More Data Makes Your Business Grow

The job of IT departments can be concisely described as having two parts: managing data and advising business people about how the data could be used. As a general rule, there's little question as to what data might actually be involved in either of those functions. But the exceptions to that rule tend to be both intellectually stimulating and economically important. In surprisingly many cases, the use of new data sources can provide a huge boost to business profitability and success.

Examples abound in both the transactional and analytic arenas. On the transactional side, some of the biggest opportunities lie in the tracking of products and other physical objects via radio frequency identification (RFID) or, in some cases, more active mobile devices. Indeed, if you're in an industry such as retail, distribution or transportation, that's probably a top-of-mind issue for you and a major part of your company's medium-range capital budgeting. Also, companies in more and more industries are developing miniature commodity-trading desks and bringing in investment transaction data to support them.

Less obvious, yet potentially even more important, are the possible sources of new analytic data. There's data that's already available for you to collect, data that you can buy and entirely new data that you would have to create. There's conventionally structured data, unconventionally structured data and data that's barely structured at all. The possibilities are varied enough that if you don't take the time to think them through, you may well miss a company-changing opportunity.

In some cases, you just have to notice data that has already fallen into your lap. Search-engine logs tell you of customers' questions and interests in their own words. General Web-visitor logs give you similar insight. You may have a lot of customer satisfaction and product-quality data sitting around to be text-mined from warranty claims, call center reports and the like. And if a solution could be found to the privacy issues, even more information could be gleaned by voice-mining actual telephone conversations.

In other cases, you can obtain valuable data from third parties. The best-known example is consumer data from credit reporting companies, which can be used for a variety of CRM purposes; many other kinds of data can be used similarly. Away from CRM, medical researchers want to look across data banks of patient records to develop new treatment insights without the cost, delay or danger of conventional clinical trials. (The privacy problems around this kind of research can and will be solved soon.)

But the really mind-blowing possibilities arise when enterprises deliberately set out to create and capture data for the primary purpose of using it analytically. Here are some examples:

Loyalty cards, especially in gaming. The casino industry has been transformed by those cards you use to tell the casino what you're doing and to collect rewards. Both when you're on the premises and when you're home, casinos market to you very precisely based on that information. Of course, this involves massive data mining, but a huge fraction of the casinos' profits comes from it.

Location-based analytics. There's something Big Brotherish about supermarket shelves that know who you are and make offers accordingly, but that technology is being tested and deployed today. Wide use of RFID will greatly expand its scope. Privacy concerns do need to be overcome, but experience shows that consumers can be bribed into giving permission for this type of effort in return for personally targeted marketing offers.

Extra customer feedback. Smart companies should and do knock themselves out to get extra feedback to use in CRM and product quality analysis alike. Here are some ideas for getting that feedback: extra incentives for submitting warranty/registration cards; online surveys with prizes/bribes for participating; outbound phone calls to customers; forums and other community-building efforts; and better customer service of any kind (online or over the phone), inducing customers to consume more of it and hence communicate better.

Price/offer testing. Marketers have long been disciplined to test multiple product prices and offers to see what is most successful. Analytics in support of these tests make the testing more valuable. You can't estimate demand elasticity if you make offers at only one price point.

These examples are concentrated in CRM and product quality for a good reason -- those are the main areas of business where statistical analysis flourishes. As the scope of predictive analytics expands, the opportunities for profitable data-creation strategies will do so as well. For more on this subject, please go online to

Curt A. Monash is a consultant in Acton, Mass, and also blogs regularly on You can reach him at

5 collaboration tools that enhance Microsoft Office
Shop Tech Products at Amazon