Like most retailers, supermarket conglomerate Ahold has a customer loyalty card program, and like most retailers, it mines that data as much as it can to help it sell more products. The data it has gathered over the years range far beyond what a certain customer buys every week. It can also determine, for instance, a customer’s price sensitivity to a certain item — that is, how much he or she is willing to spend on it.
Executives at Ahold started wondering if they could push their usage of this data even further to see which customers are most likely to buy if offered a coupon, and even why customers come into the store in the first place. “If your customers are shopping with you because you sell a certain brand of beer, it is important to know,” said Ellen Rubin, vice president of marketing for Netezza, a data warehouse appliance vendor and one of Ahold’s providers.
Deciphering the Data
It is not easy teasing out this data — developing a profile of a customer, for example, that is most likely to buy diapers and then beer when it is on sale — she told CRM Buyer.
“Retailers have data pouring in from the point of sales. So most will look for summary information and ignore raw detail. It is too hard or expensive to analyze years worth of data,” Rubin noted.
This is changing as BI applications are becoming easier and faster to use. In Netezza’s case, the company’s systems are based on an integrated data warehouse client and most of the processing is being done at the disc level, where the data resides.
By contrast, a traditional data warehouse system maintains data on storage systems, Rubin explained, and when a user runs a query that asks which customers that bought beer also bought diapers the system must access many sources for the answer. “The data must come off storage, move around and essentially get processed by a cluster of expensive servers before the piece of information you really need is handed back to the customer,” she said.
Variety of Industries Taking Note
In many ways Ahold’s story is almost commonplace now: a commercial entity wants to know something to help it make more money. It turns to a BI vendor for the answer. Now, though, as it becomes clear to industries outside of the normal realm of BI users how much this technology has changed, they too are adopting these platforms and using them to solve specific problems.
Education and the medical fields for example, have looked to the corporate world to see what they have accomplished with BI technology and then replicated it in their own environments, Christina McKeon, SAS’ BI program manager told CRM Buyer.
One researcher at the University of Louisville that had been exploring practice patterns among doctors in area hospitals turned to SAS for help, according to McKeon. Dr. Patricia Cerrito and her team of researchers began using SAS Text Miner to find cost savings as well as improve patient care by analyzing data from hospital billings, medication orders and physicians’ chart notes.
Cerrito’s work became very personal when her husband underwent open heart surgery and his surgical team made medical decisions based on her research. When she learned her husband needed surgery, she analyzed post-surgical outcomes of diabetic patients who underwent open-heart surgery and were prescribed certain antibiotics.
“She wanted to see which medications led to longer hospitalizations in these circumstances,” McKeon said. According to SAS, Cerrito has cited the applications’ data mining algorithms and the interface for managing and importing data as key to her work, as well as the way SAS integrates its text mining capabilities into its data mining application.
Other unusual deployments include a college using student data to see which ones might be at risk of dropping out. Also, a state in Germany is using BI to develop applications that map diseases in animal populations to track potential hazards to humans.
“BI has always been about getting this technology in the hands of as many nontechnical users as possible in the corporate setting,” McKeon commented. “Now that concept has spread to other, non-corporate settings.”
Next Gen Deployments for Corporates
Companies haven’t finished exploring all the ways data mining and business intelligence can help them make money.
One area that is largely still uncharted is the use of Web site data and its incorporation into a general CRM/BI strategy, Colin Shearer, senior vice president of market strategy for SSPS told CRM Buyer.
“Typically Web data has been carried separately from customer data. Yes, it includes information about customers, but it is so low level and difficult to work with, few companies have been able to use it in an extensive customer analysis,” he said.
This is one area in which SSPS has been working, Shearer declared. “We are beginning to see the emergence of techniques in which BI is able to translate these low level messy piles of Web data and then feed that data into the general mix as well.”