Oracle Contextual Intelligence (Grapeshot)


Oracle Contextual Intelligence provides prebid contextual advertising and brand safety solutions to advertisers, publishers, and platforms. The technology crawls and categorizes desktop and mobile web pages, matching them to contextual segments for use in digital advertising campaigns.

Market Differentiators

  • Holistic provider of advertising solutions spanning measurement, context, and audience.
  • A privacy-centric solution and trusted industry partner, backed by Oracle’s 40-year privacy history.
  • Proprietary technology that accurately categorizes content at unmatched scale and speed.
  • Dedicated linguistic team that curates segments in over 31 languages.
  • Truly custom and dynamic (Predicts) solutions, allowing for brand-specific tailoring.
  • Leverages leading measurement platform Moat Analytics for pre-bid verification suite.


Supported Platforms

  • Web
  • Mobile Web
  • Mobile App

Geographical Availability

  • Global

Vertical Specialization

  • Agnostic

Data Taxonomy

As a Basis customer, you have access to Grapeshot's syndicated taxonomy within the platform as well as the ability to create custom segments specific to your requirements.

Syndicated Taxonomy

Oracle Contextual Intelligence helps bridge the divide between risk and opportunity to find environments that align with brand missions, themes, seasons, and consumer behaviors. Because context isn’t "one size fits all", they offer unlimited custom brand safety categories so you can tailor your context strategy to suit the nuances of your specific brand.

  • Context: 250 editorially curated categories readily available to be targeted across the platform. This includes verticals like Auto, Business, Fashion, Finance, Health, Politics, Tech, Travel.

  • Brand Safety: These segments help increase the transparency and baseline control of any campaign, protect your brand equity, and deliver your message in environments that exclude risky content in categories such as firearms, crime, drugs, obscenity, death, hate speech, and more.

  • Language: Oracle Contextual Intelligence is able to semantically classify in over 30 languages.

  • Age Rating: This lets you target apps rated by age: 3+, 7+, 12+, and Adult.

  • Sentiment: Using these segments, you can find content that fits your desired tone of voice or sentiment around how a piece of content is written, enabling you to put message in front of your desired receptive audiences.

  • Grapeshot Predicts: Grapeshot Predicts (powered by Blab) predicts live audience intent up to 72 hours into the future. The tech stack absorbs vast amounts of conversational data and identifies “what, where, and when” audiences are going to be interacting with content or theme, for example, providing media buyers more intelligent/predictive keyword segments to target against. Targeting the current passions of the audience rather than re-targeting previously observed behavior offers a powerful way to reach a broad activated user base without needing to have seen individual users previously.

Custom Taxonomy

Segment Creation

Oracle's Grapeshot offers unlimited custom segment creation - context, brand safety, and Predicts. You can provide keywords and/or URLs that have the sort of content that you want to match. Centro account managers have access to the Grapeshot UI and can create segments on behalf of customers. Grapeshot provides a real-time preview of the content that a segment matches, allowing the segment to be tuned for maximum targeting accuracy. Having this adaptability ensures that the most flexible segment is created. Keyword dumping is not recommended.

Segment Addition

Segment creation is a manual process. Your account manager will create the segment for you, and the data team will then onboard the segment into the platform within 24 hours.


Syndicated: $0.04 to $0.60

Custom: $0.25


ODC's core technology is based on Information Retrieval (IR) science developed over the last few decades at the Computing Linguistics and Computer Laboratory Departments at Cambridge University. Unlike many semantic solutions that must utilize advanced machine-learning processing to model and generate a set of rules to a core system, ODC's two-step process – processing the page then matching against contextual segments - may be considered more flexible and adaptable, with greater degrees of human input. ODC’s Contextual Intelligence crawls and contextually analyzes content in response to requests. The technology and architecture support loads exceeding 3 million queries per second (QPS). Once a page has been crawled, indexed, and categorized, that information can then be compared to relevant contextual segments to determine if there is a contextual match.

Provider Collateral


The Science Behind ODC's Contextual Intelligence