How Customer Analytics Is Speeding The Cure For Parkinson’s

It’s no secret that large companies use customer analytics to inform their business strategies. Retailers and financial institutions are mining and analyzing data (the proverbial Big Data) about us all the time: personal information generated by our credit card purchases, loyalty programs, online behavior, magazine subscriptions, memberships and survey responses. They use these insights for everything from tailoring marketing campaigns to planning product launches and new store locations. Meanwhile, researchers and developers are offering these firms increasingly potent predictive models and software platforms.

Parkinson's Disease
Image source: Wikimedia Commons
Parkinson’s Disease

But today’s analytical techniques are still making their way through the non-profit world, as Emily Moyer, senior vice president of marketing and digital strategies at The Michael J. Fox Foundation for Parkinson’s Research, discovered when she joined the organization 18 months ago. Moyer had arrived with 15 years in corporate media management under her belt.

“I was a little surprised to find… how steep the imbalance was in analytical proclivity across the foundation,” she said. On the medical research side, PhDs and scientists in search of a Parkinson’s cure worked confidently with data mining and predictive analytics tools. But, “there was nearly none of that happening on the business side … in the fundraising and marketing pieces.”

Moyer set out to change that, as she described to the audience at the recent Wharton Customer Analytics Initiative Conference.

A Niche Market in Philanthropy

Co-founded in 2000 by well-known actor Michael J. Fox, who was diagnosed with young-onset Parkinson’s at the age of 29, and Deborah W. Brooks, a former Goldman Sachs executive, the New York City-based organization fundraises and “aggressively funds a global research agenda focused on finding a cure,” according to Moyer. The foundation’s website states that it is the world’s largest private funder of Parkinson’s disease research.

But not everything about the foundation is on a grand scale. It has a lean staff — only about 140 employees — and the small size of its prospective donor market makes for what Moyer called “one of the most difficult marketing challenges I’ve ever taken on.”

“Suddenly overnight we were able to have, on our desktops and our mobile devices, real-time automated reports. It was game-changing for senior management.” –Emily Moyer

Moyer explained that about 1 to 2 million people in the U.S. population have Parkinson’s, a number small enough to classify it as a rare disease. She noted that there is assumed to be a “friends and family” circle of three or four people around any given Parkinson’s patient, bringing the organization’s total potential donors to around 4 or 5 million. Moyer characterized this as a “tiny, tiny market” compared with the customer numbers many commercial ventures are dealing with.

She estimated that the foundation’s CRM database has identified about 15%-20% of the market, but said staff are “working on substantially improving that penetration…. Over the past 18 months we’ve been exponentially increasing our digital footprint to capture more of those people.”

Moyer pointed out that in addition to bringing in donors, the foundation also tries to attract clinical trial participants, an urgent lack in this field. “We do need people’s dollars, but we also need relevant people to participate in clinical trials so we can find a cure, and so we can develop statistically significant results.”

Predicting Donor Behavior

Over the past few months, the Michael J. Fox Foundation has assumed a new focus on customer analytics that is helping it more effectively reach both ends of its fundraising spectrum, Moyer adds. With the larger portion of donors, the foundation is looking at how to optimize and scale “low-touch, high-scale digital marketing efforts.” And with the smaller pool of major donors, they are looking at how to leverage donor data “to optimize our high-touch, personal outreach, relationship-building exercises.”

“For each donor that we’re working with, we can give our fundraising and marketing teams information about the likelihood that those constituents will return.” –Luba Smolensky

One major step Moyer took was hiring associate director of analytics Luba Smolensky (who also spoke at the conference). Moyer commented that Smolensky had the technical know-how to work with statistical tools such as R — a now widely used programming language and software environment for statistical computing and graphics –and was able to implement a business intelligence platform, a first for the organization. Smolensky’s work replaced “time-consuming, heavy-labor downloads from the CRM database, manipulating data in Excel, and pivoting every time we needed to see something,” said Moyer. “Suddenly overnight we were able to have, on our desktops and our mobile devices, real-time automated reports. It was game-changing for senior management.”

Smolensky explained that she was moving the organization away from “a standard rules-based approach” to donor data to a more probabilistic way of thinking. With the rules-based approach, the organization had divided donors into four categories based on predetermined donation thresholds. This type of analysis provides limited information, she said. “Most folks just fall into the same group [they did before]; we don’t really see a lot of upward movement or even a lot of downward movement.”

But a more probabilistic model, said Smolensky, allows fundraisers to begin to answer critical questions such as “Will certain donors return? How much will they give? How often will they give?” In her work, Smolensky draws on customer base probability models, citing in particular those developed by Wharton marketing professor Peter Fader along with Bruce Hardie of the London Business School.

She talked specifically about how she has calculated “drop-off probability” — the likelihood of individual donors staying or leaving — by analyzing the frequency and recency of their donation transactions. “So now we find ourselves in a position where for each person that we’re working with, we can give our fundraising and marketing teams information about the likelihood that those constituents will return.”

“On the high end of things … our lean development department doesn’t have enough time to reach out to the hundreds of people in their portfolio.” –Emily Moyer

Moyer and Smolensky agreed that this knowledge had put the organization in “a very different space” than they were in just a few months ago.

A side benefit of Smolensky’s work with the business intelligence platform, according to Moyer, has been the identification of underlying errors in the organization’s CRM. “[It allows] us to pinpoint where the data may be off, and go in and make tweaks.” The foundation has now implemented a 12- to 18-month plan to clean up its CRM data structures and data fields, which among other things will improve the quality of the data feeding into Smolensky’s analyses.

Smolensky also is providing three other outputs in addition to drop-off probability, using the same customer-base probability model. These include the expected number of donations from given donors, and the probability that they will make a certain number of transactions. The third will be to establish a “lifetime value” for each particular donor. These are insights into critical aspects of donor behavior, she said, that will benefit the marketing and fundraising teams in many ways.

Smolensky was asked if she was using data to figure out which small-dollar donors were likely to become larger-dollar donors. She answered that there are certain characteristics that one

1, 2  - View Full Page