A revolution has begun: Data are transforming the nonprofit world.
The new number crunching is fundamentally changing the way charities make decisions about programs, solicit contributions, and push for social change.
- The Indianapolis Museum of Art pores over attendance information and visitor surveys to determine everything from when it should be open to the best ways to spend scarce resources. The museum extended hours for its Matisse exhibit when data from previous high-profile shows predicted a surge visitors in the final weeks of the show. The change allowed more than 4,000 additional people to see the exhibit.
- The Humane Society of the United States, World Vision, and a growing number of other charities have added data analysts to their fundraising departments. One of the new hires’ most important tasks: creating detailed statistical pictures of the groups’ major donors and then scouring the organizations’ databases to find other supporters who fit those profiles. The goal: to help them focus attention on the people who are most likely to give.
- DonorsChoose, which helps teachers raise money for classroom projects and equipment, makes information about the projects funded on its crowdsourcing site available for policy makers, parents, and others to see, giving them a glimpse of what schools need money for.
Foundations are getting in on the action, too, with a handful making large grants for data projects. Last month, the John S. and James L. Knight Foundation awarded $2.2-million in grants for programs that use data to improve health and well-being, and the Bill & Melinda Gates Foundation awarded $600,000 in November for projects that combine multiple sets of data to help solve tough problems.
Some data experts envision a not-too-distant future when charities can use the information they collect to home in on individuals in greatest need of assistance and tailor their approach to each one.
Pete York, founder of Algorhythm, a data-science consulting company, is working with the State of Florida to use sophisticated data-mining technology to identify juvenile offenders at greatest risk of committing another crime and determine the right mix of services to help each one overcome the odds.
“There’s no such thing as one program that works for everybody,” says Mr. York.
Not Like Businesses
But not all organizations are prepared to fully exploit data’s promise. In many cases, nonprofits struggle to find the money and personnel to collect and analyze such information.
And some experts say the charity world and Big Data aren’t a perfect match to begin with.
Data can have a profound impact on certain problems, but nonprofits are kidding themselves if they think the data techniques used by corporations can be applied wholesale to social problems, says Patrick Ball, head of the nonprofit Human Rights Data Analysis Group.
Companies, he says, maintain complete data sets. A business knows every product it made last year, when it sold, and to whom. Charities, he says, are a different story.
“If you’re looking at poverty or trafficking or homicide, we don’t have all the data, and we’re not going to,” he says. “That’s why these amazing techniques that the industry people have are great in industry, but they don’t actually generalize to our space very well.”
Weak Statistics
Another big problem, say evaluation experts: As the nonprofit world works to get up to speed on data, organizations that are rigorous in the way they use information could actually be at a disadvantage when they’re compared with groups that are sloppier.
For example, consider two youth charities that had 80 percent of students graduating from high school in areas in which only 40 percent of students in the general population did so. Suppose only the second group’s numbers took into account that students who wanted to go to college were more likely to find the program. Some donors would be more impressed with the first group’s data.
A growing number of grant makers are getting smart about measurement, says Neil Buddy Shah, a co-founder of IDinsight, a nonprofit evaluation group. But, he says, until that knowledge is more widespread, there are going to be groups “that get away with printing out a glossy brochure and some weak statistics and get a lot of money.”
“While some people in the nonprofit world harbor doubts about the value of Big Data, many are eager to embrace it.
Two new conferences last year focused on data and analytics—After the Leap, held in Washington, and Do Good Data, in Chicago—and both sold out.
In Dallas, the Communities Foundation of Texas created the Data Driven Decision-Making Institute, a yearlong program to help area charities that serve low-income families make better use of data, expecting to offer the program twice.
But so many organizations have applied that the foundation now plans to offer it for a third year. What’s more, arts and education nonprofits have asked the fund to develop similar programs for them.
“When you’re on the ground, you’re starved for good information,” says Sarah Cotton Nelson, the foundation’s chief philanthropy officer.
Help Wanted
Efforts to get data scientists interested in using their expertise for a good cause are expanding quickly—and attracting some heavy hitters.
Last year, Rayid Ghani, chief data scientist for President Obama’s 2012 campaign, started Data Science for Social Good, a summer-fellowship program that pairs undergraduate and graduate data-science students with nonprofits and government agencies to solve pressing problems.
Eric Schmidt, executive chairman of Google, and his wife, Wendy, financed the program, which received 600 applications for 36 spots.
