To quote Steve LaValle from an IBM Global Business Services Executive Report, "in top-performing organisations, analytics have replaced intuition as the best way to answer questions about what markets to pursue, how to configure and price offerings, and how to identify where operations can be made more efficient in response to cost and environmental constraints".
I'm sure (or, at least, I hope) that's true for the top performers, but it leaves me thinking that there's probably a host of "adequate" performers (and in a world of change where lots of companies go bust, "adequate" ain't so bad) who find being top performers too difficult and get by on intuition. Nevertheless, relying on intuition is a lot more risky than relying on "fact-based decisions" and implementing a "fact based decision-making" process is hard mainly because of the vested interests of managers whose power and reputation relies on their intuition—or, possibly, their ability to bestow patronage on their cronies. This is really a situation where achieving adequacy in competition with other players with fairly dysfunctional decision-making processes isn't something to aspire to.
IBM suggests that companies that wish to become top performers need to take analytics (from Cognos and others) to the next stage, by adding Optimisation—which, of course, ties in with its acquisition of ILOG, which brought an extremely effective optimisation engine to IBM (in addition to the Business Rules ILOG is perhaps better known for). However, regardless of IBM Global Services marketing, I do find the idea of optimisation as complementing analytics very persuasive. Analytics are all very well but they are of little use unless they are used as a basis for real, operational, business process improvement—which is where optimisation has a big part to play.
IBM brought a lot of analysts together in its Berlin Analytics Solution Centre (there are other solution centres in Tokyo, Beijing, New York, Dallas, Washington DC and London), last month to present customer stories and demonstrate its ability to partner with customers to help them develop their own business analytics and optimisation solutions rather than to just sell them product licenses.
One highlight of the day for me was Brenda Dietrich (IBM Fellow and VP, Business Analytics and Mathematical Science, Research Scientist: Math Science). It is good to be reassured that there is genuine intellectual capital behind analytics and optimisation and Dietrich is a genuine Operations Research expert and rallied to any questions exceedingly well. She also highlighted the fact that if this stuff has areas still to address, they probably lie in the behavioural areas, around incentives and personality issues, where scenario visualisations have a big part to play. Perhaps the most interesting aspect of analytics research today is into the automation of the process of asking the right questions—which is the hard part of analytics in practice.
Dr Ajaz Hussain (VP Next Generation Product Assessment) from Philip Morris International R & D told us about a collaboration with IBM. Over the next 5 years PMI plans to develop new products which may have the potential to reduce the risk of tobacco-related diseases. With IBM, it is working on methods for assessment of product risk combining traditional risk assessment with new approaches based on 'systems biology'. There's a huge amount of published information about the risk associated with tobacco products and the analytics on these data sets are extremely complex and will require some high performance computing power, so PMI and IBM are working together on this challenge. If successful, this novel approach for risk assessment could have applications in many other industries.
I took a certain personal interest in a presentation from Angus Cameron (Associate Partner, UKI Communications Sector Business Analytics and Optimisation). He described the use of IBM Content Analytics for the semantic analysis of "news assets" for producing indexes, highlighting particular interest points etc. (there's a description of the technology here). It's how the BBC made its online World Cup site agile and responsive. However, while I can see how this could help a good journalist write a deeply insightful article more easily, because he or she doesn't have to bother with the hack-work of finding material related to a particular issue, I can also see how it could be used to "dumb down" journalism by automating the production of celebrity-based coverage from unskilled (or amateur) news gatherers. This highlights the real business analytics and optimisation issue: asking the right questions inn order to deliver the right business outcomes. As Angus says, we must always concentrate on what analytics does for the users, on the actual business outcomes achieved.
And that reminds me of another issue. Business analytics and optimisation isn't just for conventional business decision-making. It can also be applied to all the operational information from IT operations and the data centre—in other words, we should be treating IT decision-making as just another part of business decision-making generally—after all, these days software, to a large degree, is the business. IBM GBS is developing holistic vertical business analytics and optimisation solutions based on a wide range of its products (not just Cognos and Ilog); I hope that its Rational division is doing something similar (including Cognos and ILOG) for "software econometrics".