Edward de Bono, a noted expert on creative thinking, once said:
Business Analytics is very similar to this concept of science, and specifically, this quote applies to the concept of ‘advanced analytical’ models associated with predictive and descriptive analytics. Looking at the domain of Business Analytics in the EMA Business Intelligence Continuum, you can see that it is the middle component. Business Analytics builds on the data acquired from other platforms and often requires that data to be managed in some fashion.
I have used the EMA Business Analytics Pyramid to frame my comments. As you can see, the Business Analytics Pyramid builds from “unassuming” concepts such as “simple analytics” to the top of the pyramid and machine learning and natural language processing.
The following are my thoughts on the state of Business Analytics for 2013:
- It Is Not All About Predictive…: If you read about “advanced analytics”, you might just think that all analytics are predictive or machine learning. This is not the case… However, most of the ‘simple analytics’ that we are familiar with (i.e. COUNTs, SUMs, SORTs, GROUPBYs) are moving to dashboards as ‘right click’ features of other products. While you can still do your analytics with a 3D pie chart for top-ten customer percentage, soon that will be considered “old school” and using Naïve-Bayes to determine the likelihood of your potential customers to join that top-10 will be considered “nu skool”.
- Cubes Headed Out to Pasture: The wonderful OLAP cube always has been a cheat. Just as much as the spitball in baseball or holding in football. Cubes were designed and built because database technology couldn’t do its data day job AND all those aggregates. I firmly believe that between massively parallel processing, in-memory database and columnar storage technologies used separately or in various combinations, the OLAP cube will soon join the Rubik’s cube in the hall of stuff we used to use and now don’t remember why…
- Forecasting Is Not in a Slider: As we push more analytics into the category of self-service Business Intelligence, we need to give forecasting tools more imagination than a “slider”. This will only push folks back to the Excel spreadsheet where they have more control over variables and scenarios. However, forecasting can’t be so complex that the average user(s) cannot figure it out. Again, that will push more folks toward breaking with a central platform for the spreadmart.
- Advanced Analytics Are Here! As I said above, advanced analytics in the form of descriptive modeling, predictive analytics, and machine learning is here and in a form that business users can use. No longer are advanced analytics locked in a “vi editor” ( talk about old school… ) and a programming extension to the database engine. Advanced analytical features and products now have GUIs and wizards to help business analysts start down the path to data scientist-dom.
- More “Creepy” Will Accompany the “Cool”: Some of the best uses of analytics will be in those areas that sound “cool” until we think about how invasive they could be if used in the wrong hands. That’s the “creepy” factor – see Target and Benetton. This is where ethics and the effective use of anonymization techniques will help to avoid …. err… minimize the “creepy” factors.
What say the readers?
Do you really like and want to keep your OLAP cubes? Is predicting importance of ‘advanced analytics’ like hitting water if I fall out of a boat? Do I trust companies too much with the “creepy” factors associated with privacy?
Next week, I will cover Knowledge Delivery and the future of the 3D pie chart…. ( not that anyone would have strong feelings about that topic… ) I hope you continue the journey with me.
NOTE – For those unfamiliar with the song “88 Lines about 44 Women” by the Nails, I highly recommend you give it a try. At the very least, it was the inspiration for this series of blogs.