What would it take to stop creating bad technology in the social sector?
It’s no secret that nearly all of our large-scale workplace systems are pretty lousy. A friend of mine calls these “sociopathic technocratic” systems, and I call it Bad Tech. Bad Tech is built without regard to the actual needs or jobs of the end users (the data producers), but are required because they benefit someone else entirely (data consumers, usually the bosses and boards who need reports and metrics). In reflecting on two decades building enterprise technology systems I have to admit I have been unwittingly complicit in creating Bad Tech, and I’m tired of it. I find that I am not diminishing limitations, but rather, I am moving limitations around in a system that hasn’t been fundamentally rethought in awhile.
“Technology can bring benefits if and only if it diminishes a limitation.” — Dr. Eliyahu Goldratt, The Power of Technology
I attended this year’s SSIR Data on Purpose conference with these thoughts rattling around my brain, and the presentations were revelatory. The speakers all seemed to be grappling with the same thing — how do we work within the tensions of the digital now and change the trajectories of some of the more damaging and dehumanizing trends? No one has a solution, but I have the seeds of a new way of thinking as a solutions designer, and I want you to join me in figuring out how to do this right.
Rahul Bhargava of the MIT Media Lab identified four problematic trends that tend to arise in data collection systems:
- A lack of transparency due to centralized authority
- Extractive collection practices
- Technical complexity
- Inequitable control over impacts and use
I took this presentation as a challenge to envision how we could create data systems that deliberately stand in opposition to these hallmarks of Bad Tech. What if, starting today, the 4 pillars of system design are that all systems should strive to be:
- Transparent: fully informed by, and informing, all stakeholders
- Generative: providing direct and immediate value to those who perform the work (line staff), those who benefit from the work (constituents/beneficiaries), and those who benefit from the labor and information (executives, funders). In an ideal world, solutions will allow those who perform the work and those who benefit from the work to use their labor for the creation of new things previously unimagined by the system designers.
- Trusted: the origins, transformations, responsibility, and rules impacting data are understood and accessible to all stakeholders. Clarity combined with transparency minimizes the chance that technical complexity can be used to bully or intimidate stakeholders.
- Equitably Owned: shared ownership and responsibility for information generation, process management, and interactions within the system mean that the people impacted by system decisions will be heard.
Project Signal in San Francisco introduced me to the idea of a data trust for multi-stakeholder digital systems. As defined by BrightHive, a data trust must meet several exacting standards for governance, collaboration, ownership, and transparency. Done properly, data trusts cover the four pillars of our new systems rules: they are transparent, equitably owned, have mechanisms to ensure trust, and are aware of minimizing extractive practices. They are also, seemingly, public sector or massively multi-stakeholder systems.
So I wondered: why would we not take this approach as foundational in the social sector? What would it look like if organizations like grantmakers and affiliate-model organizations thought about their enterprise systems as trusts for all stakeholders? An organization’s operations — technology, processes, and management— are the living embodiment and enforcement of its values and strategy. Wouldn’t a data trust and the processes and governance surrounding it better embody our sector’s values of equity, trust, and transparency than what we have today?
An organization’s operations — technology, processes, and management — are the living embodiment and enforcement of its values and strategy.
What does this really look like?
I come from large scale enterprise systems integration work, so I gravitate naturally toward the translation, clarity, rule-setting, and governance activities required to make those systems work. But in bringing that approach to my nonprofit work, I believe I unwittingly also brought unexamined practices from a top-down, central-control, revenge-of-the-actuaries culture that originally allowed extremely complex, expensive systems to be built when the barriers for technology were much higher.
I found myself thinking that if I could go back in time and prevent myself from idealistically applying enterprise technology paradigms to grantor or national organization projects, what might that really look like? This is clearly a work in progress, and fairly high level. But it’s a start.
- Acknowledge the core tension of organization versus individual
Peter Senge illustrates creative tension as a rubber band stretched between two fingers. There’s a tremendous amount of energy in that stretch, but it’s also uncomfortable to hold the tension for any length of time. We tend to reduce the tension by lowering our expectations or ignoring reality. Essentially, we let one side or the other win.
