SPSS – redefining the benchmark for an analytics suite

I have been involved with data mining and advanced analytics for most of the last twenty years and throughout that period the product that has set the benchmark by which others have been judged has been SAS. Today I think we should see SPSS in that role. With the latest release of their software they are offering what is probably the strongest combination of features across the whole spectrum of elements required to define a world-class analytics suite.

SPSS are now successfully addressing each of the areas that their competitors have targeted, be that the handling of missing and skewed data sets that was the forte of Fair Issac, the productivity gain of KXEN, or the integrated enterprise management of the analytical resource of SAS. SPSS addresses all of those areas but in addition offers a number that set it apart.

Firstly SPSS have a unique capability in the area of Attitudinal surveys, because they are the dominant force in the Market Research market. When it comes to building a true 360-degree view of a customer therefore, SPSS has a significant lead on the other solution providers. End user companies, who see a need to understand their customer behaviour, are now picking up on SPSS’s Dimensions solution. This was formerly seen as almost exclusively a Market Research agency solution. With the increasing need to understand markets, solicit customer feedback, and individualise experience as part the movement towards a managed customer experience SPSS offers capability here that are unmatched.

Secondly I would point to their openness and willingness to work with others. This is typified by their openness to embrace the in-database data mining capabilities of the big RDBMS vendors, Microsoft, IBM and Oracle. As much of the effort that goes into building a data mining model is spent on data handing, manipulation and transport, the biggest productivity gains are not going to come from better algorithms but through the elimination of as much of the need to handle data as possible. SPSS have enabled people to exploit the capabilities of the RDBMS vendor’s from their introduction, enabling existing skills based on the SPSS toolkit to transfer seamlessly into the in-database mining world. This is surely the way forward and SPSS have been at the forefront of this movement.

Automation and productivity are now key issues within analytics. As analytics has become a mainstream activity the requirement for new powerful and predictive models increases almost exponentially. It is also not just new models that are required; it is vital to maintain the predictive power of those models that have been deployed. The Predictive Enterprise Services enables the automation of the lifetime management of the analytical process.

With new tools such as the Binary Classifier, SPSS have made significant steps to help make highly productive, very predictive models available to as broad a spectrum of users as possible. It automates the search for the “best” models therefore making the analyst’s work more productive and effective.

As part of the explosion of importance of analytics has come a need to both be able to handle more data (scaling), and also more types of data (diversity). As the range of data expands so the types of analysis has to evolve. SPSS now offer many new forms of analysis such as time series analysis; very excitingly, SPSS are also starting to offer a capability to look at Voice. To date most mining is limited to structured numeric data, but in the last five years text mining has been seen to offer valuable new insights, but with so much interaction with customers now being channelled through call centres the real treasure trove is likely to be in voice mining. SPSS are also offering capability to analyse things like Blogs through text mining access to Web 2.0 data through RSS, which can further enrich the understanding of customer’s experience.

As well as these headline grabbing advances, SPSS are also looking at the more mundane, but equally significant, areas for enterprise deployment such as security; and they can now encrypt fields so that sensitive data can now be passed more freely with key fields protected from prying eyes whilst the benefits of the overall model can be widely disseminated.

In a global world one of the biggest issues is how to look at the market as one when language acts as such a big divide. SPSS have come up with an answer by producing a means to create a single base and to then analyse it as one. This is the sort of innovation that really distinguishes so much about SPSS.

Whilst others introduce many advanced features with value to small niche groups so many of the ideas that SPSS are introducing have tremendous applicability to the vast majority of users. The analysis that SPSS produces really focuses in on actionable analysis of relevance to addressing the problems that we come across very day. Over the last five years or so SPSS has emerged from behind the shadow of SAS to offer a real challenge, and now the challenger deserves to be recognised as the champion.