The article from the BBC tech-business sector gives an overview of general trends within the evolving online advertising industry: specifically what programmatic advertising is and how its shifted the old pattern of demographic targeting to data-mining driven programmatic advertising.
Programmatic advertising takes advantage of the fractional seconds between website access and the site being displayed: creating a market where advertising compete in flash auctions for advertising space. This is in part why ads are different for different users. Firms utilized information that is passively collected through web-interaction (such as shopping or browsing history or Facebook profiles) as well as ‘key words’ in order to more accurately target ads to potential customers. As computer power increases these algorithms have markedly evolved, shifting the patterns of advertisement behavior.
In isolation real time interaction and data mining have few applications for advertising. However combining the two has yielded a fundamental shift in how e-marketing is done. Firms like Facebook and Amazon both aggregate and analyze information provided passively by their users, distilling personality of the individual into a profile. For example: person X likes football, ESPN, and recently posted on Facebook that they are going to a Ravens game. Facebook can sell this information to advertisers who then bid on the ability to present their ad. The difference here is that now the accuracy with which adds are placed increases exponentially as more information is gathered and analyzed. This has created a more power-driven dynamic within the e-advertising industry with information providers playing the most crucial role rather than the product itself. Instead of showing the product available to possible interested parties, interested parties are identified and targeted and the product found later.
While current e-advertising is clearly more sophisticated and more accurate than outdated traditional physical means, it is far from perfect. The wide range of disjointed information passively collected is not always useful for advertisers. The ‘key word’ problem stands in the way between e-commerce and higher levels of e-market efficiency. Without context information is often misinterpreted or pertinent information discarded. While the ‘key word’ method is perhaps the best available it is only the best in the same way that using a flamethrower is better than a match for starting a fire. The best example of this is customers being advertised similar or identical products to the ones already purchased. The information from the purchase (and lead up internet activity to the purchase) is still factored in to the programmatic advertising creating redundant and annoying ads.
The article overlooks one major factor: the constantly increasing ability to collect, transfer, and analyze data. The current issue, which boils down to a lack of full individual-personal integration into the digital world and the lack of information that entails, will likely be resolved as the next generation comes into its own alongside more personalized (wearable) tech which will provide the information currently lacking. Whatever new piece of tech becomes dominant will eventually likely serve as conduit for the customer and replace firms like Facebook in providing information to advertisers.