Thanks for these thoughtful questions, and you're right that scaling hasn't been straightforward.

1. The favorable exchange rate has significantly improved our dollar-denominated cost per infant. The Naira has depreciated by roughly 4x relative to the dollar since 2020, which is the dominant driver of that reduction. At the same time, we've increased our total cash transfer from N4,000 to N11,000 to keep the incentive meaningful for caregivers in light of inflation. We also track budget-versus-actuals in detail and review operational trends weekly. That visibility has helped us adapt to many of the realities of operating in remote and insecure areas while also identifying cost efficiencies. Managing costs at scale requires continuously responding to changing conditions such as inflation, fuel prices, security dynamics, and transportation constraints.
2. Expansion to other geographies is an open question. We've reviewed different countries, weighing factors like security, existing health infrastructure, disease burden, vaccination coverage, and population density (some top countries include Niger, DRC, Chad, and Cameroon). We may extend our coverage surveys to Niger as a near-term step because we think there may be some important similarities to northern Nigeria and potential gaps in existing vaccination data that could help answer key questions related to estimated cost-effectiveness. We're also exploring a lower-cost delivery model (see below) that could make expansion more viable in contexts where our current model would not be cost-effective to operate.
3. We think the most promising path to improving cost-effectiveness is through alternative incentive modalities. We're actively exploring how to deliver cash incentives through clinic staff or community mobilizers. If we can develop a lower-overhead transfer method that reaches caregivers directly at the point of immunization with appropriate verifications, it could unlock geographies where our current delivery model is not cost-effective enough to operate. We’re interested in understanding what becomes possible when delivery costs fall substantially by challenging assumptions and adapting the model to different operational contexts.

Mellex, thanks for sharing your concerns and questions. Likewise, we appreciate all the work Taimaka does to address malnutrition and have appreciated the lessons your team has shared with us on ORS/zinc distribution!

1. Preventing duplicate enrollment and vaccination is something we think a lot about. In 2024, we tightened our requirements so that the same caregiver must present the child at each visit after the initial visit. We also use caregiver biometrics as a proxy, indexing photos from all disbursements in a 20km radius. We have found the biometric review process to be quite accurate in identifying duplicate individuals. For example, we’ve had to exclude false positives where clinic staff appearing in the background of photos were identified and then linked to program IDs as potential duplicates. Over the past year, we’ve averaged a duplicate enrollment rate below 2%. When a caregiver is identified as a duplicate, they are blocked from receiving future disbursements. We have multiple opportunities to identify a duplicate caregiver throughout the RI schedule, which makes it difficult to bypass the system repeatedly. However, if you do hear of specific cases or clinics where this is happening, please do share that with info@newincentives.org, and we will pass those along to our Audits team to investigate. They will assess trends at nearby clinics and compare them against the duplication rates we are detecting in the area. This type of triangulation really helps us improve the system.
2. To enroll in the program, children must be under 1 year old. This reflects the period when routine immunization has the greatest impact in preventing childhood disease and mortality. We see that in many states where we work, there are still large numbers of 6 to 12-month-olds not being vaccinated, and thus our work continues to focus on reaching younger children (both at the clinics and through outreaches/campaigns) with the goal of reducing the number who reach age 2 without completing their routine immunization schedule. We are also starting to think about missed opportunities for co-coverage of essential health interventions such as growth monitoring and malnutrition services.

Vinay, thanks for advocating for this work. One of the most effective tools we've found for addressing misconceptions is consistent, transparent data sharing. Over the past few years, we've invested in equipping our team to share program and household coverage survey data regularly with key stakeholders and implementing partners. Examples can be found here and here. We also share regular updates through our stakeholder newsletter and blog.

On government buy-in: it's accurate that some actors remain skeptical, but we collaborate closely with several arms of the national government, including the logistics division of NPHCDA. We also invest deeply in relationships with state health authorities who oversee primary healthcare. We have found that regularly sharing data and reports with stakeholders across different levels of government has helped us identify blind spots, improve operations, and strengthen a culture of last-mile data use and operational visibility.

Tony, thanks for your kind words!

