Enter the Content Graph

Drawing from the successes of the past, the technology framework that will power the next generation of contextual advertising.

May 24, 2021



By Andrew Smith, Chief Product Officer



We are living in a golden age of conversation; media is the evolving centerpiece of that conversation. One common thread throughout this conversation is content: a post, a video, a photo, an article, or a TV series. This content forms the currency we create and trade in as we communicate with one another.

As content proliferates within social discourse, the content types and platforms are becoming broader and more fragmented. This means high quality content is harder to discover, less transparent, and potentially more fraught with pitfalls for advertisers. In order to make sense of it all, what we know about content must be centralized into a single architecture with the flexibility to assign, model, and exchange attributes about each individual content entity. At OpenSlate, this takes the form of our Content Graph. 

This sounds similar to an advertising technology that erupted a couple of decades ago around tracking individuals: audience targeting. And where was all the data that was collected, unified, and modeled around individuals maintained? The audience identity graph. Although under pressure today with privacy legislation, and with identity tracking being walled off or blocked by major players in tech, the audience graph is still a hugely powerful technology that can process billions of data points and can deduplicate masses of individual signals to ensure marketers can reach their intended customer.

When a similar approach is applied to content, the world of contextual advertising becomes much more powerful – enabling privacy-friendly, effective marketing at scale.  At OpenSlate, we have built our Content Graph with some incredible use cases in mind.

Drawing Connections Via Deterministic Attributes

Again, looking back at audience-based ad tech, one of the biggest challenges was understanding what cookies, logins, and device IDs corresponded to their respective individuals and households. Using a mix of deterministic data points, and with some probability thrown in, it was possible to attribute characteristics to those individuals and households.

The same is necessary for content. Advertisers require a definitive understanding of the owners and creators of content, along with other associations such as format, featured talent, series name, and date of origin, in order to draw associations across the graph.

Understanding these deterministic qualities about content entities within a single graph means that entire swaths of desirable content can be interconnected and activated across multiple platforms. This content mapping enables discovery, as well as lays a pathway to enrichment.

Data Enrichment Through Partnerships

The Content Graph becomes more powerful for the advertising ecosystem as more parties participate. OpenSlate’s trove of content data is important – but so is enriching and improving the available data via content partnerships.

Through the sharing of content information, identifiers and deterministic qualities can be used to map one content dataset to another. This enriched data lives within the graph, and allows for new products to be generated that further enhance the contextual picture. By combining datasets around content we can seek answers to questions around the authenticity, quality, origins, engagement and potential outcomes that content can deliver.

The more we know and can share about content, the deeper and brighter the contextual picture for buyers and sellers of advertising.

Data Portability, Identifiers and IDs

Data in, data out. The relationships in a content graph are the key to data exchange, making data inputs from multiple sources accessible to multiple outbound systems. Data in – from content sources like a web page, video platform, or from social media – populates new records (or entities) into the graph. Each of these content entities has an identifier pulled from the source (URL, video ID, post ID, etc) and the graph assigns a proprietary ID. If partner content attributes are ported in, deduplication can occur where the source identifier is a match, and those partner attributes are simply added to the existing content record in the graph under the graph’s proprietary ID.

Pushing data out is a similar process, in reverse. Selected records, attributes, and source identifiers can be ported to a partner to expose unique data to enrich their view of those content records. These are all assigned to the partner’s own ID within the graph. There are a few platforms out there with proprietary content IDs already in use – Gracenote and IRIS TV are two that are familiar to us here at OpenSlate. These IDs can be ingested and sent back to partners to streamline data sharing.

Building Better Contextual Solutions

Our vision is that the Content Graph will power smarter advertising strategies. When we know more about content, we can model utilizing better data, and can ensure that customers have a similar view across the digital content footprint. With this view, the true context of an advertisement can be understood. We will enter a world where discovery, activation, and measurement of content and the associated context become even more fundamental aspects of advertising workflows.

It’s time to take contextual advertising to the next level, but that look forward is also a look back. When we understand the path that audience-based advertising has taken, similar strategies around maintaining and sharing content data will only make the science of contextual advertising a more powerful tool.




Share to Facebook Share to Twitter Share to LinkedIn