An open standard built on MCP and developed by AgenticAdvertising.org that defines how AI agents communicate within the advertising ecosystem. The protocol covers audience discovery, deal negotiation, and campaign activation between publisher and buyer systems. Optable is a founding member of AdCP.

The portion of a publisher's total audience that can be identified and targeted with a known identifier. A larger addressable audience directly improves match rates, CPMs, and the value of data partnerships.

A digital marketplace that connects supply-side platforms and demand-side platforms, running real-time auctions where impressions are bought and sold programmatically. The ad exchange is the neutral clearinghouse in the middle of the supply chain; SSPs bring publisher inventory in, DSPs bid on it, and the exchange matches them impression by impression.

A platform that manages the delivery, targeting, and reporting of ads across a publisher's inventory. The ad server is the decision-making layer that determines which ad runs on which impression.

An emerging model in which autonomous AI agents execute advertising workflows, including audience discovery, deal negotiation, campaign activation, and optimization. Agentic advertising is the next evolution of programmatic, where software agents communicate directly across the buy and sell sides.

Optable's network-level infrastructure that connects publisher inventory to a growing ecosystem of AI-powered buyers through protocols like MCP and AdCP, enabling buyer agents to discover, plan, and activate against publisher audiences directly.

A multi-step process in which one or more AI agents coordinate to complete a complex task without requiring a human to manage each step individually.

Direct communication between AI agents, such as a buyer agent negotiating with a publisher's sales agent, without a human intermediary, enabled by shared protocols like AdCP.

A software system powered by an LLM that can perceive inputs, reason about them, and take actions autonomously or with human oversight. Unlike traditional automation, agents can handle open-ended, multi-step tasks, such as reading an RFP, querying a dataset, or executing a deal.

A discrete piece of data associated with a user profile, such as age range, content category affinity, purchase intent signal, or subscription status. Publishers use these to define and package audience segments with precision.

Frameworks for assigning credit to the touchpoints in a customer's path to conversion, including first-click, last-click, linear, and time-decay.

The process of making a defined audience segment available for targeting across advertising channels by pushing it to a DSP, ad server, SSP curation desk, or data clean room. Activation is the final step that converts audience data into revenue.

An AI agent that automates many aspects of audience planning: reading RFPs, interpreting campaign goals, building audience segments, and recommending inventory packages for activation, all with human-in-the-loop review. Optable's Audience Agent is the first agentic ad planning tool built for publishers, operating across identity, audience, and activation workflows in a single system.

A publisher's ability to reach its own known audience on inventory beyond its owned-and-operated properties, extending the value of its first-party data across the open web.

A measurement of how many users appear in both a publisher's audience and an advertiser's customer or prospect list, informing audience packaging and pricing.

The process of grouping audience segments into structured, marketable products for advertisers, often including defined segment criteria, scale, match rate estimates, and pricing.

A group of users sharing one or more attributes — behavior, demographics, purchase intent, subscription status — treated as a single targetable unit. Segments are the basic building block of audience strategy; attributes define them, packaging makes them sellable.

The structured classification system used to name and organize audience segments into consistent categories. A shared taxonomy is what makes audiences interpretable to buyers and portable across platforms.

An identifier tied to a user who has logged in or confirmed their identity, making it more reliable than probabilistic signals and the most valuable signal type for addressability and targeting.

Targeting based on a user's past actions, such as content consumed, purchases made, and categories browsed. This requires persistent user identification, clean first-party data, and identity infrastructure.

A system where inputs and outputs are visible but internal logic is not. Much of adtech has historically operated this way, leaving participants in the dark about how data is matched, how audiences are built, and how revenue is attributed.

An AI agent operating on behalf of an advertiser or agency to discover publisher inventory, evaluate audience fit, negotiate terms, and activate campaigns autonomously by interacting with publisher systems via protocols like AdCP.

A California privacy law giving residents rights over how their personal data is collected, used, and sold. This is a key compliance consideration for publishers and advertisers operating in the U.S.

