Back when I studied Statistics as the selected additional option of my Maths ‘A’ Level exam, the practical use of what I was learning was somewhat limited – most often the preserve of academics, political analysts, researchers, surveyors etc. Increasingly, however, statistics are being used in business….and ever more widely.
Several years ago, IBM acquired SPSS, which resurrected my interest in the use of statistics. SPSS (originally, Statistical Package for the Social Sciences) has a long history, dating back t0 the ’60s. These days SPSS is likely to be utilised by a
variety of business users for statistical analysis, data and text mining, predictive modeling and decision optimization to help anticipate change and take action to improve outcomes.
I’m most interested these days in how it helps in various aspects of forecasting, and consequently in decision-making. A couple of examples:
1. Support for high level assumptions. One of the challenges in planning and forecasting is deriving the initial assumptions. These are often based on judgement or “gut-feel”, but in today’s ever-more complex world, this is not always the best way. The allocation of resources in a world suffering this on-going recession is a critical challenge, and thus organisations are at risk of commiting resources inefficiently if they don’t understand their environment. This can be where a statistical approach can help. I’ve discussed previously how we run Best Practices in Forecasting round-tables, but one of the changes to the audience in recent years has been how many more public-sector attendees and charities. In talking to these attendees, it becomes clear that they are under increasing pressure to allocate their scarce resources on a more logical basis, taking statistical approach to predicting population profile and changes etc. There are some great examples – Memphis Police Department, Medway Youth Trust in the UK, City of Nuremberg among others.
2. Cross-selling. One of the key areas that we highlight in forecasting best-practices is a driver-based approach. In other words using the language of the Business rather than that of Finance. An example that we frequently use is a Call Centre manager. If asked just for a revenue forecast, he/ she will perhaps make a judgement call based on a previous period. If, however, the forecast takes a driver-based approach, then that same manager may be enabled to look at number of operators –> number of calls made –> leads created. Those leads created can then be used as a driver, via average sales per lead, to create an opportunity pipeline. This, in turn, will drive revenue via a Sales conversion rate.
This approach gives visibility to a number of decision-making opportunities, for example if opportunity level appears to be dropping, then the company may want to look at means of increasing the average sales per lead. SPSS can help in this by, for example, helping to identify cross-sell or up-sell potential. Again, there are a number of examples – take a look at Telenet, which transformed its customer care call center into a sales outlet by exactly this approach. You can actually view an webinar of IBM Cognos TM1 and SPSS being used together for this purpose in IBM’s Innovation Centre.
3. Run-rate forecasting. Successful implementation of rolling forecasts, will rely on a number of factors. Among these will be an ability to focus forecasting resource on key drivers/ products/ cost areas whilst monitoring less critical areas, perhaps us
ing a basis such as that shown in the diagram here. Some of the less-volatile items, though, may be valuable/ material to the business, and thus deserve some significant focus. This may, for example, be the case where valuable inventory items are treated as “run-rate”. We’ve recently seen this challenge with a number of customers who take a linear extrapolation approach to forecasting run-rate items, and are looking to us for help in how to do so in a more robust way than utilising spreadsheet functionality. SPSS, of course, performs this task incredibly easily, and integrated with IBM Cognos TM1 as part of a reporting, analysis, forecast and planning solution, delivers tremendous value to a business.
It’s great to see such beneficial uses for an area of my education that I feared would be rarely used!







