Modeled Data: Models, Models, and More Models

by Dave Hare

In the previous post we discussed the various types of selects typically available on a list. These broke down into known, self-reported and modeled data. Because the results of models differ substantially by what goes into the model and the importance or “weight” it has in that particular model, it is a good idea to refine the information with a second and possibly a third model as we detail below.

In addition to the modeled data that we may order on a list and the model our computer service bureau uses, we also use an additional proprietary model we created based on average home value and average income per zip code, to divide the data at our service into “wealthy zips” and “regular zips.” The output from the service bureau divided the names into 3 ranges or “buckets” of income (or net worth) so in the end we had six buckets to choose from; over, under and unknown.

Combining Modeled Data for Best Direct Marketing Results

Thus all of the modeled data lists used in this project used at least two models and many used three; first the high income and/ or high net worth data inferred by the list owner’s service bureau’s model, then data inferred by our service bureau vendor’s model and lastly our own.

We probably know of an individual who drove an average car and lived in an average home that upon their passing gave several million dollars to the charities of their choice. These individuals would tend to append with lower income than the individuals that always seem to spend more than they make, regardless of how much they earn.

Cons to Marketing to an “Accredited Investor” List

So to increase your success targeting accredited individuals you need to become a savvy marketer and select the individuals for your offer carefully. We have several clients who market to accredited investors. One of them wanted us to keep it simple and had us rent an Accredited Investor list; it became one of the worst lists we ever tested for this client; the data was sourced by a reputable company from multiple sources (a compiled file); but the problem was that virtually no one responded.

The hard earned lesson here was that by the time these “Accredited Investors” became easily identified and marketed as such to the many businesses that want them as clients they were quickly over marketed to and became unresponsive.

Thus the key is to find individuals who meet the high net worth, high income demographic but are “under the radar.” We do that by selecting the list criteria a different way and then putting more attention on refining the data by seeing how it compares with the several models we use.

Using Available List Selects to More Accurately Model Your Target Market

Other common selects that are available include age, gender, zip code, home or business address. Additional selects available depending on the source of the data will include both volunteered and inferred interests and known data from public records and include personal interests such as skiing, hiking, golfing, hunting, fishing, fitness, health interest, wine, cigars, active reader, active donor (what type of donor; veteran, political, animal, disease, political and how much) participant in sports, gambling, watching sports, gardening, political interest, financial interest, owners of multiple homes and distance between, investable assets, net worth, college graduate, advanced degrees, CD maturity date, Jumbo CD, supporter of the arts, collector by type (fine art, rare coins, wine, firearms, antiques) boat owner (power or sail, number of feet) model of vehicle owned, number of vehicles in the household, aggregated credit score, length of residence, single family home (by square foot and by lot size) and  more.

Thus try to find out more about your existing clients and then select your list rentals to mirror them.  Test multiple lists and track the response and conversion of each. Track and source the referrals back to the source list to lower the cost of sale.

So typically high net worth or very high income individuals who are also accredited investors are age 50 or 55+ and traditionally they tend to be male. Interestingly enough although we originally mailed a piece to the male individual in the household we find instances of the wife being the one who responded and had their name first on the investment documents.

The bottom line is that using modeled data in your marketing efforts can be effective; however, the accuracy and effectiveness of your results will climb by combining two or more modeled data sources to more accurately target your audience.