A privacy-preserving environment where publishers and advertisers can combine datasets without exposing raw data. Clean rooms enable audience overlap analysis, campaign measurement, and suppression without requiring a data hand-off.

Software that collects, stores, and communicates a user's data permissions across the supply chain via standardized signals. CMPs are how publishers operationalize consent under GDPR and CCPA.

A machine-readable record of a user's data permissions, capturing what they've agreed to, under which legal basis, and when.

Targeting based on the content of the page a user is viewing rather than who the user is, making it especially relevant in a cookieless environment.

The ongoing phase-out of third-party cookie support in browsers, driven by privacy regulation and browser policy. Cookie deprecation is the central force behind the shift to first-party data and alternative identity solutions.

A pricing model in which the advertiser pays each time a user clicks the ad, rather than per impression served. CPC ties cost directly to engagement, making it common for performance and direct-response campaigns.

A flat-rate pricing model in which an advertiser pays a fixed fee to own an ad placement for a full day regardless of impressions delivered. This is typically used for high-visibility buys like homepage takeovers or roadblocks.

The price an advertiser pays per thousand impressions. This is the standard pricing unit in display and programmatic advertising.

The ability to recognize the same user across the multiple devices they use — phone, laptop, tablet, connected TV — and link those touchpoints to a single profile.

A device-level identifier associated with a connected television. CTV IDs vary by platform and are not standardized, making cross-platform identity resolution on CTV particularly challenging.

The practice of packaging publisher audience data with specific inventory into a deal accessible to buyers via their existing DSP, allowing publishers to monetize their first-party data without exposing it directly.

A platform that unifies customer data from multiple sources into persistent, individual profiles for use across marketing and advertising. CDPs are built around first-party, consented data and are designed for customer relationship management.

A client-specific identifier assigned by a publisher or platform to recognize users within their own ecosystem, often mapped to external IDs via an identity graph.

Optable's term for a publisher's private, encrypted environment housing their identity graph, audience data, and clean room capabilities within the Optable network. Each DCN is independently operated, meaning no other party, including Optable, can access raw data without explicit permission.

A legacy data platform used to collect, segment, and activate audience data for targeting. DMPs are largely being replaced by first-party data infrastructure as cookies deprecate and privacy regulations tighten.

The process of bringing offline or first-party data, such as a CRM list, into the digital environment by matching it to online identifiers so it can be used for targeting and measurement.

A unique identifier that connects a specific buyer to a specific inventory package under agreed terms, serving as the operational unit behind PMPs, programmatic guaranteed deals, and preferred deals.

Technology that allows advertisers and agencies to buy digital ad inventory programmatically, applying targeting, bidding logic, and budget controls in real time. Common DSPs include The Trade Desk, DV360, and Amazon DSP.

An external platform an audience is activated to from a publisher's DCN, such as a DSP, SSP, ad server, or curation platform. The data exported to a destination consists of type-prefixed identifiers and identity clusters, making the audience targetable in that environment.

Optable's capability for automatically and continuously keeping activated audiences up to date across their destinations. Publishers configure a destination once and Optable refreshes the audiences in bulk on an ongoing basis.

Linking identifiers based on confirmed first-party signals such as a shared login or hashed email, producing high-confidence matches.

A device-level identifier assigned by mobile operating systems, including Google's GAID on Android and Apple's IDFA on iOS, used for ad targeting and measurement on mobile devices.

A publisher-side metric expressing total revenue earned per thousand impressions, regardless of how the inventory was sold. eCPM is the common denominator publishers use to compare the yield of different demand sources.

The percentage of bid requests that are successfully enriched with at least one additional identity signal. A higher enrichment rate directly correlates with addressability and CPM improvements.

User actions or behaviors captured in the DCN, used alongside traits to build audiences. Where a trait describes who a user is, an event describes what they did, letting publishers tailor segments around activity, not just static attributes.

