Miglautsch Marketing has pioneered automated modeling since the early 90's. Many of our clients produce four or more models per month, we have built more customer models than virtually any other direct marketing company. MMI is considered one of the world's leading modeling companies. With clients such as Guitar Center, Adobe The Bay and dozens of catalog companies, our systems have proven themselves in the real world of direct mail marketing for some of the biggest mailers in North America.  In the past 6 years we won 11 blind tests against 8 of the world's best modelers.  Click->John's Welcome to hear more.  Visit John's personal site as well

Integrity as a Modeling Variable

The direct marketing magazines are loaded with articles on the benefits of better understanding your customers.  Especially in difficult economic times, these suggestions get paraded out.   Rarely is much detail offered, especially about how many or what kinds of customer variables are most beneficial.  Usually geodemographic clusters are alluded to, sometimes specific demographics and most talk about Recency, Frequency and Monetary (RFM).  My experience suggests that only another two or three hundred are important in getting a complete picture of your customers' behavior.  Even then, they never seem to quite match what you're looking for.  You might guess that soccer moms probably drive SUV's since lots of them do in your neighborhood... but is it worth getting vehicle type when it will only match on 5% of your customers.  I never cease to be amazed at how difficult it is to get just what you're looking for.  Read More...

Werewolf Variable

Miglautsch Marketing has now created the "Werewolf Variable" (though Jim has aptly called it the "Full Moon Variable").  Its historical roots are covered below.  So far, it seems to have predictive power and certainly interesting dinner conversation potential.  Read on for a discussion of variables in general... and if your modelers don't include variables like this, perhaps you should give us a call.  :)  Information professionals use “variable” in a technical sense. They typically mean any field (column) in a database. A more philosophical use is: "Something not in the raw data, something appended or calculated, an additional level of abstraction or summarization." Unfortunately, there are an infinite number of variables for any one object or event. (Call/write if you don't believe it) A marketing executive with even moderate creativity can come up with hundreds of possible variables. Consider the famous “werewolf” variable, the distance from the customer’s purchase to the nearest full moon. The werewolf variable was coined (by me) to suggest a variable no one had yet built but anyone could understand. The moon effects births, homicides and snipe spawning, who says it doesn’t effect proclivity to buy? Remember, if you don’t build the variable, you’ll never know. My working definition for “variable” is. . . anything you can think of. The obvious problem with creating dozens or hundreds of additional variables is that your head can no more hold these than it could hold millions of rows of data. Most people can only manage list selection with about six variables. There has to be a system built to see which ones are most important in specific situations. There must be a place to hold everything, there must be calculations made and there should be graphic depictions of what is going on. It is this necessity for pulling it all together that requires modeling. Modeling is the process that brings all the variables together in a meaningful structure. Our heads can’t hold it all, so the computer can help us. But we’re not going to let the computer make the decisions. The process still requires thought. “Since one pair of human eyes has more information processing power than all of the computers in the world put together, the entire thrust of the information revolution is to leverage this power...” Dent, Harry S. Jr., The Roaring 2000's, Simon & Schuster, 1998, New York, NY The vast majority of what we do when we build a database analysis system with all the pretty charts, maps and diagrams is little more than boiling it all down to a form that makes you think. When you start thinking about turning data into money, the real problem, the reason people don’t get it to happen is because they do not think. It is not a computer problem. When people build their databases, they do not want to think. Turning data into money requires variables, variables require thought and thought takes work. Sorry, you probably were expecting magic at this point. You cannot abdicate your responsibility of thought. We may be able to build a system which will give you pictures but you still have to figure out how attractively to offer quality goods and services to interested qualified people. It takes people to know what people want.

Page 3 of 4