Contextualisation for insight

Saul Stetson

Categorising content and building user profiles and aggregated behaviour models helps marketers become more scientific in discovering who their ideal customers are likely to be, sometimes with surprising results.

Data contextualisation is the methodology that categorises -in precise detail – specific areas of content with which people are engaging. The internet is essentially an ecosystem of unstructured information but contextualisation technology is able to scan a web page, understand its content, and assign it a preset category, or topic, that will sit within a more general category chain. For example, entertainment is a general category which can then be split out into film, then action films, then a particular actor.

As a user navigates their way around the web on a day-today basis, their consumption of various categories can be added to an anonymous user profile. When these user profiles are aggregated around a central seed or action, such as a visit to a brand's website, a model can be created that offers the statistical likelihood (in comparison to the average) of that brand's customers consuming information in the categories highlighted. This is referred to as a 'lift' in behaviour and we call these interest-based audience segments. For example, people who visit cricket websites are 64 times more likely to be interested in men's suits; tennis fans are 11 times more likely to be interested in the actor Robert Pattinson; and so on.