Many in the field have written about replication as one approach to scaling impact. (You'll find several examples in "Related Reading.") In the vast majority of cases, however, it is becoming clearer that strict replication of a program by a single organization can only get us so far.

Take Year Up, a nonprofit that delivers an intensive training program to prepare low-income young adults for professional jobs. Year Up has a proven track record of results with program participants. And it has served nearly 20,000 of them since its founding in 2000. Yet some five million young adults in the United States remain underemployed or not in school. Eager to do much more, Year Up (like many other organizations) is complementing its replication strategy with another path to scaling impact—its Systems Change initiative.

Why are replication strategies often so limited? Among the explanations: philanthropy's bias for new over proven initiatives; thinning government funding flows; insufficient coordination and integration with complementary services; and variations in local contexts that bridle against prescriptive "proven" models. Consequently, many exemplary social entrepreneur-driven programs of the last few decades remain tiny compared to the scale of the problems they seek to address. That's not to minimize their impact: 20 years ago it was unimaginable that they would be at the scale they are today.

Let's focus on just one of these barriers to replication: variations in local context. Today's nonprofit organizations face complex, dynamic problems that vary over time and by geography. These variations pose significant challenges to purveyors that hope to take what worked in one place and spread it to another. Evidence of this problem is commonplace. Take J-PAL researchers' experience replicating a summer reading program that had improved reading comprehension in a Virginia school district. Its efforts to take the program to schools in California failed to generate similarly positive results. Why the difference? The researchers hypothesized it was because in California—unlike in earlier implementations—the children participating in the program were English-language learners.

Clearly both what we do (i.e., the program—which is typically the focus of rigorous randomized control trials designed to establish efficacy) and how we do it in different contexts (i.e., the implementation) matter for success. Achieving impact at scale requires rapidly learning and scaling insights about both of these factors.

As the sector has increasingly realized this, field-building intermediaries have emerged to act as platforms for "scalable learning," a term coined by John Hagel III, John Seely Brown, and Lang Davison. These platforms help a field's actors build capacity to execute, learn, and adapt. They create diverse networks of individuals and organizations working on the same issues and serve as central hubs to accelerate learning. These organizations do not focus on the replication of a single program. Rather, they identify multiple programs, practices, or permutations that yield results; prompt and capture learning about how these solutions need to vary by context; and engage networks of actors to shift practices and advocate more broadly to spread what works—and to continue the learning process. Examples of platforms for scalable learning include Community Solutions, Ariadne Labs at the Harvard T.H. Chan School of Public Health, No Kid Hungry, and ChildObesity180 at Tufts University.

To dive into one example, Ariadne Labs is a joint center between Brigham and Women's Hospital and the Harvard T.H. Chan School of Public Health. It aims to deliver better care at the most critical moments in people's lives (such as childbirth and serious illness), and it does this by creating scalable health-care solutions such as the World Health Organization Surgical Safety Checklist and Serious Illness Conversation Guide. Thousands of individuals and health-care facilities around the world use its tools. Ariadne Labs works closely with this network to implement, evaluate, and adapt tools—facilitating rapid learning across its many partners.

"Platform for scalable learning" is a role played by various types of field-building intermediaries, including field catalysts, capability specialists, evidence-action labs, place-based backbones, and hubs of aligned-action networks. Despite this range of intermediary types, organizations assuming this role share a number of commonalities in the mindsets they embrace, the features that distinguish them, and the challenges they face. Here are some of our latest observations on each of these commonalities, based on our preliminary review of organizations that appear to play the role of a platform for scalable learning.

Common Mindsets

A multidisciplinary mindset: Platforms for scalable learning engage in a diverse mix of activities, all within the boundaries of one organization. These include a mix of basic research, action research, direct service, implementation support, learning networks, communications, movement building, and policy work. For example, the Urban Education Institute at the University of Chicago (UEI) bridges education research and practice to foster greater equity and excellence in public schooling. Across four units, UEI conducts rigorous applied research, trains teachers and school leaders, operates a pre-K–12 public school, and provides research-based tools and resources to schools in 62 major cities across 34 states. Together, UEI's units produce research- and practice-based knowledge on what matters most for school improvement and student success. For other types of organizations, this breadth is a common red flag that it's spread too thin. But for these intermediaries, breadth is critical to their ability to bridge the theory-practice divide.

