Business Intelligence (BI), especially related to Big Data is a process of identifying interesting and actionable structure in historical data.  I discussed the importance of context in my earlier post “Big Data and Why Context is King” and real-time data is a way to infer the context of the situation.  Operational Intelligence (OI) deals with the situation where tactics are adjusted in response to real-time information.  The ability to collect more and more data in real time through mobile devices and emerging technology has opened the door to many more opportunities to optimize OI, not only online but in a combination of online/offline.

Over the years, I have worked on a series of evolving applications of OI to online and portable device applications.  We all tend to dwell on what we can’t do now instead of what we have accomplished.  We have come a long way in providing what people want when they need it.  The following are examples of that evolution.

E-tour utilized a large database of browsing behavior to identify common structures in who liked certain groups of websites.  This Business Intelligence set up a framework to determine which sites to deliver to a particular person and at a particular time.  Then, by monitoring the behavior of the person when receiving those websites, Operational Intelligence would determine changes to the specific websites to be delivered and adjusted on the fly to adjust the consumer experience.  Waiting for the backend data to be updated and algorithms revised to result in the same website delivery allows mistakes to persist.

I2go allowed users to answer a set of questions about the type of information they liked to listen to and develop a customized playlist.  By monitoring which tracks a user listened all the way through and those that they skipped, the system could adapt the playlist.  Additionally, random tracks were selected to attempt to further optimize the playlist.

SearchIgnite utilized historical click data to identify the structure of the relative value of a paid search position and then identified changes in the behavior of individuals from the norm that put a higher value on certain positions driven by their personal preferences. This allowed for dynamic changes in value determination.  This is especially true  when dealing with budget allocations across multiple marketing activities, such as collaborative TV and online spend.

BlinQ Media ran tests on micro targeting ads on Facebook to identify structure in who was likely to respond.  Then adaptive micro targeting was performed to drive consumers to external events.

With CloudTags, we are now utilizing who you are online (browse, buy, chat, like) with who you are in the store to deliver valued actions when you need them. The more our algorithms improve in their ability to anticipate customer needs and wants, the higher the creep factor becomes – which I will discuss in a subsequent post.

Ultimately, the combination of what both the general population and you have done in the past, along with monitoring what you are doing in the present, allows for micro-targeting to encourage specific consumer behavior. So while connecting online and offline in this way may sound initially difficult or complex, it’s really just an evolution of the types of decision science I’ve been working on in the online space for the last decade or so.