In recent years, we’ve seen an emergence of innovations to identify new prospects for giving to healthcare. And while AI is one of them, it is not the only one. Studies of donor breadth of connection, community building, social norming, and psychometrics have all opened up new possibilities.
Looking Back
Private support has been a hallmark of healthcare excellence since the Mayo brothers formed the Mayo Foundation for Medical Education and Research. As we know, patients will give back in response to positive experiences. They will also support research for eliminating diseases and improving health for their communities. Although not every patient will give. And perhaps not every patient should be asked. But knowing which patients would find joy through a philanthropic relationship is a responsibility of the healthcare fundraising community.
Early efforts into grateful patient fundraising were based in mass solicitation techniques. After receiving care, a patient or patient family would receive a letter or phone call seeking a first gift. This would begin a long relationship, hopefully growing in financial value and donor experience over time.
As multi-source wealth screening emerged in the 1990s and achieved efficiencies of scale to enable up to daily screening, many organizations shifted to a screen-first approach. Following screening and depending on wealth levels, some names were given to relational fundraisers while others received mass appeal treatment.
As we all know, wealth does not equal propensity to give. So, there was little discrimination about who should or shouldn’t be approached. This was a limitation that was somewhat addressed with the emergence of predictive science in the 2000s. Using classification or even simpler point-based measures of potential gratefulness or propensity gave a countermeasure that improved outcomes. The efficacy was realized through decreasing the cost and effort spent on those unlikely to engage while increasing efforts to reach more likely candidates.
During this time, we also saw more of an emergence of assistance from healthcare providers. Equipping physicians or other caregivers with mechanisms to give fundraisers names of happy patients or patients who were asking to give back led to increases in higher dollar leads.
Limitations
Up until the last five years, screening, a propensity measure, and a handoff system have been the primary means of lead identification. But there continued to be limitations for many programs:
- Affinity was defined by behaviors common to people who gave in the past. The analytical mechanisms did not actually uncover what was in the mind.
- Measurements were taken at snapshots in time. But affinity naturally increases and decreases over time and based on experiences.
- Predictive science was only used at the initial lead generation stage with fewer innovations in the early pipeline stages.
- Research into donor breadth and community effect, while applied somewhat in cultivation and donor relations, was not really applied at the lead generation stage.
Innovations to Accelerate Prospect Identification
Healthcare lead identification is on the verge of an evolutionary leap in addressing all four of these limitations. There are a few early adopters already rising to the occasion. Here are four ways they are doing it:
Psychometrics. This field of study concerns what is in the mind and what is the impact on behaviors. Sentiment surveys of constituents, when using certain types of framing, can capture how a constituent feels. When conducted with regularity, shifts in affinity can emerge, although the limitation of surveys is in how few people actually respond. The measurements can become the target of AI models to both expand how others might feel and connect them to external influences that change affinity.
Machine Learning. Snapshot predictive models have always had the capability of incorporating many data inputs. These might be patient experience data, giving data, third-party data, and psychometrics. But the gathering and connecting was often laborious. Additionally, their very nature left long spans of time between measurements. Now, targets such as “giving a gift” or “volunteer likelihood” can be incorporated into continuously improving models connected to multiple inputs. Scores could change daily if it makes sense, or at least every time new prospects are needed.
Pipeline Innovation. Machine learning models and other generative tools can be used to speed up the verification of leads in advance of assignment, identify likelihood to take a call, categorize philanthropic interests, nudge officer interactions, and help prepare materials. Predictive models were like printing maps from MapQuest before the fundraising journey. We now have GPS.
Community Effect. In studies of drivers of lifetime giving and major gift giving, we’ve consistently seen the breadth of connection a constituent has with the organization and with other constituents as among the top two drivers. Donors respond well to being social normed with groups of like people. In other words, “I am one of the people that does this.” Or as the marketing author Seth Godin has often said, “People like us do things like this.” Dynamic segmentation of patients into like groups—not necessarily demographic, but more effectively behavioral and psychometric—can better align people to engagement journeys. Also, leveraging donors to make friends with other donors and more people on the fundraising staff can prove to be a very non-technical innovation that could boost early engagement efforts.
These tools and innovations are here today. It is an exciting next step on the journey for excellence in healthcare fundraising. And it is one we would love to walk with you. Let your friends at BWF know if you’d like to incorporate some of these approaches into your program. We’ll do the heavy lifting for you.
