Here is Evidence Action Beta’s process for identifying and pressure-testing evidence-based innovations that reduce the burden of poverty, and then designing and building the most cost-effective programs for massive scale:
Identifying / Sourcing
We source ideas that have been initially vetted by others through one or more rigorous impact evaluations, most often in the form of randomized controlled trials. We conduct extensive desk reviews, invite submissions from researchers through calls for results, and work with top development economists and other experts
Pressure-testing for viability
Once we have a roster of pro-poor interventions that show initial, experimental promise, we put these ideas through a rigorous process to test if they can be implemented in the real world.
Think of it as product development for reducing poverty: A series of tests that help us increase our certainty that a good idea will work as a solid program that reliably and cost-effectively works for millions.
Many possible interventions fold under this pressure, so we are not afraid to end those that won’t, for whatever reason, have real impact at scale.
We carefully screen ideas in our pipeline for these criteria:
What is the global market for the intervention? How many people could benefit and where are they located?
It is feasible to deliver the product to millions of people? It’s one thing to be able to serve 500 children in 10 schools but an entirely different level of operations to deliver to 50 million children in 1 million schools.
How cost effective is the intervention; that is, what is the cost relative to the social welfare impact? Can the unit costs/cost per person be reduced so that scale becomes possible? There are many good ideas that are just too expensive relative to impact to run for millions of people.
What is the external validity? We test whether an intervention that worked in one country, for instance, works in another and can be generalized across the specifics of the original experiment.
Are we changing the underlying conditions in a location that made an intervention work in the first place? Economists call that the general equilibrium effects. For instance, are we changing wages or prices with an intervention that undermine the intended welfare benefits?
We also test for an other externalities: Are there any other unintended consequences that would affect an intervention’s impact?
We look closely at the political context for a potential program. Is there potential for funding; partnerships with local implementers or government?
What would happen if we didn’t work on this program? If the answer to that question is that someone else is likely to successfully scale the program -- or that they could do it more effectively than we can -- we will encourage and support them as best we can, and get out of their way.
Pressure testing the interventions in our pipeline is often an iterative process where several questions are examined at the same time and for different locations.
Designing and Building for Scale:
If a project passes all of these checks with reasonable certainty, we design and build a logistically sound delivery platform with the right partners and systems to scale rapidly and exponentially.
In this phase, we:
Secure the necessary financing, ideally over several years, to grow quickly and across locations;
Streamline and codify all business processes - the steps required to execute and rapidly deliver the product in various locations. This may include technical assistance protocols for local partners;
Build and roll out software platforms that increase efficiencies and reduce costs;
Develop the necessary political relationships and alliances with local partners, such as governments and implementing partners;
Develop effective tools for measuring our key performance indicators; and
Recruit the right talent for rollout and growth.