Contextual Targeting vs. Cookies Advertising
Cookie technology arrived a bit over 25 years ago as a sort of marketing miracle. Having been working in TV before I worked in digital, I looked on the tech with awe - you could actually see how many people (OK, devices) had been exposed to a campaign, and see the level of response, purchases and the amount spent, even if they had not clicked on an ad.
Early on I explained cookies to people using the - possibly apocryphal - story of a street market in Malaysia. Traders at the market had cooperated to produce three sets of bags for people's purchases, and worked out a code. Give a striped bag to a shopper who haggled hard, give a spotted bag to shoppers who managed to get a small discount, and give a plain bag to shoppers who did not haggle at all. This allowed other stall holders to assess a potential customer as they browsed, and know how to engage with them. Cookies were like this because they gave sites and advertisers some level of intelligence about web users - their probable interests, recent online history, and a likelihood that the ad would be relevant to them.
Of course a lot has happened in the last 25 years, to make cookies feel a lot less miraculous. We use more devices, we use apps as well as the web, and we have all seen the negative side of cookies. Even the most passionate supporter of the ad-funded digital ecosystem gets frustrated by seeing products from pages visited once endlessly track you around the web.
Beyond this we have seen the rise of legislation like GDPR, blocking technology, and particularly the actions of Apple diminishing the potential of cookies even further.
While it is easy to become very nostalgic about the powers of cookies, we should maybe welcome their passing, and see it as a necessary spur to try to find better technologies or other advertising methods like contextual targeting.
Just as we have more screens, we also have more data, and better AI, and it is quite possible that ads can work better with other forms of targeting.
What is contextual targeting in advertising?
Contextual targeting has always been part of advertising. Just as the medium tells you a lot about the message, it also tells you a lot about the person consuming it. It has now become much more sophisticated than ‘car ads in car pages’ (or even ‘certain car ads on sports pages’), but instead advances in artificial intelligence allow us to see the context of a page in milliseconds and place a relevant ad that is likely to work well. It is even possible to reverse engineer contextual mix by looking to see where recent audience-targeted campaigns ran, and which pages worked best, to produce future campaigns planned entirely based on context.
AI will also allow us to create similar cohorts of web users. New technologies allow users to be grouped based on similar browsing behaviour, and it is possible to imagine this working well for clearly identifiable groups like luxury buyers, or people interested in certain classes of cars.
How Contextual targeting works?
- A crawler scans the web and categorizes pages based on context and semantics.
- When a user visits a page, that page content information goes to the ad server, which then matches it with relevant ads for the keywords and content.
Contextual targeting in the Cookieless World
To find more information on this topic download the dentsu report The Cookieless World - A Guide for the New Era of Digital Marketing