CoolData blog

10 March 2011

Gifts of stock as a predictor of Major Gift potential

Filed under: Major Giving, Model building, predictive modeling, Predictor variables, regression — Tags: , — kevinmacdonell @ 6:09 am

(Image used via Creative Commons license. Click image for source)

In an earlier post, I wrote about giving-related variables and whether or not they’re okay to use in a model that is trying to predict giving itself. (My answer was “it depends”. See Giving-related variables: Keep or leave out?) Today I zero in on a specific example: gifts of securities as a predictor of major giving.

Following the logic of my earlier post, if the sample of people whom you intend to score includes non-donors, and you want non-donors to have a chance of making it onto the radar, then you must rule out ‘Gift of Stock’ as a predictor. Why? Because you want to keep any proxy for your outcome variable (the Y side of your equation) out of the predictors (the X side of the equation), as much as possible. A ‘yes’ for ‘Has made a gift of stock’ is possible ONLY for the donors in your sample, and will provide no insight into a non-donor’s potential for major giving.

But giving-related variables are frequently used to predict major gift potential. Gift count, first gift, recency, and stock gifts are all enticing predictors. You have a decision to make: Do you exclude non-donors, or leave non-donors in and forgo the potential predictive power of these variables?

For some the answer might be easy. If the vast majority of major donors to your institution had some prior giving before making their biggest gifts, and a major gift from a non-donor is extremely unlikely, then it makes sense to exclude non-donors. This makes most sense for alumni models: Alumni who are solicited every year and don’t give are rather unlikely to turn around and give a million dollars. (Although it happens!)

You can avoid having to make the decision, however, if you build two models: One including non-donors (and using no giving-related variables), and one excluding them (freeing your hand to use giving-related variables). That’s what I do. I test the output scores against a holdout sample of major donors, and whichever model outperforms in scoring the major donors will be my choice for that year.

Let’s say that at least one of your models is a donor-only model, and you’re itching to use ‘Stock gifts’ as a predictor. Hold on! You’re not done yet. You need to evaluate the degree to which ‘Stock gifts’ is independent of your DV. If the variable equates to major giving itself, it is not at all independent and should be excluded. It is merely a proxy for being a major donor.

It’s clear that stock givers are different from other donors. In the data set I have before me, alumni who have made at least one gift of stock have median lifetime giving of about $40,000, compared with all other donors’ median giving of about $170. More than 66% of stock donors have lifetime giving over $25,000, and more than 90% of them have made at least one gift of $1,000 or greater.

The fact of having given a gift of securities cannot seriously be considered “independent” of the DV, but the degree of non-independence varies with how the DV is defined. If I define it as “LT Giving over $25K”, I’m probably in the clear, because a considerable portion of stock donors (34%, in my data set) fall outside the definition of my DV. If my DV is “One or more gifts of $1K or greater,” however, I should steer clear of the stock-gifts predictor. True, not all stock donors are in the DV, but almost all of them are.

Stock donors probably represent a very small percentage of all your donors, so the variable may have little influence either way: Not a high-value predictor, but not a damaging one, either. (Given the limited number in your sample, the correlation coefficient is going to be pretty low.) Maybe if 85% of the stock donors were in my DV, instead of 90%, I might go ahead and use it. So in the end, it’s a judgment call based on what seems to make sense for your data and what you hope to get out of it.

Advertisements

2 November 2010

Finding prospects flying under the radar: A nuts and bolts approach

Guest post by Peter Wylie and John Sammis

(Downloadable/printable version available here as a PDF: Flying Under the Radar)

Let’s say you’re a prospect researcher in higher education.  You’re getting some pressure – from your boss, from some of the gift officers you work with, maybe the campaign director – to come up with a list of new prospects. They use different words, but their message is clear:

“We’ve picked the low hanging fruit. We don’t want to keep going back to the same alums who’ve been helping us out in a big way for a long time. We need to find some new people who have the capacity and willingness to make a nice gift. Maybe not a huge gift, but a nice gift.”

If you’ve been working in the field awhile, this isn’t the first time you’ve faced this problem, nor is it the first time somebody has offered advice on how to solve it. Truth be told, you may have gotten too much advice:

  • “You haven’t done a screening for five years. You need to do a new one.”
  • “Our company has gotten very sophisticated about predictive modeling as well as gift capacity ratings. Use us.”
  • “You’re not using social media resources effectively. Facebook and MySpace are great places to find out about alums who have lots of financial resources and are philanthropically inclined.”
  • “You need to learn how to do data mining and predictive modeling or add somebody to your staff who already knows how to do it.”

It’s not that any of this advice is bad, even if it comes from a vendor whose goal is to get some of your business. The problem is that you, or the person you report to, has to sift through this advice and make some kind of decision — even if that decision is to do nothing different from what you’re currently doing.

Since John Sammis and I are some of the people out there offering this kind of advice to advancement folks, we often ask ourselves: “Are we making things too complicated for the people we’re trying to help?” Often the answer we come up with is, “Probably.” Why? That’s a whole can of worms we’d rather not get into. The short answer is that both of us grew up in an educational system where precious few of our teachers and authors of our textbooks were good at making things simple and clear. And like it or not, we’ve inherited some of their tendencies to obfuscate rather than elucidate. But we fight against it as best we can.

Hopefully we’ve won that battle in this piece. (You’ll decide if we have.) Anyway, what we’ve done here is use some data from a large public higher education institution to walk you through a simple process for finding new prospects.

Before we do that, let’s start off with three assumptions:

  • You have fairly recent gift capacity ratings for several thousand of your solicitable alums, some of whom you think may be good, untapped prospects.
  • You have ready access to someone who can develop a simple score for all those alums with respect to their affinity to give to your school.
  • You have reasonably good profiles on each of these alums. That is, those profiles include information like lifetime hard credit dollars given; date and amount of last gift; date and amount of first gift; what gift officers have ever been assigned to those alums and when; the most recent occupation of the alum; and so on.

