The IAB Tech Lab's Seller Defined Audiences (SDA) Explained

Ever since Google Chrome announced it would shut off support for third-party cookies, the industry has been searching for workable alternatives. One of those alternatives comes from the IAB Tech Lab in the form of its Seller Defined Audiences (SDA) standard.

This guide covers what SDA is, how it works, why it was created, how it compares to other alternatives, and where its advantages and limitations lie.

Key Points

• Seller Defined Audiences (SDA) is a technical specification released by the IAB Tech Lab that allows publishers to monetize their audiences without needing to use a unique ID or reveal a user's identity to advertisers.
• The IAB Tech Lab developed the specification to address the declining availability of third-party cookies while offering a privacy-focused alternative to ID-based identity solutions.
• SDA builds on other IAB Tech Lab standards: Audience Taxonomy, the IAB Tech Lab Data Transparency Standard, and the IAB Tech Lab Transparency Center.
• The specification is designed to work with existing media-buying infrastructure — specifically the OpenRTB protocol and Prebid.
• Publishers can create and sell around 1,600 contextual audiences using SDA.

What Are Seller Defined Audiences (SDA)?

Seller Defined Audiences (SDA) is a technical specification released by the IAB Tech Lab that allows publishers to monetize their audiences without needing to use a unique ID or reveal a user's identity to advertisers. The spec was developed to address the diminishing availability of third-party cookies while simultaneously offering a privacy-focused alternative to ID-based identity solutions such as universal IDs.

SDA is the first specification to emerge from the IAB Tech Lab's Project Rearc initiative. Project Rearc aims to establish new standards for the programmatic advertising industry that balance the demands for privacy and personalization. At its core are user privacy and the operational efficiency of AdTech solutions.

How Publishers Create Seller Defined Audiences

Rather than building from scratch, SDA makes use of existing IAB Tech Lab standards to help publishers label their first-party data and help advertisers determine whether a user in a given SDA segment matches their target audience.

SDA rests on three pillars — IAB Tech Lab Audience Taxonomy, IAB Tech Lab Data Transparency Standard, and IAB Tech Lab's Transparency Center — along with several smaller supporting specifications.

The Audience Taxonomy categorizes audiences in a standardized way by assigning attributes to individual users. The IAB Tech Lab has defined more than 1,600 attributes within this taxonomy to help publishers build cohorts from their first-party data. Once those cohorts are large enough to provide meaningful anonymization, they can be passed into the bidstream using present objects within the OpenRTB protocol.

The Data Transparency Standard specifies how and when data was obtained and indicates its quality. Because this information is self-attested, the IAB has created a separate compliance tool to verify which sellers are appropriately labelling their data.

The Transparency Center supports the Data Transparency Standard by enabling publishers to post data labels to a centralized resource that IAB Tech Lab members can review before purchasing a given SDA segment.

The result is that advertisers can determine which cohorts a person belongs to, and the granularity of the underlying data, without any individual-level information flowing through programmatic streams. If an advertiser wants to validate whether a given audience matches their target, they can consult the corresponding data transparency label for details such as the audience data provider, segment definition, compilation method, and data source.

What Role Do IDs Play in Creating Seller Defined Audiences?

Even though SDA does not rely on unique user IDs, publishers can still build seller-defined audiences from users who are associated with an ID. The critical requirement is that those user IDs must not be connected to the SDA segments themselves — doing so would undermine the privacy model by enabling user identification.

Publishers can also leverage existing data platforms — such as a data management platform (DMP) or customer data platform (CDP) — to create seller-defined audiences from the first-party data already stored there.

What Data Can Publishers Use?

Because the IAB Tech Lab's Audience Taxonomy is organized around three pillars — demographic, purchase intent, and interest — publishers can draw on data collected from registered and logged-in users to build SDA segments.

For example, if a user provides their date of birth when creating an account, that information can populate the demographic component of the taxonomy. As the user browses the publisher's site, that record can be enriched with interest and purchase intent data over time.

How Seller Defined Audiences Work

The SDA specification is designed to work within existing media-buying processes and standards — specifically the OpenRTB protocol and Prebid.

The workflow is straightforward:

1. The publisher creates audience segments based on their first-party data.
2. The publisher passes the SDA segment IDs to DSPs via OpenRTB.
3. The DSPs examine the segment IDs and decide whether to bid on the impression.

The diagram below illustrates how seller-defined audiences are created and communicated to buyers.

Why Seller Defined Audiences Were Created

SDA was created to address the current privacy challenges in programmatic advertising and the declining availability of third-party cookies. It is one of several proposed third-party cookie alternatives, but unlike solutions such as universal IDs, it explicitly aims to balance ad personalization with user privacy.

Traditionally, publishers activating first-party data for advertisers have had two main options:

OpenRTB auctions: Publishers make data available to advertisers by pushing user-level IDs from an SSP or DMP to a DSP.

