asks and external data

One of the many things that I do every year as we roll from one year to the next is build an ask table.  That ask table has typically been based around a series of calculations that combine recency and frequency from gift history to determine what we predict to be the likelihood of that prospect increasing their giving balanced with the desire to retain them as a donor.  I have done 3 and 4 ask arrays and found no difference in returns on either number or size of gifts so have migrated toward the 3 asks - no reason to do additional work without additional results!

I calculate these asks so that our communications throughout the entire year are consistent -if you are asked in June for $78 to the School of Education Dean's fund, you will be asked in October and again in April for $78 to the School of Education Dean's fund.  The exceptions here are donors who have given and are being asked for a second gift, special projects that may have a different fund (but the ask remains constant) and prospects who are identified or qualified into the major gift process.

This year for the first time, I purchased the data to calculate asks off of two external pieces of information; a predictive model for household income and a version of that provides disposable income, adjusted for cost of living (important here in the northeast.)

Using this data as a third criteria, I created a table that calculates asks based upon an expectation that loyal donors will give more and be less bothered by an aggressive ask than a new or previously lapsed donor.  I then turned to data from the Chronicle (http://philanthropy.com/article/America-s-Generosity-Divide/133775/) to add a geographical component.  I then assumed that the average donor supports 5 charities and used the predicted income to establish how much they should be giving away.   The biggest assumption that I made is that each donor does not support all 5 charities equally - in fact they give a disproportionate amount of that to the top 1 or 2 and then the balance to the other 3 or 4.  I had to account for that by looking at that combination of gift total, recency and frequency as compared to what that should be if it was divided equally.  If that combination was greater than 20% of that expected total, I could assume that we were in the top.  If it was less than 20%, then I could assume that they were supporting us at a lower level of investment.

This gave me a good outline that was based on both what they had for capacity and history of giving, assuming that they had a history of giving that is.  For those folks who did not, we used the household income as a dividing point - for those who had household income over $250,000 we placed them into the leadership giving suspect pool to be asked for a gift of $1,000 or more.

This is an experimental exercise - has not yet produced results to the good or the bad but either way with some tweaking it provides a much richer and more robust process than simply looking at history of giving to your organization, without losing the impact or importance of that inclination indication.