Published in


Top 5 Nonprofit Data & Machine Learning Trends for 2020

Since Windfall was founded in 2016, we have revolutionized the way that nonprofit fundraising teams can access and trust data. Currently, we work with 500+ nonprofits across the nation ranging from local community organizations to major universities and hospitals. In each conversation with customers and prospects, we are impressed and encouraged about the way development professionals thirst for high quality information to drive efficiencies and uncover the next wave of philanthropic donors. Unfortunately, in a rapidly changing world there are increasing challenges for these professionals.

According to Giving USA, individual giving declined by 1.1% in 2018. This was one of the first times in history that individual giving decreased without a recession. Some of this can be attributed to recent tax changes in the United States, but nonprofits are also struggling to activate their existing constituents and replace historical established donors.

Given our unique vantage point, Windfall identified five trends in data & machine learning for nonprofits in 2020 that may help reverse this giving decline moving forward:

  1. Traditional Gift Capacity Metrics Will Be Replaced
  2. Data Freshness Will Become the New Normal
  3. Fundraisers Will Pay More Attention to Existing Constituents
  4. Sophisticated Machine Learning Will Be Democratized
  5. Increased Political Contributions Will Signal Increased Nonprofit Fundraising

Traditional Gift Capacity Metrics Will Be Replaced

Over the past 4 years, Windfall has encountered many development and research professionals that are skeptical of net worth as a metric for capacity. How can it be accurate? Don’t past gifts to any organization signal philanthropic capacity or willingness to give? What percentage can we take of their giving portfolio given past gifts to similar causes? At Windfall, we believe this has to change…and quickly.

The problem with historical gift capacity metrics is that (1) they are based on legacy proxies for wealth (2) the ability to “see” the full picture of every gift is becoming harder and harder and (3) the new generation of philanthropists don’t have a rich history of data. Let’s take a look at two metrics that most of our customers historically looked at: home values and prior giving.

Median Home Value by Market; Source: Windfall January 2020.

In many of our conversations, there are two distinct answers we receive from our customer: we look at home value as a proxy for wealth BUT we don’t trust home values as a metric for gift capacity. How is that possible? Well looking at median home values in two markets can clearly tell the story: San Francisco versus Santa Fe. This hits home for Windfall since we are based in San Francisco — just because the home values are higher doesn’t mean the wealth correlates. As a researcher, you would want to know: how much debt is on the property, when was it purchased, how much was put down, etc. If you even have 1,000 people in your database, how would you possibly have the time to vet that? How do you reevaluate every year? Every quarter? Clearly, this is a hard problem and we consistently hear that most organizations are moving away from home values.

What about historical giving? Of course, not every database with historical gifts is accurate — it may be missing a portion of past gifts. Unfortunately, especially for the affluent, Donor Advised Funds (“DAFs”) have become incredibly popular over the past 8 years. We can see this by the assets that DAFs hold by year:

Assets in Donor Advised Funds by Year

Source: Chronicle of Philanthropy; NOTE: Dollars are in billions

In 2018, DAFs took in $21.5 billion (up 24.7% from 2017) and in comparison the top 10 nonprofits took in $18 billion. Fidelity is one of the largest “nonprofits” in the country and took in over $9 billion in 2018 alone as individuals still receive the tax benefit for contributing to a DAF. When a donor makes a contribution, they can make it through their DAF and the trail of who donated to what becomes harder to track. Therefore, as DAFs are becoming more popular, development professionals lose the ability to see the full picture. For example, if I donated $100,000 historically to one organization that would publish it in their newsletter, I could now put that into my DAF and spread it across many nonprofits over the course of 10 years with the same tax benefit.

So what can we use to predict gift capacity in the nonprofit industry? Windfall recently wrote a whitepaper on gift capacity versus net worth to analyze the various approaches and pros & cons. While net worth doesn’t necessarily indicate that individuals will be philanthropic, the old metrics that nonprofit development professionals relied on will significantly change in 2020 and beyond.

Data Freshness Will Become the New Normal

When we entered the market, the common thread that we would hear about wealth screening is that organizations were screening on a 3-year interval. The data is “fresh” enough, so why would I pay for another screening this year? When most vendors update their data once a year and it’s not dynamic, our data scientists would agree with you; however, people are fluid and data decay (or working off of old data) is a real problem. What if your organization screened last quarter and someone had a wealth creation event 90 days later and your “competition” screened at the right time — you may lose a pretty massive donation, right? This is only true if your vendor updates data frequently.

At Windfall, we actually update our data every week. Changing behavior is always difficult and it did take us some time to see this change — now, more than 75% of our customers actually screen quarterly (if not weekly or monthly) to receive the most up to date data.

Customers Who Sync with Windfall Quarterly

Source: Windfall customer metrics by quarter, specifically the amount of organizations that screen in the quarter.

Clearly, as we have scaled to 500+ customers, more and more customers are opting into screening as frequently as possible. The only way we’ve been able to prove that is by showcasing our efforts to customers. In fact, one of our newest customers Ryan Bagley, AVP of University Development and Alumni Relations at The New School, summarized it best: “The one point I really want to make sure everyone leaves with: This is not your traditional wealth screening that we’ve all done and is painful, time-consuming, expensive, and stale by the time you’ve validated everyone. This is new. This is different…It can be trusted and you can use it right out of the gate.”

Freshness of data will be extremely important in 2020.

