Akkam’s Razor

When I worked for Saturn (at a local retailer) I was dismayed by the Automobile Dealers' idea of successful marketing.  The typical ploy went like this - go to your marketing firm, pick a geographic area, set a range for desirable credit card scores, refine the population list by score, design and drop a mailing, and go.  The typical mailing promised some kind of schmeeke, where you got some worthless trinkets or a chance at winning a car (which I'm sure was run in a fair and transparent manner) and maybe, maybe, some small percentage of the people walking through the front door would buy a car.  See, the point of the advertising wasn't to find people who want to buy a car - the point of the advertising was to increase floor traffic.  The standard maxim is that you can close (get someone to agree to buy 33% of the time).  So, if you get 300 people through the door, you should sell 99 cars.  If you want to sell more cars, you need to get more people through the door.  Supply side economics at it's finest.

The whole arrangement seemed wasteful to me, and I was sure the better information was available that could be used in different ways.  In my position as Financial Services Manager, I maintained my own databases, and often manipulated the data to reveal trends that might normally escape notice.  In looking at these, comparing it with what I had observed in my interviews with customers, viewing credit reports, and aggregated loan and credit bureau information, an increasingly clear profile of our customers became clear.

For example, the way sales cluster around certain "hot spots", or how certain cars ALWAYS were traded in (Taurus', Luminas, Sentras, Pulsars, Tempos, Sundances, etc.).  Surely there had to be some pattern that could be exploited.  I then came across the Polk Dealer Marketing Manager (PDMM).  It appears that PDMM is now the Polk Data Enhancement Service:

Polk’s Data Enhancement Services can help you understand your customers better—where they live, what they earn, what they enjoy doing and what their households look like. This will allow you to further segment your markets, and more directly target your marketing campaigns toward the people who most resemble your current client base.

R.L. Polk, for those who don't know, is one of the originators of database marketing.  If you've bought a product and it had a warranty card inside, odds are it went back to R.L. Polk.  The data they collected included age, ethnicity, income, household size, whether you rented or owned your home, automobile ownership status, employment status, lifestyle indicators, responsiveness to direct mail campaigns, hobbies and interests, and other demographic indicators.  R.L. Polk went a step further by merging this data with other data that they had access to (the dealer didn't have access to this raw data) including credit reports, vehicle registrations, and bankruptcies.  This in turn was appended to the dealerships own databases. 

By using the appended database, one could see how cash, finance, and lease customers differed.  You could see how many owners self-responded that they prefer to read books, play video games, watch movies, or play sports.  You could see who had a computer, or a pet, or a second home.  You could see who had adult children at home who may need a car in the near future.  The possibilities were limited only by our imaginations and creativity.

In the end, we found the project to be minimally useful, primarily because the two dealerships did not have a significant owner-base to begin with, and secondly because Pennsylvania is a closed-records state for vehicle registration, meaning that one of the primary data-sources was unavailable.  Had that been available, we could have profiled our customer demographic, selected people who were similar and had a high affinity for purchasing a Saturn, and further refined that based on where someone was in the automobile purchasing cycle (recent college grad, young adult, first time buyer, satisfied conventional finance loan, or lease expiration customers).  The hope, which was not meant to be, was that we could refine the list to get the best leads and to count on a supply of an extra ten to twenty cars per month (with one sale more than covering the break-even for the service).