Here are the steps we want to take you through:

1.     Look at the distribution of gift capacity ratings for the alums you’ve recently screened.

2.     Look at the giving data for these alums by gift capacity ratings.

3.     Have someone build you a simple affinity model using some very basic information stored on each alum.

4.     Pick a small group of alums who have a high capacity rating, a high affinity rating, and are not currently assigned to a gift officer.

5.     Look closely at the alums in this small group and identify some who may deserve more scrutiny.

We’ll go through each of these steps in detail:

Look at the distribution of gift capacity ratings for the alums you’ve recently screened.

Whenever you have a field of data (whether it comes from your own database or has been delivered to you by a vendor), it’s a good idea to make a frequency distribution of the field. (In statistics the term is “variable,” not “field,” so from here on out we’ll say “variable.”)

Here are a couple of reasons for doing this:

  • You get a big picture look at the variable. Our experience is that most people in higher education advancement don’t do this for the many variables they have in their alumni databases. For example, let’s say you asked the average associate vice president for advancement in a college or university this question: “What percentage of your solicitable alums have given $100 or less lifetime hard credit?” Our bet is that the vast majority would have no idea of what the correct answer was; moreover, they would be shocked if you told them.
  • You get a chance to see if there is anything out of the ordinary about the data that’s worth further exploration. Here’s a good example. When doing predictive modeling for a school, we look closely at the variable “preferred class year.” It’s a measure of how long alums have been out of school, and it’s a reasonably good measure of age. It’s not at all uncommon for us to encounter thousands of records coded as “0000” or, say, “1700.” Call it a hunch, but we’re pretty sure those folks didn’t graduate the year Christ was born, nor 75 years or so before the Declaration of Independence got signed. When we encounter a problem like this, of course, we ask the advancement services people we’re working with what those codes mean. The answers vary. Sometimes such codes indicate alums who are non-degreed. Sometimes they indicate alums who simply received a certain kind of certificate. Or they indicate something else. The important thing is that we ask; we clear up the mystery.

All right, Table 1 shows a distribution of the gift capacity ratings for a group of about 22,000 alums in the public higher education institution we mentioned earlier. Figure 1 displays the same distribution graphically. Take a minute or two to look at both of them. Then you can compare what you see with what we noticed.

Table 1: Estimated $Gift Capacity for Over 22,000 Alumni Divided Into 20 Groups of Roughly 5% Each

For us, two things about these data stand out:

1.     Some of the data are a little hard to believe. Let’s take a look at Group 1 in Table 1. There are 1123 records in this group. They comprise alums with the lowest five percent of gift capacity ratings. If you look at the “min” column, you’ll see that the lowest gift capacity rating is one dollar. Really? That alum must be down on his or her luck. You can’t see all the data in this distribution the way we can, but there are a total of 11 alums whose gift capacity is listed as being under $100. Obviously, you should be suspicious of such ratings. Contacting the vendor who generated them is a must. And politely staying after them until you get an acceptable answer is the right thing to do.

2.     The capacity ratings rise slowly until we get to the top ten percent of alums. There’s nothing particularly surprising about this. However, it is interesting (without showing you all the arithmetic) that, of the roughly one billion dollars of total gift capacity for these alums, over half a billion of that gift capacity resides with the top 10% of the alums.

Look at the giving data for these alums by wealth capacity ratings.

We’ve taken a look at the distribution of gift capacity ratings for the alums we’ve screened. Now let’s look at how those capacity ratings are related to the money the same alums have given to the school.

We’ll start with Table 2. The two columns on the right of the table (“Total$ given” and “Max$ given”) contain the most important pieces of information in the table. The “Total” column simply shows the total lifetime dollar amount given for the alums at each of the 20 gift capacity levels. The “Max” column shows the maximum amount any one alum has given at each of these levels.

Table 2: Giving Data for Over 22,000 Alumni Divided Into 20 Groups of Roughly 5% Each by Gift Capacity

We see a pattern that emerges from this table, but it’s a little hard to detect. So go ahead and take a look at Table 3 and Figure 2. Then we’ll offer our thoughts.

Table 3: Percentage of Alums Giving $50 or More Lifetime by Gift Capacity Level

When we look at these two tables and this one figure, two conclusions emerge for us:

1.     There is some relationship between gift capacity and giving, but it’s not a strong one.

2.     If we can believe the gift capacity ratings, there is a huge amount of untapped potential for giving, especially at the highest capacity levels.

Let’s start with the first conclusion, that there is not a strong relationship between capacity and giving. How do we arrive at the conclusion? Let’s go back to Table 2. Now if we just look at the five percent of alums with the lowest giving capacity (Group 1) and the five percent of alums with the highest giving capacity (Group 20), we see that the total lifetime giving goes from $34,062 to $2,396,810. That’s a big difference. The wealthiest alums have given about 70 times as much as the least wealthy alums. Also, the most generous alum in the lowest capacity group has given a lifetime total of $2,005 compared to the most generous alum in the highest capacity group who has given a lifetime total of $224,970. Again, we see a big difference.

But look at what happens in between these two extremes. Things bounce around a lot. For example, let’s compare the giving between capacity level 3 and capacity level 12. The total giving amount for the former group is $152,741; the total giving amount for the former group is $125,477. In other words, alums with a much higher giving capacity have given less than alums with a much lower giving capacity.

Further evidence of this “bouncing around” is apparent when you look at Figure 2 (a graphic version of Table 3). This chart shows the percentage of alums at each of the 20 giving capacity levels who have given $50 or more lifetime to the school. Notice how these percentages dip in the middle of the capacity range.

So let’s go back to our conclusion that there is some relationship between gift capacity and giving, but that it’s not a strong relationship. Yes, the overall trend of giving goes up with gift capacity, but we can in no way conclude that knowledge of an alum’s gift capacity is a good indication of how much he/she has given.