The percentage of ad requests that result in a served impression; gaps may indicate floor price mismatches, targeting constraints, or insufficient demand.

Data collected directly from a publisher's or advertiser's own users through logins, subscriptions, purchases, and on-site behavior. Authenticated 1P data is the highest-quality signal available in a post-cookie environment.

Optable's lightweight, partner-controlled collaboration tool that allows a publisher to invite an external advertiser or agency into a clean room environment.

The minimum bid a publisher will accept for an impression.

A limit set on how many times a specific user sees a given ad within a defined time window, preventing overexposure and ad fatigue.

A platform-specific identifier used in gaming environments to recognize users across sessions and devices, increasingly relevant as gaming publishers build audience monetization strategies.

The European Union's comprehensive privacy law governing how personal data is collected, stored, and processed. This requires publishers to obtain explicit consent before collecting data.

A system-generated string used to uniquely identify a record or user within a platform, often assigned before being linked to external identifiers in an identity graph.

A browser-level signal that allows users to automatically opt out of data tracking and selling across all sites they visit, without configuring each one individually. Publishers are required to honor GPC signals under CCPA.

An IAB Tech Lab framework that packages multiple regional privacy signals into a single transportable string, simplifying consent management across jurisdictions.

A single, persistent, unified view of a user synthesized from all available identifiers and attributes in an identity graph. Also referred to as a Single Customer View (SCV) or ID profile, the golden record consolidates behavioral and transactional data across devices and channels.

An email address that has been run through a one-way cryptographic function to produce a fixed-length string that cannot be reversed to reveal the original address. HEM is the most widely used privacy-preserving identifier for cross-party matching, as two parties can confirm a shared user without exchanging raw data.

A technique that lets multiple buyers compete for an impression simultaneously before the publisher's ad server makes a decision, often resulting in higher CPMs.

An identifier that groups individual user profiles into a shared household unit, typically using IP address and device graph signals, and particularly useful for CTV targeting and frequency management when multiple people share a device.

A design approach in which an AI agent proposes or drafts actions but a person reviews and approves them before they take effect, keeping human judgment in control of high-stakes decisions.

A measure of how many identifiers are linked to a given user profile. Higher identity density means a user can be recognized and reached across more environments, increasing their value for targeting, enrichment, and measurement.

Also referred to as universal IDs or alternative IDs. These shared standards define how user identities are tokenized, represented, and passed through the ecosystem as a privacy-preserving replacement for third-party cookies. Common frameworks include UID2, ID5, RampID, and Yahoo ConnectID. Optable's ID Switchboard supports all major frameworks natively, allowing publishers to activate multiple identity standards from a single integration.

Also called an ID graph or identity spine, an identity graph is a data structure that links all known identifiers for a user into a single persistent profile, grouping related signals into identity clusters that represent a device, a person, or a household.

The process of linking identifiers into a unified profile representing a single person across devices, sessions, and environments. Effective cross-device identity resolution is the foundation of accurate audience targeting, frequency management, and cross-channel measurement.

Optable's publisher-controlled layer that manages which external identity providers are active in the bid stream, distributing tokens such as UID2, ID5, and RampID into outgoing bid requests. Publishers decide which IDs are active — not Optable or demand partners.

A single instance of an ad being served to a user. The impression is the base unit of measurement and transaction in digital advertising; inventory, pricing, and performance are all counted in impressions.

A measure of the true impact of an ad campaign, isolating the conversions or outcomes that wouldn't have happened without the advertising. This is more rigorous than last-click attribution and increasingly important to advertisers evaluating publisher data partnerships.

The ability of different identity systems and platforms to recognize and exchange the same identifiers, so an audience built in one environment can be activated and measured in another.

A method of passing audience segment membership to an ad server in real-time using custom parameters, enabling audience-based deal targeting without exposing user-level data to the ad server.

A deep learning model trained on large volumes of text, capable of understanding and generating human language. LLMs are the AI foundation underlying tools like Claude and ChatGPT, and the engine powering Optable's sell-side agents.