Most large-scale digital systems are centrally funded, centrally mandated, and primarily serve the interests of a central authority and central data consumers. There are some valid reasons for this, but there is no valid reason to let that perspective dominate fully. When the organizational/central perspective dominates, individual data producers are often disconnected from the purpose of the data system, and see the data system as extra work rather than a removal of limitations to their daily routine. Out-of-balance systems like this often cause duplicative data entry, “real work” done outside of the central system, and employees who feel like they’re wasting time and are prone to burnout. This unbalanced dynamic is the origin of the Bad Tech patterns Rahul Bhargava identified.
What would I do differently? Voice this tension more clearly. As a solutions designer, I can hold the space for this tension to be explored, and help create a more neutral dialogue that minimizes power dynamics and focuses on assets and solutions. Often I find that people are scared to speak up when someone more powerful or with more money is in the room — I can help their voice get heard. What else would you recommend? What would you do differently?
2. Put governance first
In the social sector, we are running so hard and fast to just keep basic services flowing to an endless stream of need it can be very hard to value slowing down and putting rules in place. In the US, “slow” is anathema, a slur and an insult. Governance is the deliberate, thoughtful management of change, and it can be a little slower than we‘re comfortable with when it comes to technology. It is also the most important part of a healthy, functioning system. Neglecting governance is the fastest path to a bloated, opaque system that nobody trusts and everybody is a little bit afraid of.
Governance is the key to transparency, trust, and distributed responsibility. Governance is nothing more than basic co-operative practice to manage a shared resource in which everyone has a stake. Good governance — equitable governance — ensures that everyone’s voice is heard and respected as part of the change management and assessment process. When people are able to shape a system, they are more able to see their role, responsibility, and ownership of that system.
Being a more vocal advocate for governance, educating people about governance, and modeling the inquiry and conversations necessary for good governance are three concrete things I changed about my practice when I started Deep Why. I have a lot to learn about co-operative practices. What else should I be doing? What might you do to foster governance in your projects and organizations?
3.Let the systems talk to each other, and provide support
This is the part where I finally talk a bit about actual technology. The social sector is currently plagued with technology-driven inefficiencies both inside individual organizations and in our broader ecosystems of grant-makers, foundations, and other centralized data systems. Line staff whose time should be spent on mission delivery are double- and triple-entering data into systems that don’t talk to each other. They waste time capturing information that is of no use to them or their beneficiaries, solely for reporting purposes, but most organizations don’t have the analytics capacity or expertise to assess all of this extra data.
For individual organizations, establishing integrated, aligned systems is what I do for a living. There are a million tools and technologies out there, some better adapted to the operational capacity of any given organization than others. If we put governance first, we can construct integrated systems or at least minimize internal inefficiencies. This is hardly a new idea. What is somewhat new is using a co-operative, shared-resource governance model to inform our choices.
The problem that has kept me awake at night is that even if I manage to help my clients achieve internal operational alignment and efficiency, they are still at high risk for duplicated efforts and wasted time when they have to use external systems. I know small organizations that have to log into dozens of grantor systems, wasting hours each week just to report back for a few thousand dollars. Grantors have legitimate reasons to need this data, but the ways in which it is captured is often creating inefficiencies for organizations who can least afford it. A similar challenge arises when independent organizations use a program model that originates at a national organization. National needs data for very good reasons, but the affiliates already have their own technology and processes, so logging into another system is a burden, not a help.
While I am not advocating for funders or national organizations to establish public-facing systems in a true data trust fashion, I am advocating for a more thoughtful approach across the sector to minimize the proliferation of non-mission-serving demands on people’s time. I am advocating for some technology steps in addition to putting governance first:
- Please ensure that your central system has a well-written, well-documented, trustworthy API so that organizations with the capacity to exchange data automatically are able to do so.
- Please fund the resources to identify the most common candidates for automated integration (if half of your grantees use the same CRM, for example) and fund the development and support of that integration.
