Content Copyright © 2009 Bloor. All Rights Reserved.
This blog was originally posted under: The Norfolk Punt
IBM has just launched its London Analytics Solution Centre at its London South Bank offices. This is intended to help IBM customers introduce “fact based decision making” into their organisations, using IBM’s advanced analytics tools and sharing IBM’s experience. This is a Good Thing, in my book at least, but with some caveats (see later).
It’s a good thing, for a start, because it shows that there is real technology to back up the “smarter planet” initiatives IBM talks about—and similar technology is presumably appearing from other players too. We saw an interesting “operational risk” management project in which competing banks are collaborating because no one bank has sufficient data for the analysis required. This is for the not-for-profit Operational Riskdata eXchange Association (ORX), based in Zurich, and attempts to quantify the banks’ exposure to possible losses from a variety of events such as credit fraud or payment default, so they can minimise their contingency reserves while still complying with Basel II.
Another Smarter Planet project is the Smart Bay water management project, a next-generation coastal and marine monitoring and management application being built in conjunction with The Marine Institute (Ireland). This is basically an information portal (a tourist might use it to decide which Galway beach to visit) driven both by real time water sensor technology and predictive modelling.
Or there’s Deep Thunder, which is not really a weather forecasting system but more of a weather impact management system. Based on forecasting, it manages the business processes affected by the weather and the proactive distribution of resources to deal with the impact of the predicted weather events. Predicting events such as the recent Cumbrian floods is hard, but technology can ensure that flood control technology is available where it is most likely to be needed. Moreover, unusual crises can be simulated and contingency plans prepared so that, with real-time data feeds and still using the Cumbrian floods as an example, the Army could have appeared to cut a deep flood diversion channels down main streets etc. as the situation was developing, instead of deploying emergency bridges etc. after it was all over.
Plenty of other examples were available, from smart policing (Chief Constable Patrick Searer, President of the Association of Chief Police Officers of Scotland, spoke; and is clearly aware both of the opportunities provided by smart analytics and the necessity of managing the associated social outcomes) to modelling the power distribution for electric cars. So, Smarter Planet technology is real, but, and now for the caveats, I think it is important to remember that it is a disruptive technology with profound cultural and management implications.
Smarter technology might modify behaviour in unexpected ways, for instance. Mass behaviour can be predicted—but not individual behaviours. For example, highlighting polluted beaches with Smart Bay might actually encourage a significant minority of people to visit them (if you don’t like crowds, hate children, like scenic views and don’t want to swim or sunbathe, a polluted beach might be ideal for you). Of course, if the smart application works as it should, such unexpected behaviours will be spotted and any issues they cause addressed—but one can imagine “smart technology” being implemented inflexibly. Smart technology implies smart implementation—and smart people who can interpret and build on the smarts.
What I found interesting about the smart solutions presented at the launch was the way in which actual real-time data feeds were integrated with the predictive modelling. This should help with resilience and ensure that model predictions change in the light of reality. Weather systems are chaotic, for example, and local features (such as updraught “plumes”) limit predictive model resolution. Nevertheless, Deep Thunder can work at 1 km resolution partly because its predictive models and analytics are complemented with real-time weather feeds—the 21st century equivalent equivalent of throwing a barometer indicating fair weather on a wet day out into the rain to see for itself.
However, all this doesn’t address the human implications of smart technology: fact-based decision making is disruptive (often in a good way) and requires a high-maturity organisation to cope with it. Unfortunately. Lord Mandelson, who opened the centre in the morning and extolled the importance of fact-based decision making, wasn’t around in the afternoon Q&A session, to answer questions about why his government sacked scientists who pointed out that scientific facts weren’t precisely in accord with decisions it was making about recreational drugs categorisation.
There is nothing intrinsically wrong with making a decision not in accord with the apparent facts—perhaps religious or moral considerations take precedence; or perhaps you are measuring the wrong facts. Nevertheless “fact-based decision making” implies a certain transparency, with the facts available to everyone affected, thus enabling them to question decisions apparently not in accord with the facts. Many current decision makers seem to be uncomfortable with this.
Why would anyone not want fact-based decision making? Well, if your power-base comes from your patronage and your ability to slew decisions in favour of particular people or interests; or if you aren’t confident that your decisions are always good enough to stand up against the facts; then you may not be so keen on smart fact-based decision support systems.