Okay, how about our second conclusion that there is a huge amount of untapped potential for giving, especially at the highest capacity levels? We think Figure 2 provides plenty of support for that conclusion. Look at the highest gift capacity level. Barely 50% of the alums in this category have given over $50 lifetime. Not as a single gift. No. Lifetime.

If that doesn’t convince you of the untapped potential for giving among such wealthy alums, we’re not sure anything will.

Have someone build you a simple affinity model using some very basic information stored on each alum.

Now comes the tricky part. Now comes the part where we risk losing you because we get a little too technical. We don’t want to do that. We want to avoid having you end up saying, “Geez, these guys said they were gonna make this simple, but they didn’t. Now I’m more confused than I was before I started reading this thing.”

This is not a perfect solution to the problem, but we think it might work. We’d like you to find someone who works at your school who can help you. Of course, it would be great if you already had someone on your advancement staff who fits that bill – someone whose job is focused on data mining and predictive modeling. Some schools have folks like that, but most don’t. (We’re assuming you don’t, otherwise there wouldn’t be a whole lot of need for you to be reading this piece.)

Anyway, the person you’re looking for is probably a stats professor in the psychology or education department, a graduate student pursuing a degree in that area, or someone who works in what is often called “institutional research.” Ideally, the person you find should be:

  • Someone who is helpful and accommodating. This seems obvious, but, sadly, a lot of people in higher education don’t meet these criteria. Maybe a quick and easy way to decide is to ask yourself: “If I walked into a high end department store, is this a person I would want to help me?” If your answer is not an unequivocal “yes,” you should keep looking.
  • Someone who is good at explaining things in clear, simple English. Candidly, a lot of people in the technical arena are not good at this. As we said earlier, the two of us (try as we might) struggle with making things clear to the people we work with. What we’d suggest is that you look for someone who is patient with you when you don’t understand something they say. You especially want to avoid someone who acts the least bit impatient and condescending if you don’t “get” something the first time they explain it.
  • Someone who knows at least a little about major giving. The person does not need to be an expert in prospect research. But he or she should have a sense of how an advancement office works and some of the pressure that prospect researchers and development officers endure, especially when they’re scrambling to meet campaign goals.
  • Someone who has good skills with a stats software package. We think this is a must. If the person only knows how to use Excel to analyze data, that’s not good enough. The person needs to be proficient in a package like SPSS, which is widely available on college and university campuses. John and I prefer a package called Data Desk, but the important point is that your person needs to be proficient with an application whose purpose is sophisticated data analysis.
  • Someone who’s had some experience with multiple regression. You may or may not have heard of multiple regression. Don’t worry about that. Just be able to confirm that the person who helps you has a solid working knowledge of the technique. A good way to find out if that’s the case is to ask the person to explain the technique using a simple example using some of your own data.

Let’s assume you’ve found someone to help you. As we said earlier, if you follow our plan, that person will build you a simple affinity model using some very basic information stored on each alum for whom you have a capacity rating.

For the benefit of that person, we’ve described below how we developed the model for the school we’re using as an example. We’ve tried to provide just enough detail to give your person a guide, but not so much that we bog the paper down with too many words.

Enclosed in the boxes below (so you can skip over it if you wish) is a summary of what we did:

We chose lifetime hard credit giving as our dependent variable. To each record we added one dollar of giving to arbitrarily rid the sample of zero givers. We then performed a log to the base 10 transformation on this variable to reduce as much of the positive skewness as possible. 

We chose the following predictors (independent variables) for entry into our multiple regression analysis:

  • MARITAL STATUS MISSING (the alum was given a 1 if there was no marital status listed for him/her in the database, otherwise a 0)
  • MARITAL STATUS SINGLE (the alum was given a 1 if he/she was listed as “single” in the database, otherwise a 0)
  • CLASS YEAR (the alum’s preferred year of graduation)
  • THE SQUARE OF CLASS YEAR

 

 

  • HOME PHONE LISTED (a 1 if a home phone was listed in the database for the alum, otherwise a 0)
  • BUSINESS PHONE LISTED (a 1 if a business phone was listed in the database for the alum, otherwise a 0)
  • EVENT ATTENDED (a 1 if an alum was listed as ever attending an event after graduation, otherwise a 0)
  • E-MAIL LISTED (a 1 if an e-mail address was listed in the database for the alum, otherwise a 0)

Table 4 summarizes the results of the regression analysis:

Table 4: Regression Analysis Table for the Simple Model Developed for This Paper

R squared = 24.9%     R squared (adjusted) = 24.9%
s =  0.9835  with  22446 – 9 = 22437  degrees of freedom
Source Sum of Squares df Mean Square F-ratio
Regression 7213.27 8 901.659 932
Residual 21701.5 22437 0.967218
Variable Coefficient s.e. of Coeff t-ratio prob
Constant -1603.28 218.2 -7.35  ≤ 0.0001
MARITAL STATUS MISSING -0.333961 0.01845 -18.1  ≤  0.0001
MARITL STATUS SINGLE -0.472877 0.01603 -29.5  ≤  0.0001
HP LISTED 0.243367 0.01496 16.3  ≤  0.0001
BP LISTED 0.685641 0.03374 20.3  ≤  0.0001
CLASS YEAR 1.64819 0.2196 7.51  ≤  0.0001
SQUARE OF CLASS YEAR -4.23E-04 5.52E-05 -7.66  ≤  0.0001
EVENT ATTENDED (YES/NO) 0.712603 0.05089 14  ≤  0.0001
EMAIL LISTED 0.422934 0.01487 28.4  ≤  0.0001

We divided the predicted scores from the regression for alums with the highest gift capacity into twenty roughly equal-sized groups where 1 was low and 20 was high.

Okay, where are we here? In the “boxed in” technical suggestion above, the last thing we said was: “We divided the predicted scores from the regression analysis for alums with the highest gift capacity into twenty roughly equal-sized groups where 1 was low and 20 was high.” Well, what does that actually mean?

Let’s start with the specific group of alums we’re most interested in looking at. These are the 1,123 alums who got the highest gift capacity ratings. If you go all the way back to Table 1 (which you don’t really need to do), you’ll see that their total gift capacity is $405,958,000 – a lot of money.

Our regression analysis created a very granular affinity score for this group. It had 408 different levels. The alums with the lowest of these scores (according to the regression analysis) are least likely to give a lot of money to the school; the alums with the highest of these scores are the most likely to give a lot of money to the school.

That’s terrific, but 408 score levels is a lot of levels to get your arms around. So what we did is take those scores and chop them up into 20 roughly equal sized groups from 1 to 20, and (again) 1 represents the lowest scores; 20 represents the highest scores. Detailed giving data on all these 1,123 alums is displayed in Table 5 below. We can look at those data in a moment, but let’s move on to our next step.

Pick a small group of alums who have a high capacity rating and a high affinity rating.

Table 5 gives us lots of information about where we’re likely to find this small group. Let’s see what looks interesting here. Remember, everyone in this total group of 1,123 alums has a gift capacity rating greater than $116,000. This is a wealthy bunch of folks – no question about that.

We’ll start with the lowest group, group 1. These 56 alums have the lowest affinity scores of the total group, and their giving data confirms that. Look at the value for these alums in the “sum” column: $430. That means that all 56 alums, as a group, have given less than $500 lifetime to the school. That works out to a mean (average) lifetime gift of less than $8 per alum. Our conclusion? These folks may be wealthy, but both their affinity score and their history of giving have them speaking loud and clear: “Our philanthropic interests are aimed at worthy causes other than our alma mater.”

Now let’s jump up to the top group, group 20. Notice that there are exactly the same number of alums in this group as in group 1 (56), but the giving data for this top group is quite different from the bottom group. Most notably, they’ve given a total of $483,789, well over a thousand times as much as the bottom group. So here we have a group of alums who (a) we know are wealthy; (b) have a high affinity rating developed from the regression analysis; and (c) have already given the school quite a bit of money.

Table 5: Giving Data for Over 1,123 Very High Gift Capacity Alumni Divided Into 20 Groups of Roughly 5% Each by Affinity Score

Look closely at the alums in this small group and identify some who may deserve more scrutiny.

Now we can take a very close look at this group in Table 6 (below, near the end of this post). It lists the total giving and gift capacity for each of these 56 records. (Remember, each of the 56 alums has a high gift capacity rating, and each has an affinity score that says they really like the school.)

We’ll start off with a couple of alums who have already given a considerable amount to the school. What’s particularly interesting about these two is how different they look from the perspective of the possibility of very large future gifts.

  • Record #1: From the looks of things, this person is probably well known to the research staff and to the gift officers. The person has given more than $100,000 and has a gift capacity that’s not a whole lot more than that amount. We’re pretty sure the school would like to have a lot more alums like this one.
  • Record #7: We find this one pretty interesting. The alum has given a bit over $21,000 lifetime, but their gift capacity is listed as well over $13 million. Since the alum clearly likes the school, and they have considerable wherewithal to give a lot more, why haven’t they given a lot more? Maybe there’s a good reason, maybe not. At the very least this is someone who deserves continued attention both from the prospect research side of the house and the gift officer side of the house.

Now we’ll move down to five alums (Records #15, 17, 18, 20, and 24) all of whom have given less than $6,000 lifetime but whose gift rapacity ratings all exceed $400,000. Here we are probably in the neighborhood of prospects who truly are flying under the radar. They may have been assigned to a gift officer. And when a prospect researcher looks at their profiles, the researcher may say, “Yeah, we know about him.” But our experience tells us that alums like these are worth a harder look. For example, we would ask:

  • Is the alum really assigned to a gift officer, or did the last gift officer simply write the alum off as not a “good prospect” with no good documentation as to why that decision was made?
  • What does the alum do for a living? Does that occupation (e.g., investment banker) jibe with the gift capacity rating?
  • Has the alum been an active volunteer or season ticket holder?
  • Is he or she at the age where a sizeable planned gift might be a possibility?

You get the idea. With folks like these we think you should dig a little. Some of them may be at what Malcolm Gladwell calls “the tipping point.” They may be right on the verge of making a much larger gift if you do a little more research on them and send the right gift officer out to meet with them.

By the way, take a look at Record #56. This person is really rich, the internal data says he/she really likes the school, but this person hasn’t given any money. We’d sure like to know the story about this person.

At the ending of Table 6 we offer some closing comments. We really appreciate your staying with us up to this point.

Table 6: Giving and Gift Capacity Data for All 56 Alums in the Highest Affinity Group


Some Closing Comments

We’ve put a lot of tables and charts in front of you. That’s a lot of information to absorb. Several thoughts that might be helpful:

  • If you found what we’ve said here intriguing but also a bit confusing, put the piece away for a few days. Then take another look at it. It should be clearer the second time around. If it’s not, please feel free to contact John Sammis at jsammis@datadesk.com and Peter Wylie at PBradWylie@aol.com.
  • Share what we’ve written here with a colleague whose opinion you respect but who disagrees with you about a lot of things. That should make for an “interesting” discussion.
  • Whatever you do, we hope this piece encourages you and others in your advancement group to take a closer look at all the data you store on your alumni. The two of us will never back away from the importance of doing that when you’re trying to save money on appeals and generate more revenue for worthwhile projects.

21 September 2010

A pauper’s guide to electronic screening

Filed under: Alumni, Major Giving, Peter Wylie — kevinmacdonell @ 5:20 am

Guest post by Peter B. Wylie

(This is a reprint of a paper previously published on the CASE website. Click here to download as a .pdf file: PAUPER_111306. — KMD)

You’re right. I cheated a little. I chose this title to grab your attention. (Actually I “borrowed” it from an old travel book about Paris.) However … even though you and the school you work for are not impoverished, your budget for electronic wealth screenings may not be as big as you’d like. If that’s the case, I want to offer some thoughts on how you can do a pretty good job of identifying major giving prospects – ones that may not be on your development officers’ radar screens – without doing an electronic screening.

I wrote this piece to build on another article I did several years ago (“Where the Alumni Money Is”). In that one I showed that the lion’s share of any school’s alumni donations come from the oldest 25% of alums who are not listed as ‘single’ or ‘missing’ in the marital status field. The data I used were from a group of eight schools that covered the gamut between public and private, large and small, and well known and less well known.

In this paper I used a similar but different group of four higher education institutions that allowed me to go a little farther than I did with the first article. Here’s a basic outline of what I’ll cover:

  • The main question I was trying to answer
  • The bet I was making
  • How I did the data analysis for the paper
  • What conclusions I think we can draw from the data analysis
  • Some specific suggestions for “action steps” you, your IT folks, and your development officers can take

The main question I was trying to answer

In the three years since I wrote the last paper I’ve learned a few things about major giving in general and prospect research in particular (things I had only an inkling of back then). One of the biggest things I’ve learned is this: When it comes to major giving in higher education, we rely too much on the data we gather on prospects from outside sources, and we don’t rely enough on the data we have in hand on these prospects.

Don’t get me wrong. I think getting data about prospects in your alumni database from outside sources is fine. The more you know about these folks (not just their wealth), the better. On the other hand, I’m bothered that most schools ignore the huge amount of internal data they already have on their alums.

At the risk of being a little too frank, here’s the kind of process I see go on far too often:

  1. In anticipation of a major campaign, a school will spend a considerable sum on an electronic screening of several thousand alums by one of the excellent firms that do that kind of work.
  2. The firm will return a huge amount of data to the school, most of which is designed to assess an alum’s capacity to make some type of major gift to the school. (Sadly, there is a good chance this data will not be used as part of the campaign. Why this is so often the case is a bit of a puzzle to me. But it happens far more frequently than any of us would like.)
  3. In the meantime the school will not delve into the huge store of data it keeps on its alums. Data that can point an accurate finger at which alums are most likely to give and which are least likely to give. Class reunion attendance is a good example. Unfortunately, most schools don’t store this information in their alumni databases. Those that do store it don’t use it to help them identify new major giving prospects — even though reunion attendance may well be the best single predictor of giving in higher education advancement. (At least I’ve never found a better one.)

Okay. Let’s say the scenario I’ve offered here is not atypical: A school isn’t going to do an electronic screening, or it’s going to ignore the screening data even if it does one. Beyond that, the school is not going to do any serious mining of its wealth of alum data to find good predictors of giving. Can we then come up with a reasonably quick and easy way for the school to still identify good prospects for major giving?

The bet I was making

I think the answer is an unequivocal yes. Why? Because for the last several years I have become more and more convinced that if you know an alum’s age along with a few other pieces of information about his/her marital status and contact information, you can do a pretty good job of determining whether that alum is a major giving prospect.

Does that sound farfetched? All I ask is that you withhold judgment until the end of the paper.

How I did the data analysis for the paper

For each of the four schools I mentioned above, I took these steps:

One. I gathered a representative sample (no less than 5,000) of alumni records for these fields:

  • Home phone listed (yes/no)
  • Business phone listed (yes/no)
  • E-mail address listed (yes/no)
  • Listed as something other than “single” or “missing” for the marital status (yes/no)
  • Preferred year of graduation

Two. I “guesstimated” the age of each alum by subtracting their preferred year of graduation from the present year and then adding 22 – the age that a lot of students receive their undergraduate degrees.

Three. I computed the percentage of total lifetime dollars by alums in each five year age interval.

Four. For every alum I created a simple score with a range from 1-5 based on whether or not that alum had a home phone listed, a business phone listed, an e-mail address listed, and whether or not the alum was listed as something other than “single” or “missing” for the marital status field. (For a further explanation of how this kind of score works, see the paper titled “A Simple Score.”)

Five. For the five year age interval with the highest percentage of lifetime giving, I computed the mean (average) lifetime giving for every alum by simple score level.

Let’s see what these steps yielded for School A. Take a look at Chart 1. (Click on chart for full-size view.)

Just to be clear, here’s how I constructed this chart. I used a sample of over 20,000 records from this school. For all of these records I computed the total amount these alums had contributed to the school at the time the sample was gathered. This value exceeded 100 million dollars. I then computed the total amount that alums in each five year age interval had contributed and converted these amounts to percentages of the total amount. For example, 34% of this lifetime amount of more than 100 million dollars had been contributed by alums aged 66-70.

A couple of things stand out for me about this chart:

  • Well over 80% of the lifetime giving comes in after alums reach the age of 60.
  • The under 50 crowd has contributed less than 10% of the total lifetime giving.

What do these facts say about where the focus ought to be on an upcoming capital campaign? We’ll talk more about issues like this a little later on.

Now let’s look at Chart 2 which shows the mean (average) lifetime giving by simple score level for this group of alums aged 66-70. (Click on chart for full-size view.)

Here’s the algorithm I used for this score: ‘HOME PHONE LISTED’ + ‘BUSINESS PHONE LISTED’ + ‘EMAIL LISTED’ – ‘SINGLE’ + 2. It’s pretty simple:

  • If an alum had a home phone listed, they got a 1 otherwise a 0.
  • If they had a business phone listed, they got a 1 otherwise a 0.
  • If they had any kind of e-mail address listed, they got a 1 otherwise a 0.
  • If they were listed as “single” in the marital status field, they got a minus 1 otherwise a 0 (people who are listed as single in a martial status field invariably give less than any other marital code).
  • Then, for each record, I added these numbers together and added a 2 to each number so there wouldn’t be any zero or negative scores. (Zero and negative scores tend to confuse people.)

So what we’ve got here is a very narrow age group of alums (66-70) who’ve given a huge amount of money to the school. Beyond that we’ve got a very simple score that sharply differentiates theses alums with respect to how much they’ve given.

What conclusions can we draw from the data analysis?

Let’s assume that the simple score is a rough measure of likelihood of giving to the school. Let’s call it affinity. And let’s say we isolate the alums in this group that have a score of 5 on this affinity scale. As it turns out there are just over 450 alums in this group that is between 66 and 70 years old and have a score of 5. Okay, now let’s look at the top ten lifetime givers in this group in descending order of giving:

$ 5,364,619

$ 5,038,707

$ 4,072,701

$ 1,644,404

$ 1,466,562

$ 979,197

$ 520,501

$ 496,978

$ 479,202

$ 256,006

What do you think? I think all these people are under stewardship – or at least they should be. You certainly don’t need to do a wealth screening on them. Right?

All right, now let’s look at the next ten alums in this group in descending order of lifetime giving.

$ 157,439

$ 110,621

$ 107,475

$ 88,834

$ 67,281

$ 60,017

$ 49,728

$ 41,825

$ 38,154

$ 35,728

Now it starts to get kind of interesting, doesn’t it? Finally, let’s look at the next ten alums in this group in descending order of lifetime giving.

$ 35,287

$ 35,202

$ 33,186

$ 32,190

$ 28,493

$ 28,366

$ 27,225

$ 25,401

$ 21,826

$ 21,359

You’ve probably anticipated me, but here’s where I’m going with this. As we get further and further down this list of 66-70 year olds who show high affinity to the school, we run out of people who are already on our radar screens, who are already assigned and being actively appealed to for a major gift. But I think a lot of these alums should be on our radar screens. So what if we don’t have wealth screening data on them? They’ve given something more than a pittance to the school already. They are by far in the highest giving age bracket of alums. And they have an affinity score that says they like to give to the school. Somebody should be reaching out to these folks. Why? Because some of them are poised to make a major gift. It’s that simple!

And how much did it cost us to identify them? A few hours of time from a talented IT person? I think it’s worth it.

At this point, I’ve either got you seriously interested in my line of reasoning or I’ve lost you. If I haven’t lost you, let me show you the equivalent of charts 2 and 3 for the remaining three schools whose data I looked at to prepare this article. Then we can step back and get some perspective on all this. (Click on each chart for full-size view.)

That’s a lot of charts and data to look at. But let’s go back to the basic question I was trying to answer. If a school doesn’t do an electronic screening, can we then come up with a reasonably quick and easy way for the school to still identify good prospects for major giving?

Again, I think the answer is yes. Why? Let’s summarize the logic I’m invoking from the data I’ve presented from these four schools:

  • Most of the lifetime giving in a school doesn’t start rolling in until alums reach at least the age of 55. So if you’re spending a lot of time looking for good prospects under that age, I don’t think you’re making great use of your time.
  • Within this older crowd of alums, you don’t need a lot of information beyond home phone listed, business phone listed, etc. to create a very simple score that sharply separates the big givers from the small or non-givers.
  • You can use the top end of that simple score to identify alums who aren’t yet on your development officers’ radar screens but who should be.

Now I’ve already said that the last thing I’m doing here is recommending that your school not do an electronic screening. Not at all. But I am saying that if you can’t do a screening, you have an option here that can help.

Some specific “action steps”

And finally, what about all you prospect researchers? After all, I wrote this thing for you guys. Here’s a suggestion on some steps you can take:

  1. Get your IT folks to do the same kind of analysis I’ve done here for your own school. Sure, you’ll get some push back from them. They’re overworked. But you can get them to do it for you.
  2. Look up those alums in a higher age category who have a high simple score but who aren’t yet assigned to a gift officer. Some of them will look very promising. I guarantee it.
  3. Take a few of the promising ones to a gift officer who thinks this data driven decision making stuff is sort of cool and ask him/her to contact these alums.

What do you have to lose?

A PAUPER’S GUIDE TO ELECTRONIC SCREENING
By Peter B. Wylie
You’re right. I cheated a little. I chose this title to grab your attention. (Actually I
“borrowed” it from an old travel book about Paris.) However … even though you and the
school you work for are not impoverished, your budget for electronic wealth screenings
may not be as big as you’d like. If that’s the case, I want to offer some thoughts on how
you can do a pretty good job of identifying major giving prospects – ones that may not be
on your development officers’ radar screens – without doing an electronic screening.
I wrote this piece to build on another article I did several years ago (“Where the Alumni
Money Is”). In that one I showed that the lion’s share of any school’s alumni donations
come from the oldest 25% of alums who are not listed as ‘single” or “missing’ in the
marital status field. The data I used were from a group of eight schools that covered the
gamut between public and private, large and small, and well known and less well known.
In this paper I used a similar but different group of four higher education institutions that
allowed me to go a little farther than I did with the first article. Here’s a basic outline of
what I’ll cover:
• The main question I was trying to answer
• The bet I was making
• How I did the data analysis for the paper
• What conclusions I think we can draw from the data analysis
• Some specific suggestions for “action steps” you, your IT folks, and your
development officers can take
The main question I was trying to answer
In the three years since I wrote the last paper I’ve learned a few things about major giving
in general and prospect research in particular (things I had only an inkling of back then).
One of the biggest things I’ve learned is this: When it comes to major giving in higher
education, we rely too much on the data we gather on prospects from outside sources, and
we don’t rely enough on the data we have in hand on these prospects.
Don’t get me wrong. I think getting data about prospects in your alumni database from
outside sources is fine. The more you know about these folks (not just their wealth), the
better. On the other hand, I’m bothered that most schools ignore the huge amount of
internal data they already have on their alums.
At the risk of being a little too frank, here’s the kind of process I see go on far too often:
1. In anticipation of a major campaign, a school will spend a considerable sum on an
electronic screening of several thousand alums by one of the excellent firms that
do that kind of work.
2. The firm will return a huge amount of data to the school, most of which is
designed to assess an alum’s capacity to make some type of major gift to the
school. (Sadly, there is a good chance this data will not be used as part of the
campaign. Why this is so often the case is a bit of a puzzle to me. But it happens
far more frequently than any of us would like.)
3. In the meantime the school will not delve into the huge store of data it keeps on
its alums. Data that can point an accurate finger at which alums are most likely to
give and which are least likely to give. Class reunion attendance is a good
example. Unfortunately, most schools don’t store this information in their alumni
databases. Those that do store it don’t use it to help them identify new major
giving prospects — even though reunion attendance may well be the best single
predictor of giving in higher education advancement. (At least I’ve never found a
better one.)
Okay. Let’s say the scenario I’ve offered here is not atypical: A school isn’t going to do
an electronic screening, or it’s going to ignore the screening data even if it does one.
Beyond that, the school is not going to do any serious mining of its wealth of alum data
to find good predictors of giving. Can we then come up with a reasonably quick and easy
way for the school to still identify good prospects for major giving?
The bet I was making
I think the answer is an unequivocal yes. Why? Because for the last several years I have
become more and more convinced that if you know an alum’s age along with a few other
pieces of information about his/her marital status and contact information, you can do a
pretty good job of determining whether that alum is a major giving prospect?
Does that sound farfetched? All I ask is that you withhold judgment until the end of the
paper.
How I did the data analysis for the paper
For each of the four schools I mentioned above, I took these steps:
One. I gathered a representative sample (no less than 5,000) of alumni records for these
fields:
Home phone listed (yes/no)
Business phone listed (yes/no)
E-mail address listed (yes/no)
Listed as something other than “single” or “missing” for the marital status (yes/no)
Preferred year of graduation
Two. I “guesstimated” the age of each alum by subtracting their preferred year of
graduation from the present year and then adding 22 – the age that a lot of students
receive their undergraduate degrees.
Three. I computed the percentage of total lifetime dollars by alums in each five year age
interval.
Four. For every alum I created a simple score with a range from 1-5 based on whether or
not that alum had a home phone listed, a business phone listed, an e-mail address listed,
and whether or not the alum was listed as something other than “single” or “missing” for
the marital status field. (For a further explanation of how this kind of score works, see the
paper titled “A Simple Score.”)
Five. For the five year age interval with the highest percentage of lifetime giving, I
computed the mean (average) lifetime giving for every alum by simple score level.
Let’s see what these steps yielded for School A. Take a look at Chart 1.
0.0% 0.1% 0.3% 0.8% 1.0%
3.2% 3.4%
6.1%
12.5%
34.0%
20.5%
18.1%
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
35.0%
25 AND
UNDER
26-30 31-35 36-40 41-45 46-50 51-55 56-60 61-65 66-70 71-75 76 AND
OLDER
CHART 1: SCHOOL A. PERCENTAGE OF LIFETIME GIVING BY FIVE YEAR
AGE INTERVALS
Just to be clear, here’s how I constructed this chart. I used a sample of over 20,000
records from this school. For all of these records I computed the total amount these alums
had contributed to the school at the time the sample was gathered. This value exceeded
100 million dollars. I then computed the total amount that alums in each five year age
interval had contributed and converted these amounts to percentages of the total amount.
For example, 34% of this lifetime amount of more than 100 million dollars had been
contributed by alums aged 66-70.
A couple of things stand out for me about this chart:
• Well over 80% of the lifetime giving comes in after alums reach the age of 60.
• The under 50 crowd has contributed less than 10% of the total lifetime giving.
What do these facts say about where the focus ought to be on an upcoming capital
campaign? We’ll talk more about issues like this a little later on.
Now let’s look at Chart 2 which shows the mean (average) lifetime giving by simple
score level for this group of alums aged 66-70.
$176
$1,694
$3,500
$13,323
$55,732
$-
$10,000
$20,000
$30,000
$40,000
$50,000
$60,000
1 2 3 4 5
SCORE
CHART 2: SCHOOL A. MEAN (AVERAGE) LIFETIME GIVING FOR ALUMS 66-70
BY SIMPLE SCORE LEVEL
Here’s the algorithm I used for this score: ‘HOME PHONE LISTED’ + ‘BUSINESS
PHONE LISTED’ + ‘EMAIL LISTED’ – ‘SINGLE’ + 2. It’s pretty simple:
• If an alum had a home phone listed, they got a 1 otherwise a 0.
• If they had a business phone listed, they got a 1 otherwise a 0.
• If they had any kind of e-mail address listed, they got a 1 otherwise a 0.
• If they were listed as “single” in the marital status field, they got a minus 1
otherwise a 0 (people who are listed as single in a martial status field invariably
give less than any other marital code).
• Then, for each record, I added these numbers together and added a 2 to each
number so there wouldn’t be any zero or negative scores. (Zero and negative
scores tend to confuse people.)
So what we’ve got here is a very narrow age group of alums (66-70) who’ve given a huge
amount of money to the school. Beyond that we’ve got a very simple score that sharply
differentiates theses alums with respect to how much they’ve given.
What conclusions can we draw from the data analysis?
Let’s assume that the simple score is a rough measure of likelihood of giving to the
school. Let’s call it affinity. And let’s say we isolate the alums in this group that have a
score of 5 on this affinity scale. As it turns out there are just over 450 alums in this group
that is between 66 and 70 years old and have a score of 5. Okay, now let’s look at the top
ten lifetime givers in this group in descending order of giving:
$ 5,364,619
$ 5,038,707
$ 4,072,701
$ 1,644,404
$ 1,466,562
$ 979,197
$ 520,501
$ 496,978
$ 479,202
$ 256,006
What do you think? I think all these people are under stewardship – or at least they
should be. You certainly don’t need to do a wealth screening on them. Right?
All right, now let’s look at the next ten alums in this group in descending order of
lifetime giving?
$ 157,439
$ 110,621
$ 107,475
$ 88,834
$ 67,281
$ 60,017
$ 49,728
$ 41,825
$ 38,154
$ 35,728
Now it starts to get kind of interesting, doesn’t it? Finally, let’s look at the next ten alums
in this group in descending order of lifetime giving.
$ 35,287
$ 35,202
$ 33,186
$ 32,190
$ 28,493
$ 28,366
$ 27,225
$ 25,401
$ 21,826
$ 21,359
You’ve probably anticipated me, but here’s where I’m going with this. As we get further
and further down this list of 66-70 year olds who show high affinity to the school, we run
out of people who are already on our radar screens, who are already assigned and being
actively appealed to for a major gift. But I think a lot of these alums should be on our
radar screens. So what if we don’t have wealth screening data on them? They’ve given
something more than a pittance to the school already. They are by far in the highest
giving age bracket of alums. And they have an affinity score that says they like to give to
the school. Somebody should be reaching out to these folks. Why? Because some of them
are poised to make a major gift. It’s that simple!
And how much did it cost us to identify them? A few hours of time from a talented IT
person? I think it’s worth it.
At this point, I’ve either got you seriously interested in my line of reasoning or I’ve lost
you. If I haven’t lost you, let me show you the equivalent of charts 2 and 3 for the
remaining three schools whose data I looked at to prepare this article. Then we can step
back and get some perspective on all this.
0.1%
0.7%
1.4%
2.4%
4.4% 4.6%
6.6% 6.8%
8.7%
10.6%
25.0%
28.8%
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
25 AND
UNDER
26-30 31-35 36-40 41-45 46-50 51-55 56-60 61-65 66-70 71-75 76 AND
OLDER
CHART 3: SCHOOL B. PERCENTAGE OF LIFETIME GIVING BY FIVE YEAR
AGE INTERVALS
$67 $662
$8,794
$15,880
$59,809
$-
$10,000
$20,000
$30,000
$40,000
$50,000
$60,000
1 2 3 4 5
SCORE
CHART 4: SCHOOL B. MEAN (AVERAGE) LIFETIME GIVING FOR ALUMS 76
AND OLDER BY SIMPLE SCORE LEVEL
0.0% 0.2%
0.9%
1.3%
2.9%
7.1% 6.7%
12.0%
9.1%
18.3%
23.0%
18.6%
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
25 AND
UNDER
26-30 31-35 36-40 41-45 46-50 51-55 56-60 61-65 66-70 71-75 76 AND
OLDER
CHART 5: SCHOOL C. PERCENTAGE OF LIFETIME GIVING BY FIVE YEAR AGE INTERVALS
$- $4,554
$25,016
$132,414
$63,280
$-
$20,000
$40,000
$60,000
$80,000
$100,000
$120,000
$140,000
1 2 3 4 5
SCORE
CHART 6: SCHOOL C. MEAN (AVERAGE) LIFETIME GIVING FOR ALUMS
71-75 BY SIMPLE SCORE LEVEL
0.2% 1.1% 1.4%
3.4% 3.8%
7.0% 6.6% 6.0%
2.7%
12.9%
43.4%
11.5%
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
35.0%
40.0%
45.0%
25 AND
UNDER
26-30 31-35 36-40 41-45 46-50 51-55 56-60 61-65 66-70 71-75 76 AND
OLDER
CHART 7: SCHOOL D. PERCENTAGE OF LIFETIME GIVING BY FIVE YEAR AGE
INTERVALS
$- $609 $557
$77,553
$53,877
$-
$10,000
$20,000
$30,000
$40,000
$50,000
$60,000
$70,000
$80,000
1 2 3 4 5
SCORE
CHART 8: SCHOOL D. MEAN (AVERAGE) LIFETIME GIVING FOR ALUMS
71-75 BY SIMPLE SCORE LEVEL
That’s a lot of charts and data to look at. But let’s go back to the basic question I was
trying to answer. If a school doesn’t do an electronic screening, can we then come up
with a reasonably quick and easy way for the school to still identify good prospects for
major giving?
Again, I think the answer is yes. Why? Let’s summarize the logic I’m invoking from the
data I’ve presented from these four schools:
• Most of the lifetime giving in a school doesn’t start rolling in until alums reach at
least the age of 55. So if you’re spending a lot of time looking for good prospects
under that age, I don’t think you’re making great use of your time.
• Within this older crowd of alums, you don’t need a lot of information beyond
home phone listed, business phone listed, etc. to create a very simple score that
sharply separates the big givers from the small or non-givers.
• You can use the top end of that simple score to identify alums who aren’t yet on
your development officers’ radar screens but who should be.
Now I’ve already said that the last thing I’m doing here is recommending that your
school not do an electronic screening. Not at all. But I am saying that if you can’t do a
screening, you have an option here that can help.
Some specific “action steps”
And finally, what about all you prospect researchers? After all, I wrote this thing for you
guys. Here’s a suggestion on some steps you can take:
1. Get your IT folks to do the same kind of analysis I’ve done here for your own
school. Sure, you’ll get some push back from them. They’re overworked. But you
can get them to do it for you.
2. Look up those alums in a higher age category who have a high simple score but
who aren’t yet assigned to a gift officer. Some of them will look very promising. I
guarantee it.
3. Take a few of the promising ones to a gift officer who thinks this data driven
decision making stuff is sort of cool and ask him/her to contact these alums.
What do you have to lose?
« Newer Posts

Create a free website or blog at WordPress.com.