Replicating Parts, not the Whole, to Scale
Rather than replicating “one-size-fits-all” solutions across different settings, international development innovators need to identify the core aspects they can effectively and efficiently scale up.
When it comes to solving the world’s development challenges, the innovation agenda needs to expand beyond the invention of cool, cutting-edge solutions, and zero in on addressing the more mundane but arguably more important tasks of implementation and scaling. Through innovative technologies, processes, and policies, we in the global development community have figured out, more or less, how to mitigate the risks of poverty, provide clean water, get kids to school, reduce malaria, and encourage people to save. We have viable solutions. We have set up pilot projects and done field experiments. We have collected the data. And we’ve created an industry of frugal innovation. But we have not done nearly as good a job at extending the reach of those innovative solutions as far as possible; we have failed, generally, to bring solutions to scale. Though prevailing challenges in health, education, empowerment, and development are truly global in scope, innovative solutions to these challenges have tended to remain local in practice.
Most good development solutions, especially those aimed at improving the lives of the poor and vulnerable, find their origins in highly localized settings. We can’t help this, actually; innovators often demonstrate proof of concept for a novel idea in single villages and then refine interventions by bringing in a few more villages. Rarely do we roll out solutions in full-scale. So then how do we scale up? Social innovators and entrepreneurs continue to struggle with and debate this question. It seems to us that there are two dominant approaches to scaling; each is defensible on its own, but neither is optimal on its own.
- Custom Design: This approach privileges local variables and particularities, often in the extreme. Innovators design solutions for specific contexts and clients. The anthropologist in all of us favors this kind of approach, as it ensures that interventions work effectively where they are intended to work. Scaling up this way invariably takes more time and careful consideration, but people are more likely to adopt these interventions in the field. The problem, however, is that designing solutions to fit specific contexts rarely generates the efficiencies and economies of scale that are critical to scaling.
- Replication: The second approach maximizes the economies of scale. The economist in us reasons that replicating solutions makes scaling up more likely from a cost perspective. In scaling by replicating, we transfer local solutions elsewhere with little consideration of other contexts; the assumption is that what works in one locale will surely work in others. The problem, as many point out, is that scaling up by strict replication is tantamount to rigid standardization; it overlooks the imperatives of end-user adoption and utilization. The development space is littered with examples of well-meaning solutions that no one actually uses.
Replication in toto won’t always work, and custom design is too costly. We propose a practical way forward in scaling up innovations that brings together insights from both these approaches. Effective scaling, we contend, rests with the proper and early identification of the efficiency core of a proven solution: the specific part, rather than the entire solution, that we can replicate with minimal modification to fit varied local contexts. For example, the efficiency core of the Kenyan mobile banking system M-Pesa is the simple-to-use platform technology SMS input, which others can easily replicate elsewhere in the developing world. The efficiency core of the cost-effective Aravind Eye Hospital in southern India—a great example of a scalable intervention—is the high-throughput patient model (multiple surgeries in rapid succession), driven by demand-creation mechanisms such as village screening camps. This particular aspect of the Aravind eye care system, copied elsewhere, effectively reduces the cost of each procedure without compromising health outcomes.
So to scale up innovative solutions, implementers must replicate only those aspects of a proven solution that others can replicate in different settings—the efficiency core. But importantly, the core of the solution must adapt and undergo minor modifications to fit local contexts well enough so that people will actually use it. It is not about custom design, but about making a replicable intervention more palatable, attractive, or suitable for a given context.
A great example of effective mass replication and local adaptation is Sprinkles, a micronutrient intervention that has scaled successfully in six countries. This powdered vitamin and mineral supplement offers home-based nutrition fortification to young children and infants; users can sprinkle it over any semi-solid food. The iron is microencapsulated to mask its strong metallic taste, and the supplement does not change the color, texture, taste, or smell of food.
The replicable efficiency core of the Sprinkles intervention—inspired by fast-food ketchup packets—is the pre-measured, single-serving sachet, which people can easily integrate into highly ritualized practices of daily food preparation anywhere. But while Sprinkles always comes in the same sachet form, the micronutrient powders themselves vary according to local needs. In Mongolia, for example, Sprinkles adjusts vitamin D dosages to meet the nutritional needs of local populations, localizes manufacturing to reduce costs, and features local language instructions and local artwork on the sachets. These are adaptations, not custom designs.
Innovation in global development is about devising innovative solutions; it’s also about scaling up successful interventions so that more people can benefit. As we consider the post-2015 Millennium Development Goals agenda, we must pay greater attention to scaling and implementation—not how to bring good ideas to scale, but rather, what parts of a good idea to scale.