form, the platform maintains them as events arrive so that segments can be built on them.
Data replay
Data replay re-processes historical events through current logic, so a new segment is populated with people who already qualified in the past.
Without it, a segment created today starts empty and only fills as new activity arrives, which can mean waiting weeks before it is large enough to use.
Where these show up in the product
Identity resolution, computed traits, real-time segmentation and data replay are all part of the CDP & DMP. Customer match and match rate belong to audience activation. Server-side delivery is what the first-party ad server is built on.