Modeling: Time to Get On Board
Attracting new donors to your organization is critical and finding ways to do it effectively is key. Should modeling be a part of your donor acquisition strategy? Most likely the answer is “yes.”
There are several different techniques used to locate individuals in the general population who have a propensity to give to your organization. Here are a few:
Conduct Demographics and Geographic Analysis
You can simply do a demographic study of your active donor base and replicate the attributes that are most prominent (heavily represented and significantly different from the national average). So if your donor file is made up of 92% females, 76% from Montgomery County, Pennsylvania, and 66% are married, you can simply order compiled data selecting these characteristics. The pros of this method are that it is relatively inexpensive and the demographic appends you’ve paid to do can also be used for other targeting (like messaging or identifying sustainer and planned giving prospects). But this method falls short because it doesn’t look at behavior or motivation, at passion or giving history, only at “flat” descriptive attributes.
Rent Non-Donor Affinity Lists
Rather than simply selecting a “flat” demographic list as described above, you can attempt to capture those same demographics via catalogs or magazines that attract the same type of person. Married women can be reached by way of magazines like Good Housekeeping, O, and Woman’s Day or catalogs like Chadwicks of Boston and Blair, for instance. One pro of this method is that you do have the benefit of knowing you’re reaching people responsive to direct mail. However, you still don’t know their heart and whether they have a desire to forward your organization’s mission. Catalog and magazine lists won’t generally perform as well as donor lists, but are often necessary to provide a sizable enough pool of names to mail and so can play a role in your overall donor acquisition mail strategy.
Rent ‘Feel-Alike’ Lists
We’ve all heard of look-alikes. But you need your prospects to not just look like your current donors, you need them to feel like them, be motivated in the same way, care about the same things. Much as outdoor gear catalogs will rent names from other outdoor gear catalogs, so too it is appropriate for you to rent donor names from organizations whose missions are very similar, or at least parallel, to yours. This technique’s pro is that you know you are renting givers and givers to organizations similar to yours. Response rates will reflect that. The con is that this method is not predictive of giving behavior to your organization.
A fourth technique is to find clones of your donors via statistical modeling. In its truest form, modeling seeks to replicate your donors’ behavior – i.e. their response to an appeal made by your organization through the mail. This is done by comparing and contrasting the responders to a specific appeal you made to the nonresponders of that same appeal. Statistically significant characteristic differences make it into the model. The model is essentially an algorithm that assigns a weight to each characteristic based on the degree to which it has bearing on predicting the behavior. This model is then applied to a compiled consumer or cooperative file, scoring each record with a likelihood to respond.
Other modeling forms can do the same thing, but model only the characteristics (not the behavior). They do not require mail and responder files, only your donor file.
The performance of any model is largely dependent on the datasets available for use in the analysis. In order to weight characteristics, we have to have the characteristics. The more data, the better. Donor data cooperatives are strong in this area because they have not just demographics, but the transaction histories for all the donors of all the participating cooperative member organizations. This data can serve as a surrogate for your own mailer and responder data in helping to replicate giving behavior.
Like any other list source, models should be tested first. Most model providers will build your model for free and charge you only for name rental. This makes testing relatively low risk.
If you haven’t explored the world of modeling, it may be time! Tap into our experience. Contact us to find out how to explore modeling as part of your overall new donor acquisition program.