Editor's Note: This article is featured in Best of Connections 2023. Read Apra Content Development Committee Chair Jennifer Moody's editor's message to learn more about the top articles of the year. 

 

 

Last month, I was fortunate to attend the CASE DRIVE conference in Chicago. Attendees were from many corners of advancement: prospect development, annual giving, donor relations and stewardship, advancement services and events.

If you’re reading this, you’re likely aligned with prospect development — few of us have responsibilities pertaining to all these teams, but we all benefit from good data from these allied functional areas. As professionals whose success depends on leveraging technology to capture, analyze and disseminate insight, we need to pay especially close attention to emerging technologies. In case you had any doubt, the data science hype cycle is here.

The DRIVE exhibition hall was teeming with emerging and rebranded companies focused on how data and technology will transform advancement operations. We’ll see new use cases for artificial intelligence (AI) and its applications, such as ChatGPT and recommender systems, in advancement. Vendors will bake ChatGPT features into their existing products and develop novel uses for these tools. Tech-first vendors new to advancement will continue to emerge and we’ll feel their presence increase at our exhibition hall. In five years, I anticipate you’ll see many new upstarts in the AI vendor space bringing their products to non-profit advancement.

Today’s Data Science Opportunities

The most common application of predictive analytics in our line of work is called supervised learning, usually in the form of likelihood model scores for major gifts. Said another way, these scores provide information about which prospects are more likely than others to give a major gift, all else being equal. Twenty years after these products emerged on the prospect development scene, many organizations still choose to work with vendors and purchase scores, despite robust computing power, available information on statistics and free training in regression techniques.

Companies have made huge strides in the scalability of these predictive models, and it’s more common than not for these vendors to create bespoke scores that are custom to your organization’s data and constituents. Further, they’re offering ongoing retraining of models over the life of a contract. It’s never been easier for organizations to adopt customized models trained on their own data, as opposed to rules of thumb or model building leveraging data from other organizations.

AI with a narrow scope has found its way to consumers via ChatGPT, image recognition and predictive modeling. Generalized AI, however, is still in its infancy. While we’ve had success programming rules-based logic, generalized AI allows algorithms to evolve in how they react to new data — as the characteristics of a target set change, the algorithm adapts to integrate these changes and improve accuracy and precision.

For me, Rodger Devine’s presentation at DRIVE, “The Data Science Order of Operations and Hierarchy of Needs,” was the most approachable and most prescient session of the conference. Its premise is that too many organizations attempt predictive analytics and higher-order machine learning techniques without a strong foundation of data management, governance and integrity. The results are conclusions made and action taken without confidence in the collection of underlying data that the solution is built on. Instead, organizations need to be confident in the data quality, completeness and lineage before embarking on advanced analytics. It’s like skiing before you can walk.

The Build vs. Buy Debate

When a shop considers how it will practice advanced analytics, it’s common to have the build vs. buy debate for data science solutions. Should one develop data analytics capabilities in-house or purchase solutions through a vendor? For organizations without the necessary technical skills, working with an analytics vendor may be preferable.

Those with large, diversified, well-resourced shops may invest in developing data science solutions within their organizations. They may undertake exploratory projects that try to predict the readiness of a major donor to give a second major gift. They may classify proposals with a low likelihood of closing, signifying a need for extra touches to improve the proposal’s success. They may work on a clustering project of prospects to develop volunteer type personas.

Organizations that have the time and talent to invest in analytics projects see benefits from building versus buying. The benefits from building yield rich experiences, from building project management skills, to cross-departmental coordination and critical thinking. Be forewarned that many analytics projects ultimately fail to yield their intended impact, so if analysis emerges that can bring insight and new action, all the better.

Making the Data Investment

The build vs. buy decision becomes more challenging as we witness experienced analytics practitioners leaving the fundraising space, as salaries in the nonprofit sector often can’t compete with the private sector. As a manager, it may feel precarious to invest in the skills of a junior analyst, only to worry they’ll leave once learning new skill. At the same time, having team members that are skilled in R or other programming languages can unlock huge time efficiencies and insight.

We at Apra are your partner in identifying critical skills and techniques so that prospect development professionals can remain at the forefront of these new eras in our industry. Last week, Apra launched a new fundamentals course, Data Science With R.

This five-module course is for prospect development professionals who want to take the next step in their data science journey and learn how to use R, a programming language for statistical computing and graphics. As a contributor to this series, I can say that the instructors spent lots of time determining how to get learners comfortable with these concepts as fast as possible, what the most important concepts are and how to make this experience enjoyable and rewarding.

There are many free or low-cost opportunities to upskill and learn R, but I’ve heard from many that learning it in a fundraising context is preferable. Other free content uses domains like digital marketing or captured data from your phone’s gyroscope to illustrate these concepts, which is non-intuitive for many learners and can frustrate rather than excite. I hope that you’ll consider how this course might benefit you in your career.

The Road Ahead

It’s an exciting time to be part of the prospect development field, as our opportunities to leverage data and its accompanying technology evolve each day. As a member of the Apra community, I hope you’ll take advantage of the education we offer and connect with fellow members who are exploring this space.

I look forward to the promise of longer and warmer days in the months to come. I hope to see you this August at Apra’s Prospect Development conference in Indianapolis — registration is open!