A learning mindset: They view learning from their networks as key to identifying, improving, and scaling solutions that work across a wide variety of contexts. Take PolicyLink, an organization that seeks to advance racial and economic equity by working with others and Lifting Up What Works® to highlight examples of successes in low-income communities and communities of color across the country. PolicyLink grounds its work in ensuring all people have economic security, live in healthy communities of opportunity, and benefit from a just society. To accomplish this result, PolicyLink engages a broad network of leaders and organizations to understand and scale solutions and engage advocacy and policy to advance equity. The organization's work in Pittsburgh—part of its All-In Cities Initiative—is a good example of this local engagement. In Pittsburgh, PolicyLink collaborated with local philanthropy, government, and nonprofit executives to build a policy framework and support progress towards its implementation and intended results: affordable housing and good employment. Proximity plays a key role in many of these platforms for scalable learning, sometimes via networks of practitioners and in other cases by operating a direct service model themselves.

An adaptation mindset: Organizations that act as platforms for scalable learning see adapting a program to the local context as critical to determining the appropriate implementation strategy and, thus, to getting the desired results. For example, Ariadne Labs developed a Safe Childbirth Checklist and coaching intervention that helps establish basic practices—such as hand washing—to prevent major causes of death during childbirth. When implementing it with partners in Uttar Pradesh, India, the organization found that the program "achieved significant gains in the quality of care during labor and delivery, but the improvements were insufficient to reduce death rates." The reason for the lack of change in death rates, Ariadne Labs determined, was context: "persistent gaps in skills, in supplies, or in systems for care of complications" in the Uttar Pradesh health system. Now, the organization and its partners are working on improving those critical enablers of its checklist and coaching approach. This dynamic adaptation of a core idea—based on learning from different contexts—is critical to achieving impact at scale.

Common Features

Their research about "what works" is deeply informed by practice. In many cases, their own small direct-service programs serve as learning laboratories. These intermediaries pressure test big ideas in real-world situations, creating a tighter link between theory and practice that helps learning and improvement happen at a faster pace. For example, through its Brownsville and North Hartford Partnerships, Community Solutions (an organization that is working to end homelessness) engages deeply with local partners—including residents, nonprofits, businesses, and government agencies—in neighborhoods of concentrated poverty to understand the conditions that produce homelessness and to develop strategies and tools for its prevention. It shares what it learns with other communities, including the network of more than 80 partner communities in its Built for Zero initiative, a national movement to end veteran and chronic homelessness.

They establish rigorous data collection and outcomes measurement systems. Cross-site comparisons, in particular, help these intermediaries gain insight into what is (or is not) working in different contexts, and why. They use data to orient learning and inform strategy for achieving progress, enabling improvement over time. For example, in its work to combat childhood hunger in the United States, No Kid Hungry discovered that the field did not possess an adequate measure of hunger. The government tracks food security, a socioeconomic measure of access to food, but it has no measure for hunger, a distinct (though related) physical condition. Recognizing that the field needed a truer measure of hunger to understand the state of the problem and progress against it, No Kid Hungry developed a measure rooted in common sense: if children receive three nutritious meals a day, then they are not hungry. The organization tracks and closes the gaps in participation of eligible children in free and reduced-price meal programs (including school breakfast and lunch, and after-school and summer meals) to ensure children are getting three meals per day. Today, the majority of No Kid Hungry's state and local campaigns use this metric. The organization can monitor progress, see which campaigns are succeeding, and share insights from successful campaigns with efforts in other locations.

They emphasize the spread of principles and problem-solving frameworks. For example, Harvard's Center on the Developing Child shares its IDEAS Impact Framework™—a design process for developing, evaluating, and iterating on programs for children and families facing adversities such as economic hardship, child maltreatment, or maternal depression. The framework is rooted in precisely defining and measuring a program, iterating on it in fast-cycle fashion, co-creating it with researchers, practitioners, and community members, and sharing learning using common measures. Regardless of the specific nature of an intervention or where it is being implemented, the framework helps researchers and practitioners develop a more precise approach to their program's theory of change, supporting materials, and evaluation plan. All of this is informed by the science of early childhood development, and ultimately furthers a greater understanding of what works, for whom, why, and in what contexts. For example, FIND (Filming Interactions to Nurture Development) is a brief video coaching intervention focused on increasing "serve and return" interactions between caregivers and young children. Evaluations using the IDEAS framework have documented impacts on caregiver targets (increased responsiveness), caregiver outcomes (parent stress and sense of competence), and child outcomes (inhibitory control, vocabulary, behavior), with larger effects on parents with high Adverse Childhood Experience scores.

Common Challenges

Building and managing a multidisciplinary organization. These intermediaries employ individuals with a wide range of expertise and experience—often with different priorities and definitions of success. This variation can, at times, create tension (e.g., peer-reviewed academic research versus rapid-prototyping experiments)—adding to the management challenge. It is just that tension that can open up powerful learning, but it must be managed in ways that produce benefits versus unproductive conflict. As Atul Gawande, executive director of Ariadne Labs shared, "It simply is not possible to deliver system innovations capable of reaching national or global scale without people with multiple backgrounds and skills working together—designers, clinicians, technologists, data analysts, program managers, implementation scientists, and others. There is no question that this is difficult. People come to the same problem with very different vocabularies, expectations, and ways of working. We have had to spend time intentionally creating our own versions of those things, mostly through trial and error. A certain amount of birthing pain is inevitable. But the results have been remarkable. And we get better and better with every year."

Navigating the complexities of academic institutions. Many of these organizations are affiliated with colleges and universities. The benefits can be great: academic credibility and access to top-notch research talent. But those affiliations can also constrain the intermediary's ability to put ideas into practice and achieve widespread scale, as academic institutions typically aren't focused on those goals. As one leader of a platform for scalable learning noted, "Being housed within an academic institution creates both challenges and benefits. Universities aren't always nimble, so negotiating joint work with a partner can take significantly longer than it would if we were an independent entity. However, there are also incredibly valuable aspects of being here, from access to faculty and fundraising opportunities to an institutional reputation for very high-caliber work."

Finding leaders who can serve as bridges. Leaders of field-building intermediaries must be able to work across many capabilities and be trusted to authentically value and embrace them, including theory and practice, research and policy. They must be adept at developing credibility across players in different settings, from leaders of local community organizations, to funders, to academics, which is a relatively rare commodity. The leader needs to exemplify the interdisciplinary, theory-practice bridging that the overall organization needs to embody. As Rosanne Haggerty, president of Community Solutions, explains, "The key leadership challenge at an organization like this is staying focused on the goal and recognizing that many different kinds of tools, partners, mindsets, and expertise may be needed to achieve it. It is the combination of these—and the capacity to iterate continuously as you learn new information—that is at the heart of achieving results in complex settings. It is important for leaders to create the atmosphere for this kind of thinking and work to unfold—and to keep teams from locking into answers prematurely. While we have not yet established the language for these kinds of organizations or this kind of leadership, it is clear that both are crucial for impact."

If organizations can overcome these challenges, the mindsets and features discussed above combine to create a powerful opportunity for rapid learning; platforms for scalable learning can quickly identify, innovate on, and distribute "what works" to address critical problems across a wide variety of contexts.

A Critical Learning Agenda

In "Great Businesses Scale Their Learning, Not Just Their Operations," Hagel and Brown write about how the crux of competition in the for-profit sector has shifted from pursuing scale to gain efficiencies to pursuing scale to achieve better and faster learning. In today's environment, the businesses that achieve this, they argue, have the highest chances of success.

In many ways, the crucial role of accelerated, scalable learning applies to the social sector, too—perhaps even more so. If we are interested in achieving impact at scale, then, we must think more about the role that platforms for scalable learning play in elevating and spreading solutions, accelerating learning about how to adapt these solutions to diverse contexts, and sharing these learnings widely.

As many of these organizations exemplify, a more flexible, data-driven, adaptive approach may be necessary to come close to achieving impact at scale in most situations. What we need to better understand is what distinguishes excellence in this increasingly important and unique type of organization—and what capabilities, funding models, and performance measures apply best to helping them achieve their full potential.