This week, Sears surprisingly becomes an analytics SaaS vendor, and we look more closely at how the Obama campaign used analytics to gain an edge in the recent Presidential election with what one analyst calls “Small Data.” Here is my business analytics news roundup for the week of 11/12/2012.
Sears is getting into the tech startup-business with MetaScale – a BigDataAsAService firm and in the process, trying out-Amazon Amazon Web Services.
Wow. I didn’t see this coming. I’ve written recently about how retailers are helping to drive the big data locomotive, using Hadoop, TereData, and other massively parallel data processing frameworks to analyze customer transactions in near real time. But making the switch to offering those in-house capabilities as service….well, that’s bold. The good news is that Sears seems to understand that the analytical and deployment – know-how is probably more important than the Hadoop infrastructure itself. I’m going to watch this one closely.
Tony Byrne from Information Week gives a few insights on *how* the Obama campaign team used analytics to take real action at the individual voter level – and reminds us that analytic value lies in the action we choose to take, not on the size of our datasets.
I’ve been pretty vocal in my criticism of the Big Data movement – that software vendors try to convince us that we should capture and process every aspect of every transaction our business executes, then keep it forever. (I call this “BigData, Big B”) I believe that when we focus on products like Hadoop, we forget that our goal is to understand and impact individual customers and employees, and often, we can do that with sound analysis of properly groomed, much smaller datasets. Those datasets can still be very very large, like the Obama campaigns voter logs or a Fortune 500 company’s accounting journals, but those data sets are not measured in Petabytes. I call sets like this “bigdata, little b,” and believe that they offer far more value to companies interested in data-driven decision making than the kinds of data sets that get press these days.
And even more importantly, as Mr. Byrne points out, once we’ve interacted with a customer, we learn something about them – and we should update our models and take action. That’s what the Obama team did, and that’s what most of you will do.
IBM has invested $16 Billion over five years in shoring up their business analytics offerings.
Last week, I noted some researchers’ assertions that analytics would be a $50 Billion per year industry by 2016. IBM seems to agree.
IBM’s multi-year investment includes their purchase of Cognos, now IBM’s flagship business intelligence software offering, SPSS (a long-championed statistics package), several other firm acquisitions, and assumedly their media blitz touting their predictive analytics business services practice.
And IBM isn’t alone. Over the last several years, as I noted in this post, SAP has purchased BusinessObjects to shore up their business intelligence offerings and launched an R-based statistics package, and SAS seems to be everywhere nowadays.