Contextual advertising

Contextual advertising (also called contextual targeting) places ads based on the content a person is viewing. Wikipedia describes ads "selected and placed by automated systems based on websites' linguistic content" — a sports site shows ads from sports-related companies.

How it works. Amazon Ads describes two steps: automated systems analyse the content — text, images and other media — to understand its context, then relevant ads are selected to match its themes. Wikipedia describes systems that scan a page for specific keywords and phrases. GumGum describes newer approaches that use natural language processing to read meaning and sentiment, and computer vision to process images and video.

Contextual vs. behavioral. Behavioral targeting segments people by browsing behaviour — pages visited, searches, clicks and purchases, as Criteo puts it. Amazon Ads: "behavioral targeting uses past audience data and behaviors, while contextual targeting focuses on the current web page content." GumGum sums up the difference: contextual ads are placed in environments that align with user interests "with no reliance on personal data."

History and revival. Wikipedia calls Google AdSense, launched in 2003, the first major contextual advertising network. It reports a resurgence of contextual advertising driven by privacy rules such as the GDPR, the loss of third-party cookies and Apple's App Tracking Transparency.

How it is sold today. Peer39's Contextual Data Marketplace gives marketers access to pre-bid contextual data categories from multiple providers. Criteo says newer contextual products can add commerce signals from first-party data to page-level context. GumGum says its Verity technology analyses text, images, audio and video, and has Media Rating Council accreditation for content-level contextual analysis across CTV, desktop and mobile web.