Big, medium and small: when it comes to data, one size definitely does not fit all. Jacob Harold, president and CEO of GuideStar just over a year, helps Caroline Hartnell sort them out. He also tells her about the work GuideStar is doing with other organizations to make more data more accessible. GuideStar needs to become more of a curator of data than a collector, he says. What about big data? Will we ever reach a situation where funders can make decisions informed by comprehensive data? We can move towards it, says Harold.
What has changed at GuideStar over the last year?
The main thing is that we’re really trying to increase recognition that non-profits are multidimensional, complex; that any one approach to understanding their work is going to be an oversimplification. So we’re trying to bring in new kinds of information and to accelerate GuideStar’s move beyond what has been our historic reliance on the tax forms of US non-profits.
What other sources of information are you bringing in?We’re trying to build a supply chain of information about non-profits from many different sources: non-profits themselves, donors, experts, volunteers. For a homeless shelter, for example, the opinions of the homeless people who use its services are highly relevant to understanding its performance. For an environmental advocacy group its beneficiaries, who are everyone, may not know whether it’s doing a good job. In that case experts – journalists, academics or government officials – often have a much better sense of its effectiveness, which is why we acquired Philanthropedia in 2011. For certain types of development intervention, randomized control trials can be a source of information. The challenge is to create a framework that helps users understand which type of information is most relevant for a particular organization. That’s why we need not only to gather data directly but also to have partnerships with organizations like GiveWell, the Foundation Center, GreatNonprofits or IPA [Innovations for Poverty Action] as additional sources of data. And with these many sources, we’re going to need to become more of a curator than a collector.
On 25 September you formed a partnership with the Foundation Center. What will be gained by bringing the two organizations together?
The two have complementary missions and complementary datasets but some very similar operations and often similar users. So we need to figure out how to avoid overlap and waste by consolidating some pieces of what we do, notably the digitization of US tax forms, which is something we’re both doing.
Another thing that’s behind the scenes but really important is the BRIDGE Project, our partnership with GlobalGiving, TechSoup Global and the Foundation Center to create global ‘unique identifiers’ for NGOs around the world that will help to bring together all these different isolated efforts. Some other outcomes of our partnership may be more evident to an end user, such as co-marketing arrangements that allow you to save money by buying our products together.
What has the most potential but is the hardest is to co-develop products that will bring our different kinds of data together. People come to GuideStar for information about non-profits, including information about which foundations are funding them and potentially what other organizations are funded by those same foundations. We have some of that data but the Foundation Center’s data is better and could enrich our products. Conversely, if you’re looking for information about which non-profits a foundation funds and you want to know more about those non-profits, GuideStar has better data.
The Foundation Center’s audience seems to be both foundations and fundraisers, who might have quite different perspectives on what they’re looking for. What about GuideStar’s audience?
It’s very much an over-generalization, but the primary audience for the Foundation Center is grantseekers and the primary audience for GuideStar is grantmakers – including foundation programme staff, some individual donors, and also banks or law or accounting firms using GuideStar on behalf of their clients. Of course, it’s not that clear-cut: we have non-profits coming to our site looking for information on a potential collaborator, for example. The challenge is to acknowledge this basic division of labour while also seeing how the two approaches fit together. Although, I’d say that the greatest potential audience for GuideStar is non-profits because there’s so much potential for learning across organizations.
It’s important for both organizations to have good relationships with ‘input users’ as well as ‘output users’. The Foundation Center must have good relationships with foundations because it gets a lot of data from them, and it’s important for us to have good relationships with non-profits because 95,000 non-profits have given us data directly through the free GuideStar Exchange in addition to our getting data from their tax forms.
How is GuideStar funded?
Our budget is about $11 million a year. We get about $1.5 million in foundation grants and another $1.5 million from our membership programme. The remaining approximately $8 million comes from earned revenue, for example selling access to our tools to a bank or law firm or data company. Or they might want us to provide a specific cut of information – either a snapshot or a continuous flow of information through an API [Application Programming Interface].
Most of the services you provide benefit mainly US non-profits and foundations. The exception, I think, is the BRIDGE Project, which might particularly benefit international funders?
Yes. GuideStar USA is explicitly focused on US non-profits, but we do see ourselves as part of a larger philanthropy community, which is one of many reasons we’re engaged with the BRIDGE Project.
There’s also GuideStar International, which was acquired by TechSoup in 2010, so it already has a close relationship with a US-based organization. We definitely want to have better brand alignment so there’s no confusion between the two of us. In fact, we’re having a meeting later today which will give me a clearer picture of exactly where they are right now and of their relationships with in-country partners in Israel, India, Belgium, Luxembourg and elsewhere.
In the long run the hope is to have a coherent data system for philanthropy around the world. That’s not a project for any one organization. The BRIDGE Project may serve as a seed for that and GuideStar International would certainly be part of it.
What has happened to GuideStar International?
The expertise and talent and connections of GuideStar International got folded into TechSoup Global. But I think there is an opportunity to figure out how you go to the next level, which is sure to take cross-organizational collaboration.
I wonder if the reduced profile of the brand might have caused the momentum behind the GuideStar idea to slacken?
Organizing data about non-profits across the globe is immensely difficult; it will take years of international cooperation to solve this puzzle. Within the US, though, I’m glad to report that ‘the GuideStar idea’, as you say, has an extraordinary sense of momentum: more than 10 million visits to our website annually, 800,000 on our email list, more than 60 partners using our data, most of our programmatic indicators up more than 30 per cent last year. With all that said, we still face a resource challenge. How do you pay for the investment in cultivating new partners?
For many years, you were part of the grantmaking community, and now you are providing services to them. How has that change of vantage point affected your view of them?
In many ways, what I’ve been doing since I left the Hewlett Foundation a year ago feels very consistent with what I did while I was there; I’m interacting with many of the same organizations, the same people. There are some differences, though. The power dynamic, for one, but also, I came from the programme side of Hewlett, whereas the foundation staff who use GuideStar currently tend to be foundation operational staff members – from the legal team, the grants management team, the IT team, the communications team. It is less often used by programme staff.
Do you aim to bring together all the different sorts of information into a product that is useful to programme staff, too?
Absolutely, yes. We want to think about how we either adapt our existing products or create new ones that really solve problems for programme staff. That’s not going to be just information about non-profits but also data about what works, about different issues and, as I mentioned earlier, about where other funders are funding.
Will all these different types of information all need to be within one organization, or do you need to bring all the relevant information on to a common platform? So, for instance, users won’t need to think to themselves, ‘I’ve looked at GuideStar, now I’ll look at Charity Navigator’ and so on?
Definitely. We’d like to see an organization like Charity Navigator as an input into the supply chain. We don’t have that arrangement with them now but we’ve talked about it, and might some day share their data through our channels. We would be much more comfortable providing their Charity Navigator 3.0 data, with a greater focus on impact, than we would have been with their 1.0 data, because I think it’s getting much better. And, yes, we do want to figure out how to bring all of this together, because it’s just a waste of users’ time to have to float among so many different sites.
Do you think your understanding of the funding world from Hewlett is helping you as you bring these things together for the funding community?
My experience and my relationships are both helpful, but there’s one way that it’s dangerous. My experience is from a large, very well-run foundation, so I need to keep in mind that not all foundations have the same resources. And when we think about funding sources, we’re not just thinking about foundations but also about individuals, corporations, even government, and they require different approaches.
This really gets to the second half of what we’re trying to do, which is to think about how the information is distributed to end users. A lot of that’s going to happen on our website, but, as we’ve discussed, a lot will happen on other websites. A grants management software company is going to design something that is appropriate for foundation staffers and an organization like VolunteerMatch is going to provide data about non-profits that is useful to individual volunteers. Instead of having one user interface that you try and make work for everybody, you have different ones that are appropriate for different user groups but all based on the same data.
We also recognize that people as well as institutions have different analytical and emotional frameworks for giving. The Hope Consulting Money for Good research suggests that 10–20 per cent of people come to philanthropy with a very rational and analytical mindset. It’s not a majority but it’s still a significant number. At the same time, our Money for Good II project ‒ also with Hope Consulting ‒ showed that the overwhelming majority of funders and donors want impact information, and we’re really focusing our efforts on this kind of data.
The Hewlett Foundation is particularly interested in building up data and information. Is there a danger that this will cloud your perspective on what foundations and other funders might be looking for and how easy you need to make it for them?