A modeled audience built by identifying users who share behavioral and demographic characteristics with a defined audience, extending reach beyond known users while maintaining relevance.

An identity inference engine that links anonymous or unknown signals to a likely email address or other identifier by checking against third-party data providers. Optable's Matcher orchestrates multiple providers in a prioritized waterfall, scoring each result for accuracy so that only high-confidence identities are activated in the bid stream.

The percentage of records in one dataset that can be linked to a corresponding record in another dataset. In publisher-advertiser collaboration, a higher match rate allows for more precise targeting and measurement.

An open standard developed by Anthropic that defines how AI agents communicate with external tools, data sources, and other systems, providing the technical foundation for agent-to-platform and agent-to-agent interactions.

An attribution approach that distributes conversion credit across multiple touchpoints in a user's journey, rather than crediting only the first or last interaction.

An approach to multimedia advertising that coordinates targeting and messaging across multiple channels, including display, video, CTV, audio, and more.

A protocol that governs how bid requests and responses are formatted and exchanged in real-time bidding (RTB), which is the millisecond-level process by which a buyer's bid is evaluated and an ad is served each time a user loads a page.

The internet outside of closed, proprietary platforms, where independent publishers own and operate content accessible via programmatic advertising infrastructure.

A durable identifier that remains stable across sessions and environments, used to maintain continuity of a user profile over time.

Any data that can be used to identify a specific individual, such as name, email address, or phone number. PII is subject to strict protections under privacy laws, so it is typically hashed or tokenized before being shared across systems.

An open-source header bidding framework that enables publishers to offer inventory to multiple buyers at once before their ad server makes a decision.

A category of tools and techniques including clean rooms that enable data analysis and collaboration while limiting the exposure of individual-level information. PETs are increasingly required for compliant data partnerships.

An invite-only programmatic auction where a publisher offers premium inventory to select buyers via a Deal ID. Publishers control who participates, at what floor price, and against what audience or contextual criteria.

Linking identifiers based on statistical inference from signals like IP address, device type, and behavior. It extends reach where deterministic matches aren't available, at the cost of some accuracy.

The automated buying and selling of ad inventory through software and real-time auctions rather than manual insertion orders. Programmatic spans open real-time bidding, private marketplaces, and guaranteed deals.

A deal type reserving a fixed volume of impressions for a specific buyer at a set price, combining the precision of a direct deal with programmatic automation.

Targeting new users who have not yet engaged with a brand, typically at the top of the funnel, often using lookalike or modeled audiences to find likely prospects.

A record of where an identifier or data signal originated and how it was collected, allowing partners to verify that data meets their compliance requirements.

A publisher-assigned identifier that allows publishers to recognize their own users across sessions without relying on third-party cookies or device IDs.

The millisecond-level auction that runs each time a page loads, in which buyers submit bids for an individual impression and the highest bid wins the right to serve.

A brief from a media buyer outlining campaign objectives, target audience, budget, and flight dates, to which publishers respond with inventory proposals, audience packages, and pricing.

A partner API that translates a confirmed identifier into an encrypted, transactable ID safe for use in OpenRTB bid requests. Where matchers infer identity from anonymous signals, resolvers take a known identity and convert it into specific formats required by different platforms or channels.

The percentage of a publisher's total traffic that can be assigned a known, persistent identifier as the foundation for assessing addressability.

An advertising business run by a retailer that monetizes its first-party purchase data across its owned properties and the open web. Retail media is one of the fastest-growing channels and a leading example of first-party data monetization at scale.

Also called remarketing, this entails serving ads to users who have previously interacted with a brand or visited a site, based on known identifiers, to re-engage them further down the funnel.

Revenue earned per thousand pageviews or sessions. This is a page- or site-level measure of monetization rather than a per-impression one.

The measurable growth in publisher revenue attributable to a specific change, typically expressed as a percentage increase in eCPM or RPM.