Richer contextual signals in the bidstream help agencies and publishers alike

"Programmatic buying has long relied on a combination of identity-based and contextual signals to guide decisions. Identity has been central to that model, particularly as first-party data strategies have matured and deterministic approaches have become more widely adopted. Contextual signals have also been part of the mix, but they have often been applied more narrowly, supporting brand safety, suitability or filling gaps where identity is not available."

"That division of roles has worked, but it has also limited how contextual data is used. In many cases, it has been applied at a high level, grouping content into broad categories that are easy to manage but not always precise enough to reflect what is actually happening in the moment. As a result, buyers and sellers are often making decisions based on signals that only tell part of the story."

"One of the clearest examples of this shows up in how news inventory is handled. Entire categories of content are frequently excluded because they sit within labels that are treated as inherently risky. That includes large portions of the news ecosystem, where overblocking has become a common outcome of applying broad taxonomies to complex environments."

"In practice, this means that content which is entirely appropriate for advertising, and often highly relevant depending on the message, is either filtered out or undervalued. The tools used to evaluate it are not nuanced enough to distinguish between different contexts within the same category. For publishers, this creates a persistent challenge. Valuable inventory goes under-monetized because it cannot be expressed in a way that aligns with how buyers are making decisions."

Programmatic buying has relied on identity-based signals and, to a lesser extent, contextual signals. Identity has become more central as first-party data strategies mature and deterministic approaches spread. Contextual signals have often been used narrowly for brand safety, suitability, or to cover gaps when identity is unavailable. This separation limits contextual data’s usefulness because it is frequently applied at a high level using broad content categories. News inventory is a clear example, where entire categories are excluded due to labels treated as inherently risky. Overblocking results in appropriate, relevant content being filtered out or undervalued. Publishers under-monetize inventory, and agencies lose access to environments with strong engagement and clear audience intent.

Read at Digiday