1. We’ve had some key (and difficult) pivots along the way. During our early years, one of our first program iterations utilized CCTs to prevent mother-to-child transmission of HIV. While effective, it wasn’t scalable because the number of HIV-positive women we were identifying on the ground didn’t match with publicly available data (learn more here). More recently, we pivoted from ORS and zinc co-pack distribution to ORS only. This decision was made after reviewing the available evidence and consulting with implementing partners and state health authorities (learn more here). Following the evidence sometimes means we fold an entire program; sometimes it means we make a program design pivot. But each time we make that hard decision, we become more committed to following the evidence even when it’s not the popular choice.
2. One thing that may not be obvious from the outside is how much of our work revolves around operational visibility, feedback loops, and continuous learning. Running a program across thousands of clinics means we are constantly trying to understand not just what should be happening, but what is actually happening on a given clinic day. One lesson we’ve learned while scaling in difficult environments is that policy, reported implementation, and operational reality can diverge meaningfully. Over time, we’ve built systems that help us identify discrepancies, surface patterns, investigate anomalies, and continuously refine the quality of implementation. That operational learning process has probably been one of the biggest drivers of our scaling journey. The key is not just collecting data, but building systems that surface discrepancies, create feedback loops, identify operational gaps, and continuously improve execution quality.
3. Our team noticed the same conversations happening at Skoll. Government engagement is extremely important and must be embedded throughout the program. However, we think about it not primarily as a funding pathway, but as a legitimacy and sustainability pathway. We operate under state health authority approval, and that integration is core to how the program functions. There's been real discussion in global health over the past decade about the limits of donor-funded programs and the need for government ownership at scale. We take that seriously. But we're also realistic that government fiscal capacity and priority-setting in northern Nigeria right now make direct government funding of the CCT program unlikely in the near term. One way we are considering further integration is to explore a lower-cost model in which incentives are provided by clinic staff or community mobilizers rather than by dedicated field staff. This could also unlock new geographies within Nigeria and beyond if we can develop a model that is suitable for different operational contexts.

Simon, thanks for this question. We will soon have new results on this. An RCT of the program found a modest but statistically significant effect on first-time clinic visits: children in treatment areas were 5 percentage points more likely to have ever been taken to a clinic than children in control areas (see page 37). We're now working with our research partner to assess whether our coverage surveys detect a similar pattern in facility visits across the population. That analysis is ongoing and will be published on our website.

Thanks, Toby! And thank you for all you do to cultivate these conversations on the EA Forum.

1. In Nigeria, we work across 7,000+ public health clinics that already provide childhood immunizations. Embedding our program within existing health infrastructure is essential and wouldn't be feasible in environments without a relatively stable vaccine supply chain and widespread clinic access. Northern Nigeria also has low vaccination coverage, high disease burden, and large birth cohorts, creating substantial room for impact. In higher-coverage contexts, marginal returns may look quite different.

   We also invest heavily in addressing vaccine hesitancy: participating in village meetings, discussing questions from fathers, and engaging religious and traditional leaders to mobilize their communities. What that looks like varies considerably by setting.

   What I think does generalize is the core logic: small financial incentives that offset practical barriers, such as transportation costs and lost income, can meaningfully shift caregiver behavior around routine health visits. At the same time, the importance of operational visibility and verified delivery at scale feels broadly underappreciated. How those principles are operationalized, at what cost, and through what systems would need to be worked out carefully in each new context.

2. Honestly, this is something I think about a lot. Our cost-effectiveness has improved significantly over time due to both operational efficiencies and favorable exchange rates, but we recognize that either can shift, and the program’s CEA could change at any time. There are three ways I think about managing this.

   First, we are continually trying to simplify systems, reduce low-value operational burden, and identify where additional complexity is no longer improving program quality. We do this both to improve our cost per infant and to build the operational discipline that makes the program resilient to external changes.

   Second, we've been investing in making our program genuinely reversible, such as developing the systems and governance to responsibly exit a geography when the evidence indicates it's time to do so, rather than continuing out of inertia.

   Third, we're exploring how to improve cost-effectiveness by layering additional interventions onto our existing platform (such as ORS distribution) and testing lower-cost incentive models. The latter could unlock new geographies within Nigeria and beyond, and potentially serve as a graduation model for areas where our current approach is no longer sufficiently cost-effective.