Selling hope

As one industry leader puts it, “[development officers] aren’t selling a product; they’re selling hope.” Selling hope requires a lot of finesse. On the one hand, there’s the advantage of marketing a “product” that furthers social good. On the other, there’s the challenge of soliciting today’s money for tomorrow’s improvements. And that’s not calculating for the additional challenge of marketing what amounts to an invisible purchase (there’s often no immediate or tangible product in return).

Selling hope also requires fortitude and stamina. It can take years to nurture a meaningful relationship (and ergo gifts). And non-profit budgets are invariably lean, meaning small development teams are often charged with delivering large—and operationally crucial—gifts.

Given these challenges, how can today’s fundraising professionals best optimize their time and donations? Hint: it’s not by adding more staff. Rather, non-profits are increasingly turning to the same techniques being used by their corporate cousins, namely data mining, automation, and predictive modeling. A few recent use cases:*

1. Smaller, smarter portfolios

Until recently, some development professionals at the University of Iowa (UI) were managing as many as 120 donors or prospects. "Trying to keep track of that many people was beyond out-of-control," says Janet Weimar, associate director of the University’s Center for Advancement.

To get a handle on the workload, the team first analyzed five years of giving data (focusing on gifts between $50,000 and $5 million) and catalogued individuals in one of three ways: "qualified," "actively cultivated and solicited," and those “being stewarded for gifts already made”. They determined that there was a lot of deadwood on the lists and culled about 1,700 names. Some fundraisers’ portfolios shrank by two-thirds, allowing them to use their time to cultivate those supporters with the greatest potential.

2. Untapped data gold mines

UI’s development staff also looked outside their own data sets to identify other valuable information across the organization. For example, they found that the university’s 1,800-seat Hancher Auditorium (which hosts Broadway shows, dance recitals, orchestral concerts, and lectures), had years of ticket-sales data had never been analyzed. The team subsequently sorted through several years’ worth of sales data, categorizing ticket buyers by number of performances attended and the amount spent. For example: Did they pay more for front- row seats versus in the balcony?

They found that many of those who frequently bought premium tickets also gave generously to fundraising efforts. "This ticket data was really somewhat of a lead indicator of how much somebody was giving," says Brad Cunningham, the development office’s head data-cruncher.

As a result, some names were added to prospect lists, and some were dropped. Now as the team interacts with various university entities, he says, they look for people "sitting on an Excel spreadsheet" or any other untapped source of information on public engagement with the university.

3. Small staff, huge help

As we touched upon earlier, the cultivation of donors demands finesse and a high degree of empathy. But too light of a touch can compromise your organization’s ability to demonstrate interest in its supporters.

Such was the case for a Denver, Colorado based non-profit that partnered with HAI to gain a better understanding of the factors that influence repeat donations. Like most development offices, they shied away from requiring donors to provide any information other than name and an email or street address, in order to make giving as easy as possible. This customer-focused strategy left the client without a clear understanding of their supporters’ demographics, giving history, and possible motivations. In addition, it was very difficult to match multiple donations to a single contributor. Because donations were accepted via multiple formats, it would have been an arduous task for their development staff of one to comb through various lists in an attempt to make correlations.

Working closely with the client, HAI was instead able to create an algorithm that brought together all donations by an individual donor into a single record. Even better, the cleaned and organized data was paired with third-party geodemographic information. Hundreds of new variables (from average apparel and book spend, to neighborhood demographics, and beyond) were merged with the client’s internal data, which enabled HAI to build a predictive model three times more accurate than rules-based sorting. Its insights are now being used to prioritize outreach efforts, a boon to the foundation’s development officer. “As much as I’d like to be able to, I’m only one person and there’s just no way I can reach out to every potential donor,” she says. Now that she has lead scores generated with advanced predictive modeling, she doesn’t have to.

*Use cases #1 and #2 from The Chronicle of Philanthropy’s article, “How a University Used Data Analytics to Improve Fundraising – and You Can, Too”, by Brennen Jensen.