Staying on topic with contextual

Pete Danks, VP, Product at Magnite

May 23, 2023 | 5 min read

While uncertainties remain around Google’s Privacy Sandbox, and other identity alternatives face questions around long term prospects, one area that most agree on is that context will likely play a significant role as a pillar of addressability.

Contextual as a pillar of addressability

Contextual targeting is an already viable, sustainable and effective complement or alternative to identity-based addressability that fundamentally embraces the privacy-centric paradigm without any reliance on user data. Modern contextual targeting has been transformed by innovations such as; natural language processing (NLP); automated image/video analysis and speech-to-text engines have led to more accurate and automated content analysis. These insights provide the foundations upon which many audience identifiers are built on such as interest signals. 

Contextual across channels

Contextual intelligence is starting to be leveraged across channels from web-based contextual to CTV/OTT, audio and gaming. For web-based contextual understanding Magnite uses natural language processing (NLP) and contextual modeling processes to automate the extraction of themes and keywords of media owners’ pages. These are then matched to the IAB’s taxonomy or to more bespoke segments, to power custom contextual segments, brand safety and suitability, as well as campaign planning and real-time targeting.

On CTV, contextual intelligence comes from metadata and content signals such as genre, series and channel, and while these can be included in bid requests there are still challenges to achieve the transparency and scale seen in web-based contextual. For instance, the broad nature of content signals (e.g. genre), the need to tag all content and metadata, and concerns around data leakage can all be obstacles.

Automatic content recognition (ACR) data is bolstering contextual capabilities in CTV and audio, where tech can identify what is playing and match it to an associated content signal, in order to build audiences, manage frequency, measure viewability, and do attribution. As smart TVs continue to proliferate there’s an opportunity to leverage ACR data more effectively – increasing its availability, standardizing its use and measurement, and enriching it with other data to improve ad serving.

What this means for media owners

Curate contextual inventory

Curation of inventory packages across flexible execution options (open market, programmatic guaranteed, private marketplace deals, etc) using deal IDs provides buyers access to contextual inventory while alleviating concerns around possible data leakage by stripping the bid request of metadata publishers do not wish to share.

Standardize and differentiate

In addition to mapping audience segments based on standardized content taxonomies, publishers should look to highlight their unique audiences, mapping content, keywords and behaviors into their own custom taxonomies.

Protect their data

Media owners should seek tools to control what data they share with buyers and prevent data leakage. For instance, Magnite’s Data Lock capability offers signal filtering capabilities in CTV environments to dictate what content information is shared with DSPs.

What this means for advertisers and agencies

Contextual now and in the future

Many buyers are already regularly using valuable content parameters in display to make more informed bidding decisions that can improve scale and accuracy, while reducing waste by targeting relevant content. As contextual data becomes more available across channels, buyers will see similar benefits elsewhere.

Working with the supply-side

Buyers should explore supply side activation paths that leverage a portfolio of data enablement options that includes using contextual data for accurate audience mapping and forecasting, inventory and audience curation, and detailed insights.

Beyond inclusion/exclusion

Advertisers can leverage both standardized taxonomies as well as some publishers’ custom taxonomies to access even more granular and accurate contextual intelligence and targeting opportunities beyond basic keyword inclusion/exclusion.

What’s next for contextual?

Going forward, contextual will be an important pillar of addressability and publishers should take steps to maintain control of how and where their contextual data is leveraged. While publisher contextual signals are often sold by third parties, the best way to guarantee accurate contextual data is when accessed from the publisher themselves as there’s a risk with third parties miscategorizing such signals. Though scale is one potential drawback of buying contextual from individual publishers, the standardization of contextual signals is creating possibilities for buying against contextual data from multiple publishers – through SSPs – where publishers maintain control over their data.

Collaboration will be critical for contextual activation to ensure consistency between signals the publisher provides and the ability for the buy-side to understand and bid on the right signals. For instance, while web-based contextual benefits from standardized IAB taxonomies, there’s a need for standardization specific to TV that can then be adopted across video platforms too.

As addressability challenges increase due to third party cookie deprecation, Magnite is helping publishers and buyers use other signals by enriching bid requests across omnichannel inventory with contextual and behavioral signals. Innovation around contextual targeting will continue with real-time contextual curation and targeting, as well as concepts such as contextual lookalikes, where we can infer signals based on a user consuming similar content to a previous user.

Want to learn more? Why not sign up to Magnite University to learn more about contextual targeting, key ad tech industry trends such as retail media and ACR data, and more.

Contact Us