February 20, 2007

Integrating Constituent Data

Increasingly, organizations are realizing the impact of student life, patient experience, and membership on philanthropic behaviors. At a recent CASE conference, I presented the entire life-cycle of constituents as it pertains to prospective donor identification. My goal was to encourage increased partnership between various campus/organization departments. However, much of our conversation focused on the integration of data between these various departments. Since this is very much on people's minds, I thought this article would be interesting to this community.

Integrating all your constituent tracking applications so they can share data is highly desirable, but how should you go about it? We walk through the pros and cons of three basic methods of integrating data: Manual Import/ Export, Integrated Packages, and Automated Connectors.

Read More

Labels:

February 16, 2007

Model Selection for Major Giving

This past couple weeks have been active on Prospect-DMM, the discussion group for data mining in fundraising. One discussion called into question the use of certain modeling techniques for major giving. I thought my brief response might be useful to this forum.

From Prospect-DMM:

For identifying major giving prospects, it makes sense to try various avenues.

A reason binary logistic regression or a decision tree on a binary variable (C5) maybe effective is because of the nature of major giving donors.

Certainly, a pathway to major giving via increasing annual support is a pattern found among major donors. These donors might be making gifts informed by cash flow (what can I afford to give this year?). And, using different ordinal or linear techniques makes sense.

However, there tends to be a large group of major donors--usually the majority of many files I review--that have very inconsistent "pre-major giving" gift behaviors. They tend to give gifts out of assets and are motivated by an investment frame of mind. Since they are very different from the overall donor population and tend to be a small pool, categorizing these large outright donors as a 1/0 makes a lot of sense.

Only using factors related to levels may miss many new opportunities (it also may not - each data file is different). Likewise, undocumented planned gift donors and future volunteers are difficult to predict using other modeling techniques.

I would rather be equipped with a large arsenal of techniques and cater my approach/es to the specific modeling need. In major gift modeling, this may require diversifying the models to find as many new leads as possible.

Labels:

Awarding Advancement Services

I think it is wonderful that Ithaca College recognized the contribution of an Advancement Services professional, Lori Watkins, to their campaign. Congratulations Lori!

Although the focus of this service is to highlight analytics, I am always pleased to see fundraising organizations reap the benefits of investment in infrastructure. Services, research, and operations are critical to fundraising success.

See the award posting

Labels:

Putting the Business Intelligence Puzzle Together

Analytics is a key to understanding constituent behaviors, tune processes, and project future behaviors across many industries. The following article is a bit one-sided in promoting Advizor Solutions, but the presentation of business intelligence is worth a the quick read.

Combining data mining, statistical and numerical analysis and data visualization, the use of business intelligence applications is, as a result, spreading fast. What is differentiating competitors in the field is the BI application's ability to perform compound, ad hoc queries on large data sets on the fly and graphically display and distribute results in an intuitive, easily comprehensible format.

Read More

Labels:

Data Mining in Financial Services

This account of the integration of data mining with the insurance and financial services industry has much overlap with the approach to fundraising. Although many fundraising organizations are now using data mining to predict donor behavior, few are digging into their data to profile existing donor segments.

"We wanted to be able to analyze profiles and preferences of existing customers and then predict buying behavior," says Alt-Simmons. "And, we wanted to make sure our marketing efforts were what they needed to be."

Read More

Labels:

February 1, 2007

Communicating Data Mining Results

A challenge that faces almost every fundraising analytics professional is explaining models and analysis for non-technical front line staff. I generally use analogies and stories liberally. By comparing a propensity model to a credit score, or comparing the incorporation of many independent variables to TVs with more pixels, I find it easier to get over the deployment hurdles.

A Whole New Mind by Daniel Pink is a must read for all data miners. Andrew Sallee of William Jewell College pointed it out to me as we were discussing statistics at the SPSS users forum. We couldn't help but notice the varied backgrounds of analytics professionals. From our anecdotal observations, these backgrounds seemed heavy in the arts. This book will shed some light on not only the personalities predisposed to discovery, but also techniques for telling the story.

This article, "Science-speak 101: Researchers sometimes need help to explain complex work in simple terms" might also shed some light on the matter.

Some business counselors urge scientists to use allegories, metaphors and everyday images to make their technology understandable. "I don't want to say they should dumb down their stories, but that's what it amounts to sometimes," said Mark Long, president and chief executive of IU Research & Technology Corp., which runs a life-sciences business incubator near Downtown Indianapolis.

Read More

Labels:

Building a Predictive Model

In a previous post I described the standard process for building a predictive model. I found this older article, "The non predictive part of predictive modeling." It provides a bit more detail on the project outline. Based on my experience, I would agree that the actual modeling rarely exceeds 10% of the total project time.

Some catalogers may be intimidated by the techniques required to build a statistics-based predictive model. But actually generating the predictive model - that is, creating the scoring equation - makes up about only 10% of the entire six-step process. The remaining 90% encompasses the nonpredictive part of predictive modeling: developing a sound research design, creating accurate analysis files, performing careful exploratory data analysis, implementing the model, and creating ongoing quality control procedures.

Read More

Labels:

Boomers and Planned Giving

Increasingly, nonprofits are targeting planned giving appeals to younger audiences. Lately, when I have built custom planned giving models, my clients requested that I control for age as well as consistent giving. Older, regular donors have long been the population for targeting. Since those variables are constantly targeted, it is difficult to find other factors without controlling them.

Families with more than $10 million to give away often create family foundations rather than give to established charities, but taxes and paperwork discourage that for those who are wealthy but have smaller estates - in the $1 million to $2.5 million range. At the same time, only 42 percent of people have wills.

Read more

Labels: