Likelihood Scoring for Fun and for Profit

Dear readers, I am sure you will forgive my inattentiveness to this blog as I prepare for a big day next week. Now, I must be off to continue some likelihood scoring to help me sort through a pile of potential puzzle solutions...

And if you ask, "Why, what do you mean by likelihood scoring?", then read on:

When you have a large list of possible outcomes, say answers to a puzzle, or potential donors in a geographic area, it is helpful to devise a system to narrow that list down. In other words, you are trying to find out how likely a particular potential outcome is to be true.

You can do this through statistical data modeling or by building a more crude score based on gut instinct. Here are some pointers for building a crude likelihood score.

Let's think about a list of potential donors, and how we might build a likelihood score to predict who is the most likely to give a major gift. You want a combination of ability to give a major gift (wealth) and desire to give a major gift (inclination or affinity toward your organization).

Let's take wealth indicators first. What do you have in your database, or what might you easily append to the info you already have? (See this post for some tips on how to join two Excel lists using the VLOOKUP function.)

Consider:

• Capacity ratings -- in-house or purchased from a vendor
• Geodemography based on home address -- use a wealthy zip code list, or Social Explorer
• Previous major giving activity

Now think about inclination/affinity:

• Recency of last gift (I know, recency is not a real word.)
• Frequency of giving
• Volunteer involvement with your organization
• Attendance at your organization's events
• Other likelihood scores created through in-house data modeling or purchased from a vendor
• Season ticket holders
• Alumnus or other status indicating a prior close relationship with your organization

Turn each indicator into a yes/no question, e.g. "Has this person donated within the last two years?" Assign each positive indicator a point value, e.g. yes = 2 points. Total the points, and you have just built a raw likelihood score that you can use to rank a prospect list.

For extra credit, consider building an RFM score and using that as a piece of your likelihood score.

Please do note that this kind of scoring is in no way scientific. It's just a very simple way to quickly prioritize a list of prospects.