How Data Mining Can Increase Phonathon Acquisition

illustration of an article


This is the second in a series on donor acquisition by Jessica Neno Cloud, CFRE, the author of Successful Fundraising Calls: A Phonathon Scripting Workshop. Cloud is the assistant director for fundraising initiatives and planned giving at the University of Southern Mississippi Foundation. She practices evidence-based fundraising with a focus on return on investment, and has a variety of innovative and effective techniques to share.

You may also be interested in the first article in this series: “How Data Mining Can Increase Direct Mail Acquisition.”

The leaky bucket is a common metaphor in annual giving. Organizations will not retain 100% of donors from year to year. In fact, most institutions have retention rates far below that. Without acquisition work to bring on first-time donors, most of our organizations would flat-line within a short period of time.

However, acquisition work is expensive. The best acquisition results usually come from the phonathon program, where conversations can be tailored to the prospect’s interest and have a human touch.

Acquisition mailers usually have poor results, but surely there is some low-hanging fruit in your database that would prefer a mailing. (To learn more, read our article on “How Data Mining Can Increase Direct Mail Acquisition.”).

There are also those stalwart non-donors out there that we may want to avoid and not waste precious acquisition resources trying to convince.


Many people use the term “data mining” interchangeably with “segmentation.” While they are related, segmentation and data mining are not the same thing.

Segmentation involves breaking your entire pool of prospects into smaller groups, based on demographic or giving similarities, in order to utilize targeted messaging or for testing and statistical analysis purposes. For example, breaking your fall mailers into groups based on donor type (lybunt, sybunt, etc.) or by college in order to use different letter versions or so that you can report results easily to on-campus clients (or both).

Data mining involves using data points in your database to limit who is solicited, thereby saving costs and improving effectiveness. This can also include using information in your database to go beyond segmentation to create customized ask ladders for prospects within segments in order to maximize returns.

Data Mining for Phonathon Acquisition

I recommend setting up future donor segments to be as large as possible so that the call center manager has the greatest amount in flexibility in creating smaller calling pools underneath.

Typically, phonathon does not have the capacity to call all future donors segments to 65 or 70% complete, so data mining is necessary to decide how best to “spend” those completed calls in each acquisition segment.

Let’s say you have a large Business Future Donors segment with 5,000 records in it. So far you have 500 completed calls (10% penetration rate). Now you have a sample to work with. Let’s say the plan is to take this segment to 30% complete. How should you “spend” the remaining 1,000 completed calls you planned to make (out of a total of 1,500)?

If you are on an automated system, it is likely the software program has a suite of analytical reports that you can run on past and current calling projects. This is the information gathering phase of your data mining. Run every report you have available on only that Business Future Donors segment. I would run a majors report, a ZIP code report, and area code report, a state report, and a grad year report.

Once you have this information on calls you have already done in this segment, grab a highlighter and look at the participation rate on all of these different reports. Let’s say your overall participation rate so far in the segment is 6.75% and you see that alumni in this segment with marketing degrees are running at 8.9%; highlight that. You might find 10 area codes and 4 ZIP codes that are performing above 10%. Or three targeted grad years that are doing similarly well. Highlight all those trends.

Write down all of these data points that you highlighted. This will be your criteria for a query of “Business Future Donors low hanging fruit.” You’ll use this to create a calling pool of prospects underneath the broad segment. I always called these “Business Priority Future Donors” for two reasons: 1) They are the best prospects for the rest of that segment this year and 2) Callers love knowing that they are calling primo people.

Caveat: If you see that Business Future Donors alumni in Washington, DC are running 50% but there are only 2 of them left to call, it is probably not worth your time to add those to your “low-hanging fruit” query. Remember you have 1,000 more calls to complete. So you will need to construct a group of 1,500 – 2,000 prospects. But let’s say that Louisiana is running 25% for this segment and the count 450 prospects available for calling. Add that to your list.

Now, for the rest of your calls in this segment you might run 9.5% rather than the 6.75% that you did during the first 500 calls. This will bring your overall participation for the segment up. You’ll acquire more donors while saving calling hours.

One more caveat: Because phonathon is very dynamic, it is best to use test data from the current year to do your acquisition data mining. You could do this before the year begins, using for your analysis the same reports but on last year’s Business Futures segments. However, keep in mind that you acquired the low-hanging fruit from that group already. So while Florida prospects might have been where it was at last year, all of those folks are now in your First-time lybunt segments this year not future donors. Therefore, Florida may not be a hot-spot for acquisition anymore.

Once you understand the theory and strategy behind this kind of data mining (creating targeted calling pools for phonathon) you can pull in new data points into the calling program and let those inform your calling pool creation. If you have your database screened for Annual Giving Likelihood or projected annual gift capacities, you can add those in flexible data fields in your calling software for all prospects.

Once that data is in the software, early results can be analyzed by those likelihood ratings and if they correlate, you can create calling pools based on ratings in tandem with other information to make those calling pools even more targeted and powerful. To go one step further, gift capacities can be used to drive ask levels for prospects so you can be sure you aren’t under-asking. For example, you might have a “Priority Business Futures High End” and a “Priority Business Futures Normal Ask” pools with differently scripted ask amounts. With this strategy, your remaining calls will have a higher participation rate and average gift.


Read Jessica Cloud’s Book

Successful Fundraising Calls offers you a full scripting workshop and tutorial for successful solicitation calls. Author Jessica Neno Cloud critiques and revises 6 real phonathon scripts submitted by institutions across North America.

In this book, you will:

  • Learn the pillars of writing effective scripts
  • Review sample LYBUNT, SYBUNT, young alumni, and future donor scripts
  • Explore Cloud’s in-depth critique of the samples
  • View revisions of the sample scripts

The only resource of its kind, this monograph brings a low-cost scripting workshop directly to your office.

To order copies for your office — or to view a free sample: