Data Analytics: The Gift that Keeps Giving

Amnesty International Belgium - Flemish Section streamlines fundraising with automation and predictive analytics for more effective donor engagement.
Customer Experience

Non-profits have faced myriad fundraising challenges in recent years and numerous industry reports outline declining gifts and donations. To meet those challenges, non-profits must maintain close relationships with prospects and donors. In the absence of data analytics, however, those challenges intensify as fundraisers blindly search for new donations.

Amnesty International Belgium -- Flemish Section (AIB), a branch of Amnesty International, a nonprofit organization of more than 7 million people who advocate for human rights, relies heavily on donors (which account for 93 percent of its annual income) to help fund its campaigns for refugees, reproductive rights, and more.

Like other chapters, AIB raises funds through telemarketing, petitions, email campaigns, and mailers. When the organization recently wanted to streamline its fundraising efforts, it discovered the critical need for an analytics tool to further its outreach efforts.

Ilja De Coster, senior fundraising consultant at AIB, noted that the chapter's fundraising efforts were hindered by operational inefficiencies and a lack of data analytics. "One of the issues was with calls for reactivating donors," De Coster explains. "The lists of who to call were updated manually and there was too long a gap between when the lists were made and when agents called people. Sometimes the donors had already reactivated by the time they got a call, which is a waste of resources."

The organization also struggled to get insights into its fundraising performance since there was often a delay in working with programmers to produce the data analytics. De Coster searched for an analytics tool and found RapidMiner, a self-service platform that offers data mining and predictive analytics.

Two years ago, De Coster integrated RapidMiner with the organization's CRM database to begin streamlining its data on active and inactive donors. AIB's database includes about 20,000 active donors and 10,000 inactive donors. The organization's first goal was to clean up its database and get a better understanding of how many donors they had and how many were dropping off, De Coster explains.

The organization discovered that roughly 10 percent of a call list for reactivation campaigns included duplicate names or people who had already reactivated their memberships as donors. AIB also automated the process for updating its call lists. Prior to adopting the new platform, the reactivation call list was updated about every other month but now it's updated on a daily to weekly basis.

"Our marketing department was often frustrated because they were spending a lot of time on manual work and couldn't get to more creative things," De Coster notes. "Now they have more time and the improved reporting makes it easier to track our results." As for the impact on donations and donor engagement, the organization has tested a few sample groups but it's too early to extrapolate the results to a larger group, according to De Coster. "Donor retention is something that we need to monitor for at least three years before we have a real sense of the impact," he says.

The organization is also experimenting with predictive models. AIB is analyzing the data that it has about donors including their age, gender, donation size, their level of engagement and combining it with publically available demographic information like household incomes to identify potential donors, changes in donation behavior, and churn predictions.

Predictive modeling is an "ongoing process" that is part of the organization's efforts to manage its daily fundraising operations while finding new ways to improve its engagements with donors, De Coster explains.

"We're always trying to find a balance between experimenting with new approaches and giving those things time to work," he says. "Instead of trying to do everything at once, we're slowly taking the stuff that we're testing and doing it on a bigger scale."