There’s always a danger of our vision being clouded by our own personal experiences and background. That’s part of why I’ve been on what we’ve called a listening tour. I’ve met with staff of probably 20 different foundations, from the president to grants management administrators, to try to understand what they want, need, hope for.
Did anything that came out of the listening tour really surprise you?
One thing that surprised me was the extent of the operational/programme distinction in foundations in terms of our users. On the positive side, I didn’t realize that there were so many foundation operational staff who use GuideStar every single day. On the other hand, programme staff rarely came to our meetings at foundations. That was an important signal that we haven’t yet figured out how to be useful to them.
In September’s Alliance Chris Worman wrote about small data, for example 100 or so funders coming together to create a funding map that shows what’s going on relating to their funding priorities in a particular geographical area. Then, earlier this year, you wrote in the Harvard Business Review about medium data. So are there three levels of data, small, medium and big? And what is medium data?
What I mean by medium data is clarity about what you are trying to accomplish and standardization at a basic level about goals and strategies. So I think there is some overlap with the small data concept. Both are about the immediate world in which you’re working. But in the article there is also this definition of small data as the amount of data you can conveniently store and process on a high-end laptop. I don’t find this particularly helpful because you can put quite a lot of information on a laptop!
One thing about medium data is that our aim is to make it comprehensive. Take your example of a hundred or so organizations coming together to create a funding map: I would like to take that idea and make it universal. It would be shallow, but it would still have great value because it’s wide and standardized.
Take Charting Impact, our partnership with Independent Sector and the BBB Wise Giving Alliance, which is now fully integrated into our data collection tool. To me, it’s a quintessential example of medium data: it’s a data standard; it’s meant to be comprehensive but it’s relatively straightforward. It doesn’t go very deep, but it enables a common framework to emerge.
Another way of putting it is structured storytelling. We have a structure that enables us to make comparative decisions and be at least somewhat analytical while still creating space that allows the diversity of the non-profit sector to emerge.
Is big data what we’re all aiming for?
I think it will come and it will enable the kinds of insight that you can’t get from small or medium data. But we could easily fall down the rabbit hole of big data without ever getting the basics right, and that’s what I really want to avoid.
Let’s think about a vision of a perfect world for funders. Say you are a foundation in Boston and you’re interested in funding education in Tanzania, you could go to a single source and find out about the state of education in Tanzania, what the needs are, what other funders are supporting and where the gaps are, so you could see exactly how and where to intervene effectively. Is that a world you’re aiming for, or is that unattainable?
That is a world that I’m aiming for. When IBM uses the term big data they are talking about something at a scale that we can barely even wish for in the non-profit sector at the moment. But can we get to a place where people are able to make decisions that are informed in a structured and at least somewhat comprehensive way as you just described? I think we can move in that direction.
What are the realms of big data that we’re never going to get to in our sector?
We’re never going to have enough data to know, for instance, which advocacy organization contributed the most to which campaign.
Is that just because our sector is more complex than the business world?
It’s all about the unit of analysis. If the unit of analysis is the non-profit, that could be a billion-dollar enterprise like Save the Children or Oxfam. Going back to your Tanzania example, if you wanted Oxfam’s contribution to education in Tanzania as opposed to Oxfam’s global contribution, that would not help you. We may not get to breakdowns by programme within individual organizations.
Some organizations will provide the data by geography or by issue, but they won’t cross-reference it. You can find out how much they spent on education or how much they spent in Tanzania but you can’t find out how much they spent on education in Tanzania. These are the sorts of things that we can move towards. Maybe we’ll get down to how much is spent on education in Tanzania, but we won’t be able to break that down by intervention type. That means we’re not able to cross-reference it with the social science research that tells us which types of intervention have been effective and which haven’t. So we’ll have these comprehensive tools eventually, but there’ll always be a level of detail we can’t reach.
Even if you got to that ideal world, you still wouldn’t just go and make your grant, would you? I suppose the next step would be to find out who to talk to.
That’s exactly right, and that’s an extremely important point. Information is meant to inform, it’s not meant to decide for you. That’s the human bit, and we never want to take that away. We want to make it more alive, richer, and ultimately more effective.
For more information