The 'Big Brother' Advertising Framework

Advertising still remains the number one income source for most websites. Whether you are a blogger who gets a handful of visitors or a company such as Facebook, it is likely you rely on adverts to generate a revenue. Nowadays you can start making money within minutes, all you need is a Google AdSense account and a couple of your best friends to get you started. Most solutions rely on content from your pages, but users visiting your website will not always be interested in what you have to say, and thus what the adverts have to offer.

Much like the popular TV show ‘Big Brother’, let’s imagine an advertising system that knows a lot about you and what you like and are likely to buy in particular. To do so, the system relies on a centralised database that collects information from ‘partners’. Any website can opt in to participate and act as a ‘partner’. To start, you first identify your website with a set of keywords that best describe your business. Each time a user visits any of your pages, you forward information to the centralised database. This would usually contain the IP address of your visitor, keywords associated with the page and statistics, such as average time on the website, referrer and others. The advertising system aggregates all this information and builds a comprehensive profile of keywords based on you, the ‘partner’, and pages your user visits. As the user goes from website to website, the system knows more about them and things they search for, read and chat about.

At this point it might not seem as a big improvement. Most platforms already do most of the above today. Let’s go a step further. Imagine a user browsing the web looking for pages with information on Halo 3 and XBox. We can safely assume the user is interested in purchasing an XBox console to play Halo 3, but we cannot tell for certain whether they already own one. Advertising ‘partners’ can submit this kind of information back to the centralised database. If a ‘partner’ is an on-line shop, upon each successful order a list of keywords that match the purchased products is submitted back to the database. At this point the advertising system can offer suggestions for Halo 3 expansion packs or XBox modifications, such as an HD player.

Using data-mining the aggregated information about a user can be sorted in a timeline which can later be analysed and used to provide up-to-date suggestions. If a user visits a page which contains the keywords ‘hd dvd’, but the user has been looking for ‘blue ray’ in the last week, we can assume they are interested in an HD player of the latter kind and serve the appropriate advert.

With this kind of system privacy must be at the top of the list of concerns. Harmful websites may request adverts for a particular IP and based on the results, build a profile for a given user at their first visit. For this, and many other reasons, no personal information should even be kept on record. Details such as first name, surname, date of birth and others should always be rejected by the system. A community, much like the one driving Wikipedia, should keep an eye for harmful websites and abusers. Registrations should be allowed only by referral. Limits on the amount of adverts served per ‘partner’ should be in place where appropriate.

More start-up companies are founded every day and they need a better source of income. Old companies are already making their best to optimise websites and make more sales, but there is only as much one can do. While contextual advertising might be hot today, I believe person-targeted advertising will soon replace it.

What really decides consumers to buy or not to buy is the content of your advertising, not its form.