Fundraisers Will Pay More Attention to Existing Constituents

Similar to the for-profit world, development professionals are always on the hunt for the next large prospect. These prospects can’t possibly be in your CRM, because you know your constituent base, right? So where can we find the elusive leads? Let’s look for new prospects!

This is still a common objection we hear from prospects. In the old-world, we would tend to agree, as Ryan Bagley mentioned, wealth screening used to be time-consuming and super expensive. Each time you screen your database you have to curate the top 10–20% of donors to screen and get information. We think that’s crazy at Windfall — how do you know who is in that 10–20% pool? Your top donors? If you had to pick, why would you pick that population if you really know them?

Since Windfall focuses on affluent US households ($1MM+ net worth), we analyzed the data across 500+ customers and interestingly enough only 17% of affluent households were donors. That means that 83% of affluent constituents were not donors — this is a massive pool that our customers can now activate. You already spent time acquiring them in some capacity and probably are more likely to recognize your organization’s brand.

Instead of spending 5x to acquire your next new donor, Windfall believes that in 2020 nonprofits will double-down on existing constituents as long as there is a cost effective way to screen the entire database on a regular cadence.

Sophisticated Machine Learning Will Be Democratized

According to Crunchbase, there are 16,632 companies focused on AI/machine learning. That’s an incredible amount of talent focused on solving business problems. In fact, recent studies have shown that only 20% of the data science positions can be filled by “real” data scientists (not analysts, engineers, etc). While we won’t get into what makes a good data scientist here, at Windfall we have pretty high standards for what that means and the skill sets required.

For nonprofits, often times, there is a desire to get more sophisticated in the approach. The reality is that nonprofits have competing priorities just like any other organization and it becomes incredibly difficult to manage the team, streamline processes, and focus on the core mission of the organization.

In a few recent webinars, Windfall has surveyed some of our most sophisticated prospects and customers and collected results from 125 organizations. When thinking about machine learning algorithms, like a constituent’s propensity-to-give, we asked what organizations are currently using:

Source: Windfall survey of 125 nonprofits (customers & prospects)

While this is still a massive desire for these larger, more sophisticated organizations only ~27% have data science resources. Others are still using legacy approaches like heuristics or RFM (recency, frequency and monetary) to better prioritize their approach. While these are better than nothing, it simply does not do the trick.

Windfall believes that nonprofit organizations will embrace vendors in this journey to democratize data science. The important part for nonprofit leaders is: who do you trust? This is an important mission and data aggregators are not data science companies, so traditional vendors may not necessarily have the core competencies to handle large initiatives. Major questions for vendors include:

  • What data are you using for the models? What is the freshness of that data?
  • Is this modeling exercise one time or ongoing? If ongoing, how much will it cost?
  • How do you know the model works? What metrics are you using and sharing with us to evaluate the strength?
  • Are we measuring the performance over time to see if it works? What happens if it isn’t working? How do we iterate together?

Data science is really hard but has pretty amazing results. We hope to see the adoption rise dramatically in 2020.

Increased Political Contributions Will Signal Increased Nonprofit Fundraising

We are entering a supercharged election year with both sides donating in record amounts. Windfall has measured how affluent households drive contributions in elections, specifically the most recent 2018 election cycle.

Source: Windfall 2018 analysis.

While individual donations declined for nonprofits, there is a large opportunity to tap into those affluent households that are opening their wallets and providing contributions to political candidates. This is a great signal since these contributions are made without any tax incentives, signifying liquidity at that household.

Depending on the type of nonprofit, we have seen that political affiliation and political donation activity has correlated with donations over time. Windfall not only analyzes FEC data, but also state & local contributions. This is important since local elections are drawing more attention from voters (and donors). For example, let’s look at Virginia. In 2019, 2.4 million voters turned out for legislative elections, which was almost 1 million more than showed up in 2015 (1.5 million).

As nonprofits look for other screening metrics (other than wealth) to help them identify and prioritize constituents, we believe political donations will be another signal to help them engage with potential prospects / existing donors.


At Windfall, we are extremely excited and compelled to help nonprofits become more data-driven and leverage the power of machine learning. Over the past 4 years, we have seen tremendous changes in the nonprofit development ecosystem and behavior change that highlights the evolution of the overall market.

On our medium blog, we write many articles investigating current trends, changes in data science, and how to leverage third party data to make more informed decisions. We’ll keep track of our trends over 2020 and see how we did at the end of the year.

To find out more about our services and how we may be able to help your organization, feel free to reach out to our team.

The analysis and article was co-authored by Arup Banerjee, CEO, and Jeff Kamei, Senior Software Engineer, using a combination of third-party research, internal metrics, and surveys.

About Windfall:

Windfall helps you identify, understand, and engage the affluent. We provide you with precise net worth data on affluent US households, allowing you to make informed data-driven decisions.

For more information about Windfall please visit our website:

Insights from Windfall

Recommended from Medium

Why I’m not buying any property right now…

How to save 1.2 million people money on their mortgage

Household Net Worth Soars During Pandemic

Secure 2.0: Saving Millennial Retirement

Fortune: Why the coronavirus market is particularly dangerous for individual investors

Why Money Is Worth All The Hype?

Show Me The Money

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Arup Banerjee

Arup Banerjee

CEO & Co-founder @Windfalldata

More from Medium

ROI for data science led “Insights”

Learning the Udacity A/B Testing Course through its Final Project (Part 1/3)

Verify This! How Cedar Uses Data to Optimize the Check-In Experience

What Women Talk About