Just Data New Learning The Deep WIth

Conceptualizing better data work in non-profits: Prompt 2 – Data Collaboratives

This posts belongs to a series, Conceptualizing better data work in non-profits: First steps toward practice, in which I’m developing 5 prompts for orienting non-profit discussions about data work that can be organized toward justice. Each prompt is meant to provoke and ground ideas for organizational change that would include transformations in data work. The previous prompt leaned on the Reggio Emilia model of documentation. Prompt 2 offers Data Collaboratives.

I’m currently in the midst of writing a glossary entry on Data Cooperatives, so I’ve had them on my mind. I made an important connection the other day though. It should have been obvious, but the power of individualism is such that it wasn’t (at least not to me!). Though in some of my previous work on data, I suggested that we need to attend to the greater ecosystem and economies around data work, including asking funders to make changes to their demands around data, I’ve missed an essential point.

Data work in single organizations is almost always about individual clients. Other approaches to data work, like studying systems problems and social determinants, or participatory action research, invite us to consider not individual clients, nor even clients in aggregate: they invite us to look at the environments and systems in which clients live. In doing so, these approaches can skip focus on those harmed by oppressive systems and put blame squarely where it belongs.

Unfortunately, this isn’t particularly palatable yet to funders, private or public. They want accountability, and accountability apparently comes through tracking and reporting on “progress” by clients. This becomes the greatest challenge for many organizations that want to do differently with their data, but ultimately need to report to funders to keep the doors open. Yes, we need to change the ways funders are thinking about accountability. But I also want to offer another, viable prompt for thought.

The framing that I hadn’t quite yet stumbled upon was this: focusing on data in individual organizations is as harmful and ultimately inaccurate as focusing on individual clients. Clients/youth move through 3, 5, 10, or more organizations/institutions on a weekly basis. Each of these organizations has impacts on their lives. Despite numerous requests from agencies I’ve evaluated over the years, I can’t attribute change for individual participants to mainly/only their organization.

Collective impact models have tried to change this. They bring many organizations together across a geographic region (usually) and attempt to coordinate goals and thus objectives and data measurement / management /sharing. Unfortunately, these models come with many drawbacks. Mainly, they usurp power under an umbrella organization and cause a sort of superorganizational mission drift toward a particular goal, ultimately determined by outside funders and board members that live in wealthy suburbs.

Data cooperatives have the opportunity to become something collective impact cannot be in its current form. A cooperative is an entity made up of its members – and more importantly – governed by them. Participation of organizations and clients should be woven into the fabric of the organization. This means a data cooperative, which could collectively tell stories about changes for clients, and more interestingly, changes to neighborhoods, communities, and systems, would be capable of speaking to funders and policymakers. Organizing together around data would provide a vehicle for becoming organized generally, sharing resources, working toward shared goals, and influencing policymakers and funders to change accountability systems.

Data across organizations would provide loads of insights not available in individualized systems. This might include ideas like Nora Bateson’s Warm Data Labs, which would allow analysis of the connections between systems (say the places young people get lost, needs go unaddressed, etc.).

Importantly, a data cooperative would maintain the diversity of organizations working in an area together. Rather than elevating some organizations that fit funders’ ideas of “good” or “effective”, a cooperative would have to value better each of its member organizations. This diversity helps a cooperative weather the changing tides of the non-profit industrial complex’s new, most recent fad in philanthropy.

Non-profits have the notorious reputation of wanting to stay in their own lanes so they can compete with each other for limited funding. A commitment to clients, including providing better services, reducing harms, and advocating for serious change at all levels, should move us toward greater collaboration. A data cooperative provides a perfect vehicle. The Cooperatives movement has provided time-tested models for organizing, operating, and governing. Sharing resources around data work can help organizations tell their own stories. By sharing data, organizations create greater value for themselves and others. And I suspect, by thinking about data on this greater level, organizations might start to consider other models for data collection and reporting that can focus more on failed systems and less on tracking individual clients and outcomes. In other words, data cooperatives in non-profit organizations may be a model that could help transform the industry’s relation to data and also to systems change.

In case this isn’t yet clear: this is not the same as collective impact, or as attempts to build shared databases used by many organizations. The difference lies first in how these things are funded and who decides what outcomes matter – in the case of cooperatives, this is bottom-up and community-based. The second difference is who designs and governs these systems. The third difference is that cooperatives must first benefit their members and the social good – not funders, not governments, and not the businesses that are currently the profiteers of shared data systems (and that’s the vast majority of current data systems in use by CI). Finally, while CI models theoretically increase cooperation, they usually enforce a new competition between organizations to perform to desired metrics, closing down organizations that don’t fit the current metric du jour and promoting those that do. A cooperative provides real models for deep collaboration among organizations, and builds toward a collective good.