Decision Science, Prospect Research & Management

Today nonprofits of every mission and size continue to leverage analytics to inform individual constituent persona behavior, predict outcomes of campaigns, and assess program performance. The benefit of fundraising analytics for them is foundational: apply individual-level insights to large populations. Any fundraiser could filter through a list of 100 known names and identify top prospects, but our brains are not correctly wired to filter through 10,000 or 100,000 names.

Broaden Your Focus

A majority of the analytics in practice today contains a micro-level focus: projects targeted towards individuals or small groups predicting independent outcomes. This approach can often limit insights into greater opportunities to gain efficiency and best align resources. In short, focusing on only key top donor segments or individuals provides limited opportunities to harvest all available insights and impact within a nonprofit’s database.

There is much to be learned from all types of constituents: top donors to non-donors, alumni to friends, patient families to event-only givers. Every constituency has “tribes,” and greater understanding of who they are, what makes them unique, and how to best utilize resources to engage them can be transformational.

Define Your Personas

There are two fundamental analytical techniques to highlight distinct personas in any constituency, which can then be used to segment and filter to support a variety of engagement and fundraising efforts. While both persona approaches offer strengths and have some gap areas, they are realistic options no matter your database size. These two approaches also align with the two fundamental types of model building: supervised and unsupervised.

Supervised indicates that we get to decide what the distinctive factors should be—what we value most, what we may observe or believe to be distinctive, assign them different weights etc. Unsupervised allows for the data to sort itself out, letting constituents group themselves by demographics, engagement, giving, and other activities that reveal unique “tribes” within your constituency.

Persona Score: Descriptive and Supervised

This is the most commonly developed persona score in fundraising analytics today. It is descriptive because it seeks to describe a constituent record’s profile based upon tracked and recorded data points. It is supervised because an analyst, or a small working group, or perhaps even leadership have predefined behaviors and outcomes they consider distinctive and that are modeled via data from your database. For example, many higher education institutions value alumni greater than past parents, but past parents who are also alums are greater than both non-alum parents and past parents. These relationships can be easily modeled through data.

The benefits of this approach are in the weighting and selection of criteria; organizations can measure and track what they value. The gap of this approach is we don’t know, what we don’t know. Every database has distinguishing data points that may not be natively recognized but may prove helpful to assigning tribes.

Clustering: Predictive and Unsupervised

This is the less widely used approach, yet it offers some unique advantages to common persona or connection score. It is predictive, as it employs any manner of clustering algorithms (K-means, two-step, Kohonen, etc). It is unsupervised since the algorithms let the data, ergo the constituents themselves, group and align by the characteristics that can best segment, not by an analyst, small team, leaderships opinions, or judgements around what are distinctive behaviors.

The benefits of this approach are the observational qualities: your constituents essentially “self select” themselves into distinct tribes. This often allows organizations more flexibility in application as some key data elements (such as responsiveness to channel or a demographic status) may not be statistically significant in a cluster, allowing to overlay and say, “Tribe A responds to this channel, Tribe B does not.” The gap of this approach is many of the data points that organizations value may not be distinctive (alumni status for example), and this can create dissonance for broad integration and implementation.

Become Your Organization’s “Anthropologist”

Creating an understanding where all constituents fall within your organization’s database has almost limitless application and value. The “least” often has as much business value as the “best” from any measurement (giving, activity, etc.) allowing even greater focus on large populations and resource alignment to where application and energy will resonate.

BWF has uncovered and recorded the tribes for hundreds of nonprofits of every mission and size, and we welcome the opportunity to visit this customized approach with your constituency.

Want to learn more about leveraging analytics and uncovering personas within your constituency? Contact Alex Oftelie.