A solution can be evidence-based and perfectly translated without producing the same result elsewhere. From its Latin America team, Busara argues that contextualization is a strategic capability, not a final adaptation.

In public policy and international development, scaling rests on a reasonable promise: if a solution worked somewhere, we should be able to learn from it and use it elsewhere. Without that possibility, every organization would have to start from scratch. Knowledge would not accumulate, countries could not learn from one another, and responding quickly to urgent problems would be far more difficult.

The thing to remember when translating solutions is that evidence generated must consider the context in which it was generated, and is relevant.  

The result of an intervention is dependent on many variables, for instance, how people respond to it, the incentives it offers, how institutions perceive it, and the material conditions around it, to name a few. When any of these conditions change, the result is subject to change as well. 

That is why the strategic question is not simply: ‘How do we implement a solution that works somewhere else?’

The more important question is: ‘What would have to be true for this solution to work here?’

Context does not surround the mechanism, it is part of it

Consider a reminder that increases attendance at medical appointments. It is tempting to conclude that the message was what made a difference.  But there are so many other factors in play: forgetting the appointment could be the main barrier but people must also be able to fit it into their schedules; have accessible ways of reaching the health facility; and trust the service waiting for them.

If the main barrier is losing a day’s income, finding transport, or anticipating discriminatory treatment, improving the message may have little effect.The causal diagnosis must reflect so. 

The same applies to interventions in financial inclusion, education, social protection, and citizen engagement. A solution may change information, incentives, or individual decisions, but its result will depend on what the system allows people to do next.

This leads to a fundamental distinction: a mechanism, a method, and an intervention are not the same thing.

A behavioral mechanism may be relevant across contexts. A research method can travel and be adapted. A hypothesis can be tested in several countries. But a specific intervention combines all of those elements with a channel, an institution, an operational sequence, and a particular population.

That combination must be validated again.

Contextualization is not ‘making it feel local’

Contextualization is often understood as changing the visible features of a solution: translating materials, replacing images, adding familiar examples, or updating the name of an institution.

These  changes may be necessary, but they do not examine the logic of the intervention.

Rigorous contextualization distinguishes at least three things:

First, whether the behavior we want to change responds to the same barriers, motivations, and norms identified in the original context.

Second, whether the institutions and frontline staff responsible for delivering the solution have the capabilities, incentives, and room for judgment required to do so.

Third, how other actors in the system will respond when the intervention changes existing processes, expectations, or relationships.

This is not about adding ‘cultural sensitivity’ to a finished solution. It is about testing whether our explanation of the problem remains valid.

Evidence does not lose value when we recognize its limits. It becomes more useful, because we can identify which parts of the learning are transferable, which conditions must be reconstructed, and which elements require a different solution.

What we learned by modeling a system in Guatemala

Our work on gender-based violence in Guatemala shows why this distinction matters.

Rather than analyze an intervention in isolation, we combined behavioral science, qualitative research, and systems thinking to understand how victims and survivors navigated the health and justice systems.

The process involved more than 80 actors, from public officials to frontline staff. We then built an agent-based model to explore 29 intervention scenarios to  anticipate how different decisions might interact with the wider system.

The objective was not to make a familiar solution appear more Guatemalan. It was to understand how confidentiality, trust, institutional coordination, and frontline practices could change its effects.

The analysis showed that measures generally recognized as good practice could produce unintended consequences if implemented without the right conditions. Changing one part of the system could alter other actors’ decisions, create new risks, or strengthen the very cycles the intervention was intended to weaken.

Context then, did not determine how to implement and intervention, it determined which intervention made sense to implement.

This is a critical distinction for any organization seeking to scale. Especially when we aim for a solution to be integrated into a system sustainably, and lead to accurately anticipated responses.  

Regional presence as learning infrastructure

Working from context does not mean assuming that someone from the region automatically understands every country, community, or institution. Latin America is not a single context. Its countries, and the territories within them, have profoundly different political trajectories, public capabilities, markets, languages, and forms of organization.

Regional presence offers something different: infrastructure for learning.

It allows relationships to extend beyond a single project. It makes it possible to work alongside researchers, implementers, public officials, and communities that understand systems from different positions. It helps build institutional memory, compare experiences without flattening their differences, and sustain learning through implementation.

It also connects knowledge produced in Latin America with experience from other Global Majority contexts. That connection should not operate as a one-way transfer of solutions. It can become an exchange of methods, questions, and perspectives among regions that are often excluded from traditional centers of knowledge production.

Regional presence does not replace research. It makes research more continuous, collaborative, and responsible.

Why Busara is in Latin America

This conviction led Busara to establish a presence in Latin America and build a team that lives and works in the region.

Today, our LATAM team and portfolio work with organizations across the region to investigate problems, design and test solutions, support implementation, and measure results. We also seek to innovate in the methods used to understand behavior within complex systems.

Our presence is not intended to turn Busara’s experience in Africa and Asia into a catalogue of answers for Latin America. It connects that experience with the capabilities, relationships, and perspectives of the region to produce something new.

This is not only about applying behavioral science in Latin America. It is about developing behavioral science from Latin America.

That matters to our partners because it improves decisions about what to test, what to fund, and what to scale. It also matters to the field: the region is not be merely a destination for ideas created elsewhere. It is a source of evidence, methods, and new ways of understanding behavior.

Contextualization is not the additional cost of taking a solution to scale. It is a condition for scaling responsibly.

The challenge is not to make an imported solution look local. It is to determine, with evidence, what can travel, what must change, and what can only be discovered by being here.