How Artificial Intelligence Improves Predictive Analytics for Nonprofit Fundraising

Artificial Intelligence for nonprofits is yet to be widely adopted, but the beginnings of its growth can already be seen. In a recent survey of 212 nonprofit leaders, 89% of nonprofit professionals believe that AI can make their organizations more efficient. AI’s first impact on nonprofits will be seen in fundraising by enabling better, faster and more affordable predictive analytics.

What Is Predictive Analytics?

Analytics is a broad term that means gaining insights from data. In 2010, IBM published an article that described three kinds of analytics:

• Descriptive Analytics: Insights into how things are now or were in the past based on data. 
• Predictive Analytics: Insights into how things are likely to be in the future, based on past or current data. 
• Prescriptive Analytics: Insights into what actions are most likely to change things favorably in the future.

Without necessarily knowing these labels, nonprofits have long used analytics in fundraising. Here’s some examples:

• Descriptive Analytics: Understanding who past and current donors by analyzing donor records. 
• Predictive Analytics: Developing “personas” of likely future donors by examining who past and current donors are. 
• Prescriptive Analytics: Developing fundraising campaigns most likely to acquire new donors or re-activate past donors based on the success or failure of past campaigns. 

What Is AI (and Why Shouldn’t You Use It)?

With all the excitement (and hype) around AI, it’s easy to forget that AI is a means to an end, not an end in itself. In fact, AI refers to a wide variety of applications — the only commonality is that they are systems that perceive their environment and take actions to maximize their chances of success.

Put simply, AI is a “how” not a “what.” Nonprofits benefit from AI only when it’s used to accomplish a task that provides value and does so in a way that is better, faster and/or more affordable than current methods of accomplishing the same task. In fact, many tasks that can be accomplished by AI are often better accomplished without AI. (For example, autonomous driving, while it can be done with AI, is still better done by humans — at least for now.)

There are, however, some tasks that AI excels at. One of those is predictive analytics.

How AI Enables Nonprofit Predictive Analytics

Nonprofits typically perform predictive analytics manually. They look at their donor records and apply statistical analysis, past experience, and personal expertise to make informed judgements about who are likely future donors. The process involves looking at huge volumes of data, trying to determine what matters in the data, then testing those determinations over time. It is time consuming if performed in-house and often costly if outsourced to a professional.

AI makes predictive analytics for nonprofits faster, better, and more affordable. AI can sort through more massive volumes of data than any human could and, through machine learning, develop predictive models faster and with far more complexity and power than what a person could do on their own. And AI does so with more power to digest and identify relationships in massive quantities of data without human bias. These AI-powered predictive analytics produce returns that pay for themselves.

Examples of AI-Powered Predictive Analytics for Nonprofits

Donor Acquisition

A nonprofit can use records of current “best donors” to build a predictive model using AI that can find likely future donors in any list of prospects. A nonprofit recently used both an AI predictive model and wealth-screening to select over 20,000 prospects for engagement in a direct mail campaign. The result: The predictive model developed by AI identified prospects that results in 114 new donors. The number of new donors resulting from the prospects selected by screening: zero.

Donor Engagement

A nonprofit can use donor records to build a predictive model using AI that can identify which current donors are the best prospects for further engagement, such as a higher gift ask. A nonprofit recently used a model built by AI from past donor records to predict which current donors were the best candidates for a major gift ask. The result: a major gift donation rate 56% higher than an identical campaign that did not use the predictive model.

Donor Retention

A nonprofit can use donor records to build a predictive model using AI that improves donor retention by identifying which donors are most likely to respond to particular campaigns, allowing nonprofits to target the right message to the right recipient. A nonprofit recently used an AI-created predictive model that resulted in a 35% increase in email open rates.

AI holds much promise for improving nonprofit fundraising by enabling predictive analytics at speed and scale. However, this promise should be tempered by a recognition that AI has its limitations. First and foremost, AI lacks judgment, an understanding of context and empathy — human factors that are critical for nonprofit fundraising success. AI predictive analytics should be used to enable — not replace — nonprofit fundraising professionals. Human beings make AI more powerful by imparting judgement on and teaching context to the machine’s predictions.

Second, AI requires an ethical framework, particularly when it is used to make decisions about people. Both the organizations that build AI and those who use it share the obligation to use AI responsibly.