- Please fund the resources to identify whether middleware will enable a large percentage of your grantees to automatically connect to a system that transforms their data into your reporting requirements. If you find this is the case, please fund that middleware and the necessary support for its maintenance.
- Please construct methods for bulk upload where APIs and middleware are inaccessible to your partner or grantee organizations. Many nonprofits use Excel, and they need help to efficiently using your central system.
To be clear, the burden is not solely on the funders and national organizations. In a data trust model, all parties are responsible for their piece of the work, because a data system is a shared resource that creates value for all. Individual organizations must also be willing to participate in the rules and governance for these shared systems, which at first may be more time-consuming than just keying the data into yet another database. Individual organizations must be willing to take advantage of the offered automations. It is a dramatic change in the way we currently operate in the sector, masquerading as a small step into modern technology.
The era of massively accessible consumer technology is putting pressure on all industries to rethink our enterprise technology approach. When we make better use of published APIs and choose technologies that can be made to talk to each other, we can gain efficiency, and we are taking necessary steps. But this is not sufficient to create transparent, trusted, generative, and equitably owned systems. I believe that change comes from our minds, hearts, and practices, not the technology itself. Technology is and can only ever be a tool to represent what we believe are limitations and priorities in our work together.
As a solutions designer, it’s my role to explore and educate about practices and models that may not be familiar to other project stakeholders. I can certainly suggest the alternative of collaborating within a data trust model when asked to “build a database.” I have not been an advocate for common-sense automated connections between systems, and I will correct course immediately. What other barriers need to be removed to pursue this idea? What role can you play?
Is this realistic?
While most organizations believe they are designing their data system as a shared resource, I have only actually seen it once. A client designed their entire data system as a true operational support tool, with and for their line staff.
- All of the screens and interactions were built for one reason — to let staff do their jobs better. Staff co-designed, tested, and approved the system.
- If they couldn’t build the functionality, they integrated a tool that was already in use and familiar. The mandate was “no double data entry.”
- Reporting and executive needs were represented, but not dominant.
- The organization was operationally mature and had strong alignment between their strategy and staff activities before they built a new data system. It’s not a substitute for governance first, but the rules, roles, and processes were crystal clear to everyone, enabling trust and transparency.
I think that it wasn’t technology that allowed this client to succeed as much as it was a shared understanding and fierce commitment by all parties that the data belongs to the beneficiary, is to be used in providing services to the beneficiary, and is captured to allow those services to be provided. Reports, management, and executive activity all emerges from data that is captured as part of the daily work. Nobody feels like they’re wasting time or energy on irrelevant tasks.
This is why I put governance and shared purpose ahead of technology in my list. When people share the commitment and understanding of who the data belongs to, what it’s used for, and what their role is in stewarding it, the technology to support that vision becomes easier to create.
I’m not sure what this looks like with foundations, grant-makers, and national organizations with affiliates yet. I simply haven’t seen it up close. I know that initiatives like the Humanitarian Data Exchange from the UN’s Centre for Humanitarian Data are well ahead in creating a data exchange, but I’m unclear how closely it models a data trust structure. And of course BrightHive is partnering with organizations to execute on data trusts. I do know that we have the technology to enable secure and massively interconnected systems; we know how to govern these systems, and we know how to design them. We collectively commit a lot of time, money, and energy to inefficient systems, so diverting the time, money, and energy to systems that remove the inefficiencies seems reasonable.
Let’s find out together.
I have spent so much of my life advocating for the experience of the end user when I work with an enterprise system, trying to make life just a little bit better for the line staff who has to do all the real work. But I have never once really understood that my efforts might just be making an extractive system marginally less uncomfortable while the extraction is taking place. I keep participating in the creation of Bad Tech, and I’m ready to own that responsibility and help propel our sector toward good, useful, usable tech that lets the amazing humans doing amazing work get on with that work.
This article is just a beginning, and I intend to keep exploring the practical realities of creating data systems that live up to the four pillars of transparent, trusted, equitably owned, and generative. Please share your experiences so we can collectively move away from unnecessarily redundant and inefficient data practices in the social sector and toward using these powerful tools as ways to improve our reach and impact.