Funding and Doing
How The Agency Fund Invests in Social Ventures
By Paul Wang
Imagine a social venture with a world-changing idea. What would they need to scale the idea – to continually improve and grow it, reaching not just hundreds or thousands, but millions or even billions?
At The Agency Fund, we focus on two elements to unlock cost-effective impact at scale. First, a key – and often overlooked – ingredient for human flourishing is agency, or people’s fundamental ability to envision, navigate towards, and realize futures they desire. The principles of agency are evidence-based, and apply across sectors and populations (see our white paper). To enhance agency, programs we support often feature elements of mentorship, coaching, and/or counseling. These programs can rely on in-person interactions, while many – in part or exclusively - engage people via mobile technology.
Complementing the agency focus, we also focus on the power of learning and building for scale by enhancing our partners’ data, technology, and research capabilities. This focus takes inspiration from the tech sector combined with the observation that mobile technology is important for most of our partners. Generally speaking, such technical capabilities are nearly always vital for social ventures to achieve cost-effective impact at scale.
To execute our mission, we use a unique and innovative “funder-doer” approach where – in addition to grants – we provide direct in-kind resources to organizations.
Unlike most philanthropies, more than half of our team comprises data-related technical staff (e.g., economists, psychologists, and computer engineers) who spend the majority of their time working alongside organizations in our portfolio. Others may also combine funding and doing (the World Bank comes to mind), but - overall - it is not a very common approach.
In our experience, a combined focus on agency + data-driven learning enables promising social ventures to make their programs considerably cheaper, adaptable to context, more impactful, and scalable.
We feel these are keys to unlock outsized social value, and are very happy to unpack more of our thinking to the world.
Unlocking program improvement through data and technology
Take our partner Shamiri Institute, a Kenyan nonprofit that equips young people to provide mental wellbeing support to high school and university students. Shamiri is keen to scale its approach, since rigorous evaluations have revealed exciting impacts on depression and anxiety. However, they know their program needs to be more affordable, simpler to use, and more adaptable – to contexts across Kenya and beyond – before it can scale.
To help address these needs, software engineers at The Agency Fund are collaborating with Shamiri to pilot smartphone apps and a digital data system, and a behavioral scientist is helping analyze existing program data, and design and assess program innovations. With more efficient data capture and experimentation, Shamiri can optimize their operations via many channels including participant tracking, nimble program management and quality assurance, iterative experimentation, predictive analytics and detailed program costing. Our year-long collaboration aims to establish these capabilities in pursuit of a new dimension of data-driven learning at larger scale.
Most organizations only scratch the surface of what data and technology can provide. The reasons behind this are myriad, systemic, and could motivate many additional blog posts. Suffice it to say that these capabilities require technical expertise that can be scarce, expensive, hard to find, and/or otherwise inaccessible for many organizations. More broadly, the current funding landscape often does not adequately incentivize the development of even basic data and technology capabilities. We hope to help alleviate these constraints to democratize the use of data and technology.
A funder and doer
But why did we build our own technical team, rather than simply give grants focused on technical support and/or partner with a service provider to build these capabilities? The answer is twofold.
First, providing technical services ourselves allows us to work dramatically faster, cheaper, and with greater consistency. Organizations often wait 6-12 months to line up funding and service providers for a technical project. With our own technical team, we can act as soon as a promising opportunity arises. Structurally, working through grants or third parties adds a layer of overhead and increases communication and coordination challenges. We have learned that fewer cooks in the kitchen is significantly cheaper and easier to implement.
Our partner SameSame, for example, is a new nonprofit that uses mobile technology to support the mental health of LGBTQI+ youth. We have worked with SameSame since its founding in 2022 to integrate adaptive experimentation into its routine programming – using an approach similar to A/B testing called “multi-armed bandits.” Such early and dedicated technical support allowed SameSame to integrate data, technology, and learning into its operations from the start, and the venture is now on track to introduce large language model AI capabilities into its programming, with continuous evaluation and treatment allocation. At their young age, it is difficult to imagine how SameSame could have acquired such capabilities this quickly via more traditional philanthropic processes.
Second, beyond these direct impacts with grantees, our funder-doer approach enables us to learn about the common challenges that organizations face, and build scalable products and platforms that address these pain points. Our partners often need highly similar solutions – i.e., routine data analysis, data visualization, rapid experimentation, automated program costing, mobile chat capabilities, and AI – and once built, these solutions can often be easily transferred to other organizations. In many cases, the technical solutions we build can even be transferred to organizations that work in very different sectors or settings.
A good example is Dalgo, an open-source platform that we supported with Tech4Dev, that automates data ingestion, storage and transformation for development organizations. In the tech sector, there are standardized processes for extracting, transforming, and loading data. Dalgo was built after observing that established Extract-Load-Transform (ELT) platforms do not work out of the box in the development sector since most NGOs work on a different data stack (including ODK, paper surveys, and other field data capture tools). Once built, transferring these technologies and practices is free and enables organizations to easily use program data for many downstream applications, including experimentation, predictive modeling, and customization. Working with just a few organizations on these processes revealed the opportunity to democratize such capabilities across social ventures.
Being a philanthropy enables us to accelerate spillovers via additional channels. For example, we organize fellowship cohorts of social entrepreneurs to identify how their new ventures can harness data techniques from the start. Similarly, we bring multiple organizations together for “collaborative sprints” – intense, in-person workshops to jointly develop data and technology solutions – to harness collective intelligence, cross-pollinate ideas, and – ultimately - build scalable tools and solutions. We also match-make and facilitate collaboration between organizations that would clearly benefit one another but might not otherwise connect. All of these activities are made far easier given our position as a philanthropic organization.
Rapid impact, lasting change
A key pillar of our funder-doer approach is its time-bound nature: we do not intend to embed in organizations forever, and all our technical support partnerships are capped to be as short as possible without sacrificing impact. We feel that time-limited efforts to introduce new capabilities are most appropriate – nobody wants these efforts to drag on, because if they did, it would indicate that we had failed to build actual capabilities.
Our bet is that our partner organizations will not only be able to optimally use and maintain new technologies and data capabilities; we believe that these changes will also increase their operational efficiency and impact, producing higher growth and attracting more funding.
This is a bet on our ability to identify partners who will maintain the platforms we help build, but also on the broader philanthropic sector. We hope other funders will recognize the organizations we work with as operating on the frontiers of technology and cost-effective impact, and resource them accordingly in the ambitious ways we cannot.
It is still early days, but we feel that we are making important progress in cracking this most promising of nuts: data and technology for cost-effective impact at scale, around the world and in virtually every sector. We have already seen remarkable progress, and we are excited to see what opportunities and partnerships we can realize next. Stay tuned to see how we fare in the coming months and years!
