Decision support is an old-fashioned term for BI. However, when it comes to making strategic decisions (should we acquire so-and-so? What R&D strategies should we adopt?) conventional BI doesn't really help. On the other hand, some software back-up in helping to view options, assess risks and estimate potential rewards is a seriously good idea. Of course, you can hire consultants (expensively) for a one-off strategic decision but what about all the other strategic decisions that you have to make? More expensive consultants? Or, of course, you can model everything in Excel, which is fine as far as it goes, except that it was never designed for this sort of task and it doesn't go far enough.
Or you could go look at SmithBayes (named after Adam Smith and Thomas Bayes). This company was spun-off from Mclaren a couple of years ago. Now, you may not know this but the leading Formula 1 racing teams model different race scenarios prior to each race and then decide on relevant strategies in advance. Then, during racing, these scenarios have to be continually updated and revised so that appropriate decisions can be made based on calculated risk/rewards relative to the various options that are available. For example, should the driver go all out for a win or settle for 3rd place? Of course, they don't always get these decisions right but the point is that the computer models provide them with the information to make informed decisions.
Anyway, that's what the SmithBayes Platform was originally designed to do but, since it's spin-off, it has been targeted at strategic business decisions. One point I find particularly interesting with respect to the product's background is that we tend to think about strategic decisions as being long-term things but, as the Formula 1 example proves, strategies can be very short term in appropriate circumstances. Thus, SmithBayes has customers in the Aerospace sector, for example, using the company's platform to support new product development in areas where you are talking about a decades of development and deployment yet, on the other hand, it has customers in FMCG where an issue might be one of how best to respond to competitive marketing campaigns. In the latter case, the customer is now able make appropriate responses to such threats in a week when previously it took three to four months.
As far as the product itself is concerned, it was originally written in Delphi but version 2.0, which is released this month, has been completely re-written for Windows. It provides a highly graphical environment, using drag-and-drop techniques against so-called ‘smart objects’ that you use to build your model. These objects may drive from either internal sources (databases, spreadsheets and so on) or, in the latest release, externally (competitive actions, government regulations, industry forecasts and so forth). All of these are treated as time series so that they are subject to change. They also have associated quality indicators, which indicate how reliable these objects are: are they historical facts, forecasts from a trusted third party, a finger in the wind or whatever? Objects with different quality indicators are treated differently. The model itself re-calibrates over time (something which is very difficult to do with Excel) and it will be automatically updated if a constituent object changes (say, an updated forecast is published), along with the raising of relevant alerts.
The company only has offices in London at present though it already has one US-based customer and is talking to several more potential clients: if it succeeds in winning some or all of these then it is likely that it will expand into the States. As I think this is a cool product for which there is no real competition I expect SmithBayes to have to open in the US sooner rather than later.