Finding an online advertising compromise

Tracking cookies and automated profile of internet users allow commercial concerns to create targeted advertising for users of social media and other websites. However, the privacy issues surrounding such approaches to marketing worry many people. Researchers in the UK have examined the implications of an alternative approach that would give users back some control of their personal information but allow companies to offer them advertising that would be precisely targeted and perhaps meaningful to the users. The solution lies in what computer scientist Reuben Binns of the University of Oxford refer to simply as a Self-Authored Interest profile.

Binns explains that much of the content we see on the internet, advertising and product and service recommendations in particular are fed to us on the basis of algorithms that track our internet history and behaviour. Needless to say, this causes friction between companies and consumers and has led to the emergence of ad-blocking software, the need for "do not track" and cookie crumblers for browsers. These are perceived as allowing users to take back some control of their privacy.

Many users do not want to see any advertisements at all and many do not want to be tracked. However it is usual that free online services rely on revenues from advertising and would not survive if they could not display advertisements to users. Those sites make more money if the advertising is targeted to users' behaviour and interests and many run systems that block the advertisement blockers and circumvent the privacy rules.

A compromise is needed with which both company and consumer would be happy. Binns suggests that the self-authored profile might be the answer. He has tested this approach against behavioural profiling and found that it has many benefits. "People respond more positively to product recommendations when they are derived from SAI profiles," he says. "Moreover, the mere belief that a recommendation comes from an SAI profile is also associated with more positive responses."

Any advertising system is likely to be imperfect. Non-targeted ads are annoying to many users and do not necessarily result in sales leads. By contrast, targeted advertising works better in presenting offers that a user is more likely to be interested in but they only work if the advertiser has access to more private or personal information about the people they hope to advertise to. If users are given some control over what information the advertisers get to see, then they are, it seems, more responsive to advertisements targeted to them based on that information, Binns suggests.

The next step will be to test whether such an approach can overcome consumer scepticism as well as be effective for the advertisers. Different trusts models will also need to be built to test the legal limitations and the effectiveness of this approach to privacy with targeting.