Achieving a large impact requires bringing effective solutions to scale. Yet before recommending the scale-up of a program or policy, we need to understand if it works in multiple places, how it works, and why it works. We also need to understand the nuances and specific challenges of maintaining effectiveness at scale.

There are then four phases along the path to scale:

  1. Program design: Identify existing solutions or design innovative ones to be evaluated.
  2. Proof of concept: Evaluate if a particular solution to help the poor is effective in a given setting.
  3. Validate and/or adapt the concept: Use field replications to evaluate if and how the solution is effective outside of the original context
  4. Mobilize for scale: Promote or incubate effective solutions. Provide advice and technical assistance to fine-tune the implementation of a solution, ideally integrating it into existing systems.

The vast majority of IPA’s research to date has been in the proof of concept stage. Although an effective solution can still have an impact in that one context, (as in this case study on motivating community health workers), if we want to have an impact on millions of people in different contexts, as in the case of the ultra-poor graduation model, we need more studies that validate or adapt the concept.

In validating and/or adapting a concept beyond the original context, we show how a solution can work effectively in different places and at scale. These kinds of studies ask questions like “Does it work in Ghana as well as it does in India?” or “Does it matter who runs it, the government or an NGO?” or “Does the impact or cost-effectiveness change when we do it with 100,000 people vs. 5,000?” Studies at this stage help us to hone in on the most effective mechanism to advocate for.

Going from the proof of concept stage all the way to embedding the concept into existing government systems to impact millions can take a decade or more. So when we see something works well in more than one context, we start advocating for it more broadly, while still working to answer questions that are outstanding on how to make it the most cost-effective solution at scale.