I recently got an overview of InfoCentricity and their analytic product, Xeno. InfoCentricity is an analytic solutions company focused on helping customers find the insights buried in their data. They see a gap between BI-like analytic tools and data mining/analytic tools like SAS or S where you need to be a programmer. Their objective is to provide ease of use (so that programming is not required) but to also deliver analytic power for analysts. The company was founded in 2000 and has a strong historical tie with the retail banking industry and to credit scoring. They offer Xeno and analytic services that are mostly in support of Xeno but they also provide analytic services such as developing custom models.
Xeno is their analytic workbench. It’s a nice product with a thin-client metaphor. There is no software install/maintenance as it is completely web-based and it is designed to build better models, faster by supporting rapid iteration. It is data agnostic and designed to support a collaborative, team-based modeling process. It supports the usual predictive modeling techniques, some unique ones like “multiple outcome optimization”, trees, clustering and reporting/analysis. It is aimed at modelers but they do have some non-modelers using it and they are working on extending it to develop an easy to use tool for non modelers e.g. in retail/catalog. As they add this they plan to continue the collaboration focus so that Xeno supports not just collaboration between modelers but also collaboration between models and non-modelers. The tool is not focused on data preparation – most of their customers use SAS for this I suspect – but is aimed squarely at a collaborative modeling environment.
The product has a nice, web based interface. It provides quick analysis of imported data and a wide array of reports. The interface is very much a point-and-click one with lots of automation including automated classing and binning and predictive ranking. The idea of the automation is to get you most of the way while supporting manual fine tuning. The tool stores versions, shows the steps in the process and manages the development and evolution of models nicely. They have a tree builder and a clustering tool and these tools provide guidance to make sure clusters, for instance, are moving in the right direction. While aimed at modelers, the tool was approachable and open without any of the opaque modeling scripting one sometimes sees.
From a company perspective they think their differentiators are analytic brain power, a commitment to knowledge transfer, innovation, collaboration, data flexibility and ease of use. Their focus areas are risk management (credit, collections) and marketing (response, revenue, churn, segmentation).