Another group, DataKind, which matches data scientists with charities for pro bono projects, has grown from an all-volunteer effort to a nonprofit with six full-time staff members in a little more than two years. Late last year, the group won a $250,000 grant from the Knight Foundation.
And yet nonprofits, like for-profit companies, often struggle to hire professionals to help them gather and parse the numbers. McKinsey & Company estimates that American business now offers from 140,000 to 190,000 more data-analyst jobs than it has people to fill them.
Turnover among data analysts and data scientists is high, but that’s not necessarily a bad thing for charities, says Mr. Ghani, who also serves as research director at the Computation Institute at the University of Chicago.
“You don’t need a person to spend 30 years at a nonprofit,” he says. “You need short stints. That’s what most companies have done. The average tenure for people like that is a year or two years.”
Mr. Ghani thinks nonprofits will have to rely on a mix of full-time employees and data professionals who volunteer their expertise. He says he gets at least one email a day from data specialists who want to lend their talents to a good cause but aren’t sure how to do so.
Need to Collaborate
Many groups will need to change the way they work to benefit fully from the data revolution, says Andrew Means, founder of Data Analysts for Social Good, which holds meetings and online trainings for charity leaders and analysts who work at nonprofits.
Collaboration among nonprofits is key, Mr. Means says. Right now, he notes, most measurement happens in single organizations working alone. But in data science, he says, the bigger the data set, the greater the potential for learning.
“If only 100 people used Google, we couldn’t get that much good information out of Google,” he says. “It’s the fact that hundreds of millions of people are using Google that makes it valuable.”
Mr. Means also worries that many nonprofits have yet to make the leap from collecting data to taking action on it. He says he sees data professionals who are excited to be working in the nonprofit world grow frustrated and eventually leave because their organizations’ leaders ignore their analyses when making decisions.
“Even the organizations that hire these people tend not to listen to them,” he says. “The nonprofit sector doesn’t always do a great job of listening to evidence over instinct.”
Quantity vs. Quality
Foundations increasingly demand evidence that the programs they support work.
The problem, says Nancy Roob, head of the Edna McConnell Clark Foundation, is those demands have not been accompanied by money to help charities evaluate their programs.
“On the one hand, funders want to invest in what works, and they want nonprofits to deliver good results,” she says. “And then on the other hand, there’s very little funding that’s available to support nonprofits to build the capacity that they need to do that.”
Clark has long focused its grant making on youth charities that can prove their impact. Two years ago, it started PropelNext, a grant program to help promising groups collect and use data to improve their programs and develop performance-management systems.
Too often, the data that foundations ask for don’t help nonprofits improve their programs, says Stu Taylor, director for performance measurement at International Development Enterprises.
He says that collecting that information takes valuable time and undercuts his efforts to get employees to embrace assessment—not because donors demand it but because the data can help improve programs.
Says Mr. Taylor, “It does make it more difficult when you say that one day, and then the next day you’re emailing them back to say, ‘Oh and by the way, I need that report on my desk by tomorrow with all of these indicators that we all agree are not all that relevant.’”
With so much need and with limited charitable dollars, it’s critical for the nonprofit world to turn to data to figure out what really works, says Paul Niehaus, co-founder of GiveDirectly, an international antipoverty group that relies on an independent evaluator to collect data on its work and gauge its impact.
“Things that we thought we saw with our eyeballs or things that made sense in our head have often turned out to be untrue in the real world,” he says. “So they have to be tested.”
3 Ways the Indianapolis Museum of Art Uses Data

40%: The percentage of visitors who attend high-profile exhibits in the final three weeks.
Action prompted by the data: The museum extended its hours during a recent Matisse exhibit when it realized its standard hours wouldn’t be enough to accommodate the expected surge of visitors. Tickets sold out during the extended hours, which allowed more than 4,000 additional people to see the show.

46142: One of the zip codes in the metropolitan Indianapolis area that had the lowest attendance rate.
Action prompted by the data: Placed billboards in Greenwood, a town bordering the Indianapolis city limits, which has increased museum visitors from the area.

The museum asks visitors to fill out surveys about their experience that include both open-ended questions and very specific ones, such as whether they viewed an exhibit’s accompanying video.
Action prompted by the data: Each survey question helps the museum make smart choices about how to tweak exhibits and spend scarce dollars, says Meg Liffick, director of public affairs. “If we are seeing that one out of every 100 visitors sat down and watched the video, then what do we need to do in order to make the exposure for the video better?” she says. “Or is it worth it to have that video?”