

Systems of intelligence (SoI), the most complex class of big data applications, apply machine learning to anticipating and analyzing human interactions. However, as yet no end-to-end packaged applications the likes of the ERP model have appeared. In their absence, writes Wikibon Lead Big Data and Analytics Analyst George Gilbert, ad-tech applications provide an architype that customers can use to develop their own. Advertisers, ad agencies, consumers and sellers (a.k.a. publishers) – the advertising marketing ecosystem – use them to collaborate in real time to optimize customer messaging. In ad tech:
These are highly interactive systems, not pure batch mode Hadoop applications. They allow both buyers (including advertisers and ad agencies) and sellers (such as Google and Facebook) to track and collect rich information about millions of individuals across the Web. In 100 milliseconds, while a Web page loads on the individual’s screen, they bid on and sell access and deliver ads tailored to the individual’s interest.
Ad tech disintermediates the advertising sale, removing middlemen that could absorb as much as two-thirds of the price of each ad and isolate the sellers from buyers. Behind the scenes, e-commerce applications must either manage billing and other accounting functions directly or integrate with the participants’ systems of record to support those functions. Publishers have to manage their full ad inventory across all digital distribution channels.
These systems have been in operation – and are constantly evolving – for several years. They are behind the ads that pop up when an individual researches a product and then sometimes follow that individual across the Internet for days. And, writes Gilbert, they provide a model that companies in other areas can use to develop their own machine intelligence-based systems to automate interactions with their customers.
photo credit: Photon via photopin cc
THANK YOU