Predictive and precarious

So the news breaks that SPSS has pulled off a marriage with an IT industry behemoth. While many have speculated that SPSS was hoping to get bought by Microsoft, the big surprise is it’s the sober suits at IBM that have carried off the bride. Now, the real reason for the big name change is clear: the due diligence had surely shown there was an unresolved court claim by SPSS founder over the continuing use of the SPSS name. The remedy – dump the name. Quick.

Remember, we went from catchy, resonant names like SPSS, Clementine and Dimensions to whole sentences so little adjusted to memorability that you needed to go on a one-day course to learn how not to confuse PASW Reports for Surveys with PASW Reports for Surveys (Server), or, come to that, PASW Web Reports for Surveys and PASW Reports Professional for Surveys. Just try, for a moment, saying ‘predictive analytics software reports profesional for surveys’. Like me, you’ve probably already forgotten why you even came into the room, let alone started reading this article.

Where was I going..? Oh yes, IBM acquires SPSS for its predictive analytics capability. Research magazine gets an interview with Jack Noonan, CEO of SPSS. He manages to get Predictive Analytics into the first sentence. The problem is, this does not have a whole lot to do with market research. Sure, market research can and should play a role in any predictive strategy. Forecasting is an important function for MR, but it is not its only role. Indeed, as SPSS do rightly remind us, some of the best predictions are to be made by analysing actual transactional data – real, recent behaviour – rather than asking people what they might do. There is still too much research activity expended on asking people what they did, when there is better data to tell you precisely what they actually did.

Unfortunately, what SPSS means by ‘predictive analytics’ has very little to do with market research as it is practised. The press release, published earlier today, focused on the synergies between the SPSS product line and IBM – in business analytics, information mining and predictive models based on very large amounts of business intelligence data for the corporate enterprise client.

Dimensions, by old or new name was not mentioned. However, neither did any of these words appear in the substantial 1700-word press release: ‘survey’, ‘questionnaire’, ‘market research’, ‘question’, ‘answer’. Neither did ‘social research’, not even ‘research’ at all. Even ‘statistics’ only put in a solitary appearance in the “About SPSS” paragraph. It seems ‘data collection’ is the new euphemism for what the MR profession does. The entire announcement had nothing to say to market research – a market which appeared to account for around a fifth of SPSS’s revenues. SPSS were – and still are (just) – a major player in the market research software market.

So what is the future likely to hold for the Dimensions products and the market research customer? They were already becoming marginalised within SPSS, with very little remaining expertise in market research within the company. For IBM, running a portfolio of market research products is likely to be of very little interest. It is likely the software will be unbundled and disinvested – sold on to another developer – or it will be folded more completely into the IBM product family with its MR edges beaten out. There won’t be many MR customers of the Dimensions platform that are revelling in this news, unless they happened to own a whole bunch of SPSS’s NASDAQ stocks.