BI in Action

Sponsored by:

Become a BI in Action Sponsor June 20 & 21 2007!

• Agenda
• Sponsors
• Partners
• Blog
• About
• Contact Us

BI in Action Blog

October 05, 2007 Mortage Mess: Where Was the Business Intelligence? As you probably have seen in my blog postings and everywhere else, the era of Competing on Analytics is upon us. Businesses now have the tools and the drive to look at who, exactly, their customers are, what they want, where they are going, and if they will go amiss at some point in the future. Financial services have been the leading force in adoption of business analytics. However, such business intelligence seemed notably absent from the currently unfolding subprime mortgage debacle. Anyone following the business news over the past few weeks has seen the grisly stories of mortgage companies taking huge losses or going out of business. In addition, financial services firms with stakes in the market are also taking their share of mega-losses. For example, at the time this post was being written, one of the powerhouses of the financial services space, Merrill Lynch, announced $5 billion in losses thanks to investments in debt obligations tied to the floundering subprime mortgage market. Washington Mutual also took a 75% drop in profits as fallout from the Mortgage Meltdown summer. That's big, big money. However, if our business intelligence tools and platforms are so good, why didn't any of them see this coming? Could effective predictive analytical tools have helped steer these companies away from these losses? For example, couldn't these systems have predicted the impact that the current waves of mortgage-rate resets would have on certain markets? Why weren't these systems able to flag risky loan applications that would not be able to bear resets to higher rates? I am reminded of what happened to Cisco Systems during the dot-com and telecom busts in 2000 and 2001. Cisco, with the world's most sophisticated and intelligent supply chains -- able to see customer demand well into the future -- ended up with $2 billion worth of overstock, because it kept on producing and didn't see the market softening ahead. Our analytics systems may be better and better at spotting new opportunities and market anomalies, but they are only as perceptive as the executives that actually make the decisions from the reports they produce. When a mania overpowers markets, it overpowers both rational and programmed decision making. Companies can be cautious, but, at the same time, can't have their lunches eaten by nimbler competitors wiling to take greater risks. We have the systems that can make good tactical calls -- telling decision-makers whether customers are good or bad credit risks, if they will stay on for extended periods of time as customers, and the likelihood that they will respond to certain market promotions. We can tell what items will need to be shipped through a supply chain weeks before demand actually hits. The next challenge is developing systems that are outward-facing, that can take a broader look at the big picture, and the impact these external forces will have on opportunities. And, be able to look at the impact that new players have on the competitive landscape. with that, analytics that not only spit out reports, but make actionable recommendations. Posted by joemckendrick in Predictive Analytics  |  Permalink  | Comments (2)  | TrackBacks (0) May 21, 2007 Analytical Software and the Pre-Crime Division A couple of years back, in a movie called "Minority Report" (based on a book by Philip K. Dick), the Washington, DC police department had established a "Pre-Crime" division that was able to apprehend individuals before they committed (or even thought of committing) their heinous acts. Ethics and legalities aside, the movie actually made a fascinating use case for predictive analytics, except for the fact that at the core of the police department's system was a bizarre setup involving psychic mutants who were kept in a sensory deprivation chamber and had their brains hardwired into the computer system. This was how the original story was written, of course. But it would have been really cool and believable if the police relied entirely on pattern-matching and analytical software to develop their pre-crime scenarios. (That would have given the movie's main actor, Tom Cruise, a good reason to jump up and down on a couch or two!) Minority Report was supposed to be based in the year 2054. A new story in the New York Times, however, shows that we may be closer to this reality than we imagined, thanks to analytical technology. The NY Times relates how the Richmond, Virginia, police department had been able to deploy BI solutions to help bring its crime rate down -- 20% last year, more this year. The department employs software that culls through the department's data stores, such as 911 calls and police reports, and overlays other streams of data, such as neighborhood demographics, payday schedules, weather, traffic patterns and sports events — to predict where crimes might occur. For example, the NY Times relates, the technology "pointed to a high rate of robberies on paydays in Hispanic neighborhoods, where fewer people use banks and where customers leaving check-cashing stores were easy targets for robbers. Elsewhere, there were clusters of random-gunfire incidents at certain times of night. So extra police were deployed in those areas when crimes were predicted." Likewise, the NYT article describes how similar analytical software is helping retailers such as Wal-Mart Stores and Kohl’s use advanced computing and math to more accurately predict what sizes of clothes should go to what stores. Harrah’s and other casinos decipher slot-machine results to optimize customer traffic and profits. Stockholm and other cities use traffic data and patterns to determine “congestion pricing.” In the financial industry, Capital One and other banks mine all kinds of transaction data to identify, and stop, fraudulent transactions. Whirlpool, which sells 25,000 washing machines a day, also now automatically scans warranty reports as well as manufacturing, supplier, sales and service data to try to further trim its warranty costs and improve quality. The company says it has trimmed by 30 to 90 days the time required to detect and fix parts or manufacturing problems that cause defects. Similar technology is being put to work for more mundane applications as well, as as tracking employee productivity within enterprises. Tracking e-mail traffic, instant messages and other digital communications — stripped of personally identifiable information — is helping companies understand how work and ideas flow through their internal social networks, NYT reports. And, thankfully, no mutant psychics were harmed in the making of these implementations. Posted by joemckendrick in Predictive Analytics  |  Permalink  | Comments (1)  | TrackBacks (0)