PMP deals: Publishers establish a private marketplace deal (PMP) whereby they pass a Deal ID from their SSP to a DSP. The Deal ID can be linked to specific audience attributes.

Both of these approaches carry privacy and business-related concerns — concerns that SDA is specifically designed to address. It is for these reasons that the IAB Tech Lab has stated it will not be creating an ID solution nor will it advocate for the "broad collection, use or sharing of email addresses or phone numbers as IDs across the ecosystem."

SDA emphasizes strengthening the value of a publisher's first-party data while improving user privacy. Publishers who adopt the SDA standard can activate their first-party data for advertisers without sharing user identification information with external platforms.

How Seller Defined Audiences Compare to Universal IDs

Universal IDs were initially developed to provide a shared identity layer so that users could be recognized across the advertising supply chain with less cookie syncing overhead. Some universal IDs use deterministic data — such as email addresses and mobile IDs — while others rely on probabilistic matching. Crucially, universal IDs are not limited to third-party cookies; they can be constructed using first-party and third-party data sourced from both online and offline environments, including point-of-sale (POS) and CRM systems.

The defining characteristic of universal IDs, however, is that they revolve around identifying individual users to enable targeted advertising. That user-level identification is precisely the crux of the privacy concern in AdTech.

SDA takes the opposite approach: users are placed into audience segments based on demographic data, interests, and purchase intent, but their actual identity is never revealed to advertisers or their DSPs. Individual identification is not part of the model.

How Seller Defined Audiences Compare to Google Chrome's Topics API

Both Topics API and SDA form cohorts and share anonymized data with advertisers, but the similarities are largely surface-level. The key differences are in how audiences are generated and what signals they use.

Topics API:

• Is based on device-managed audiences created through a Google Chrome API.
• Collects information about users' browsing interests (topics) and shares those signals with ad businesses.
• Assumes that all targeting and measurement will be handled within the browser.
• Allows advertisers to target ads based on users' interests.
• Generates audiences automatically from a user's browsing history — no manual effort from publishers required.

Seller Defined Audiences:

• Are based on first-party data formed into cohorts directly by publishers, not browser-generated signals.
• Are created using IAB Tech Lab standards: Audience Taxonomy, the Data Transparency Standard, and Transparency Center.
• Allow advertisers to target based on a broader set of signals — demographics, interests, and purchase intent.
• Require publishers to manually create and maintain the segments.

Unlike Google's cohort-based Topics approach, SDA sorts online visitors into groups using a publisher's own first-party audience data rather than browser-derived data.

Advantages of Seller Defined Audiences

The primary goal of SDA is to give publishers and data providers a standardized method for defining targetable audiences that buyers can interpret and bid against confidently.

Consumer privacy is preserved. The technology avoids user-level IDs for ad targeting. Users are placed into cohorts and their identity is never exposed to AdTech businesses.

Reach at scale is maintained. Buyers can target the same cohorts across multiple publishers, and the technology offers standardized labelling and purchasing against first-party data cohorts across multiple browsers and devices.

Adoption complexity is manageable. Because SDA uses OpenRTB, Prebid.js, and established IAB standards, the toolset is already familiar to most publishers. Integration does not require building from scratch.

A quality advertising ecosystem becomes viable. Publishers can create valuable audiences that advertisers can target without exposing user-level data. SDA strikes a reasonable balance between targeting granularity and privacy protection.

Challenges and Disadvantages of Seller Defined Audiences

SDA is a well-designed standard and more privacy-friendly than many alternatives to third-party cookies, but it is not without its limitations.

Third-party cookie delay. As long as third-party-cookie-based targeting remains available, the buy-side has limited incentive to invest time and resources in alternatives like SDA.

Shift in control. Third-party cookie IDs give advertisers — or rather their DSPs — direct control over audience creation: a DSP can assign a third-party cookie to a user simply by receiving a request from a publisher's site. With SDA, control shifts to publishers, who own the first-party data and build the audience segments. This is an advantage for publishers but a relative disadvantage for advertisers accustomed to controlling their own audience construction.

Insufficient data for niche publishers. Publishers with smaller or more specialized audiences may not have enough data to reliably determine audience attributes and build meaningful seller-defined segments from user interactions on their properties.

Reduced competitive differentiation for advertisers. Because the audience taxonomies are standardized, all advertisers can potentially target the same cohorts. The lack of customization makes it harder for individual advertisers to gain a competitive edge.

Low testing rates. Few publishers, AdTech companies, data providers, or advertisers are currently running tests with SDA, so concrete performance feedback is limited.

Low demand-side priority. DSPs do not currently treat SDA as a priority — largely due to the factors above — which creates a chicken-and-egg problem: testing rates are hard to increase when one side of the market is hesitant to commit.

Despite these challenges, SDA represents a meaningful contribution to the ongoing effort to build a programmatic ecosystem that can function effectively without user-level identifiers. As third-party cookie deprecation eventually moves forward in earnest, the relevance of standards like SDA is likely to grow.