Hospital’s Analytics Found Donors Large And Small

Database and direct marketers have long used analytical methods such as the Recency Frequency Monetary Value (RFM) model and logistical regression in order to identify trends and effectiveness in their efforts. But what would happen if an organization took tried and true analytics and applied them to new areas?

Two higher-ups at Memorial Sloan-Kettering Cancer Center (MSKCC), in New York City — Director of Direct Response Kim Walker and Campaign Strategic Research Director Kate Chamberlin — spoke during a panel called “New Horizons in Analytics” at the Direct Marketing Association Nonprofit Federation’s 2012 Washington Nonprofit Conference. They highlighted some of the hospital’s projects in taking “analytics principles and technology and applying them in new areas.”

MSKCC has approximately 6 million donors who mostly give small amounts in high volume. The hospital’s managers wanted to implement a more robust major donor program that identified potential big givers for extra attention. Chamberlin and her team of analysts focused on major gifts and principal gifts, $25,000 to $1 million and greater than $1 million, respectively.

Chamberlin used an integrated approach, drawing from the areas of analytics, prospect research, and development and information technology. First, she analyzed the donor base. The majority of the donors are from New York City and New York state. MSKCC donors are generally more wealthy and educated than the New York state population as a whole, and major gift prospects have usually already donated but have not given major gifts.

In creating a major gift propensity model, it is important to separate capacity from inclination — those who can give more than $25,000 might not necessarily want to do so. The first step is creating a capacity estimate. Chamberlin’s team calculated four estimates and used the largest: three times the largest gift a donor has ever given, total of everything given to date, a fraction of the median household income, and a percentage of total real estate value.

Most databases identify 1 percent as having major gift capacity, but for the propensity model, Chamberlin further narrowed that number down to living donors who have given at least $50 in the previous five years. She then used a logistic regression model, assigning points for variables such as the amount of the first gift given, the number of gifts, if the donor has a business or seasonal address, if the person has an executive job title, and the velocity of giving, i.e., has the number of gifts per year increased. She then provided a modeling software program with those variables, which assigned a probability ranking to each prospective major donor.

MSKCC found that 85 percent of major gifts came from that top one percent of the donor base. The regression model identified at least two principal gift givers; one had just sold his company, and prior to doing so had donated a total of $3,000, with $300 being the most he’d ever given. The other was an executive in a large corporation who helped secure a sponsorship deal for $1 million. Walker used Chamberlin’s research to identify lapsed donors and found that, while the rate of response did not change, the amount of gifts “increased substantially.”

Chamberlin also debunked some myths surrounding many major donors, such as that they are usually new to a donor file (80 percent of MSKCC’s major donors have given before), they do not make tribute gifts (53 percent had made a gift in a loved one’s name) and they do not give to events (23 percent have given to MSKCC’s events, such as its stationary bike relay Cycle for Survival).

Analytics can be used for more than just major donor identification. Chamberlin used another direct response staple, control group studies, to see if stewardship phone calls to high-end donors have any effect. Her team identified 700 donors; 200 got the standard treatment of direct mail, and the other 500 received a phone call thanking them for their support in addition to the direct mail. The phone calls brought in an extra $60,000 and increased the average gift from $92 per person to $215 per person.