The following excerpt is from Prospect Research in Canada: An Essential Guide for Researchers and Fundraisers, a recent top-selling publication by Hilborn's Civil Sector Press.

Analytics and data mining have many definitions, and most reference "exploring data to uncover meaningful patterns," or "insights to drive decision-making." Before deciding what is meaningful insight, it is helpful to clarify what question you are trying to answer or what decision you are trying to make. Otherwise you won’t know what findings are meaningful, and which are just noise. There are many interesting questions and avenues that can be pursued, but some of the more common and relevant questions to explore about your donors include:

• What are the characteristics of my highest-contributing donors? How can I use that information to identify other potential high-value donors? (i.e. major gift prospects)

•Which donors are most likely to give again next year?

• Which donors are most likely to lapse?

• What are the characteristics of my most loyal donors? How can this information be used to target future acquisition activities?

• Which of my non-donors (such as alumni, subscribers, members, etc.) are most likely to respond to a charitable appeal?

• What events or other factors appear to influence giving levels – up or down?

Each of these questions requires a different approach, and more importantly, a different way of defining which donors are "ideal cases" to investigate and research. One of the key premises in analytics is to analyze historical data (ideally recent history) to find patterns that can be used to predict and inform the future. To do the review, however, requires defining either a group, or specific variable to compare against other criteria to look for trends.

Here are three potential questions and suggestions for analytical approaches:

1. Which of my non-donors (such as alumni, subscribers, members, etc.) are most likely to respond to a charitable appeal?

Suggestion: First, determine "cases to study," in other words people from this group who have already given in response to an appeal – preferably a recent one. Pull a list of everyone in the main group (all alumni, all subscribers, etc.), along with one or more fields to identify the people you’re interested in (i.e. lifetime giving, or giving in the time period since that appeal).

2. What are the characteristics of my highest-contributing supporters? How can this information be used to identify other potential high-value donors (i.e. major gift prospects)?

Suggestion: Include all recent donors (i.e. who have given in the last three years or so. Major donors don’t necessarily give every year). Include some giving fields so that you can sort out which are the "top donors," such as total lifetime giving, and total giving in the last three years.

3. Which donors are most likely to give again next year? Which donors are most likely to lapse?

Suggestion: Analyze all donors who gave two years ago, comparing those who gave again last year to those who did not (renewed versus lapsed). Include total giving and/or gift count for each of the last two years to be able to define the two groups.

Getting to know your data

The best information about a constituent’s relationship with your organization is (or should be) in your database. This information is specific to the constituent, is presumably provided by the donor or someone close to them, is organized in some fashion for analysis, and – most importantly – is completely related to that individual or entity’s relationship with your organization.

At the very minimum in Canada, for individual donors you will have name, address, gift amount(s) and date(s), which are needed to be able to process receipts. There may also be some information about the gift payment type, what appeal or solicitation (or lack of) was attached to the gift, and potentially what the gift is intended to support.

Start simple, with some basic biographic data and summary-level giving information in one file/report to get your feet wet.

SAMPLE: Biographic data

• Donor ID: the unique identifier that will be used to link data, and should be on every individual file pulled.
• Donor name and title (Mr., Mrs., Mme., etc.)
• Home address and phone
• Business address and phone
• Email(s)
• Employment information
• Gender
• Age
• Marital status
• Spouse and other relationships

Be sure to include the relevant data fields needed to address your question. Don’t try to pull everything for your first project.

This article is excerpted from the chapter on "Data Mining and Analytics" in Prospect Research in Canada: An Essential Guide for Researchers and Fundraisers, edited by Tracey Church and Liz Rejman.