This article originally appeared in the Spring 2008 issue
of the Stanford Social Innovation Review.
In virtually every for-profit industry, stiff competition breeds economic efficiency. Success hinges on producing more goods or services at a lower cost without compromising quality – that is, on increasing productivity. The classic example of the benefits of productivity gains is Henry Ford’s adoption of the assembly line, an innovation that allowed his company to build more reliable cars at a faster pace and at a lower cost—transforming the automobile from a plaything for the wealthy to an everyday tool for ordinary Americans.
But in the nonprofit world, organizations have few external incentives to operate more efficiently. Indeed, many people are actually suspicious of efforts to increase nonprofit productivity. At a recent meeting of 20 nonprofit leaders, for example, we presented some of the Bridgespan Group’s findings about the growth in size of nonprofits that serve young people. When we broached the topic of economies of scale—that is, the idea that larger organizations can produce more goods and services for less—most attendees said that they simply could not increase their productivity without compromising the quality of their services.
If these skeptics are correct, then the only way that nonprofits can increase their impact is by raising and spending more money. Yet for-profit companies routinely preserve and even improve the quality of their goods and services without increasing their costs. If nonprofits could be equally savvy about how they spend their existing dollars—that is, if they could increase their productivity—they could get more bang for their buck.
Some nonprofits, such as Teach for America, have achieved exactly this result. Since its inception in 1989, the organization has recruited and trained 17,000 college graduates to teach in some of the country’s neediest schools. Between 2001 and 2005, the cost to develop each Teach for America teacher increased 12 percent each year. But at the same time, the percentage of Teach for America teachers who hit their performance benchmarks increased about 24 percent each year. And so although the organization’s cost per teacher increased during this period, its cost per successful teacher decreased by approximately 10 percent each year.
Teach for America is not the only nonprofit that has increased its productivity. Our recent study reveals that Jumpstart and Year Up, two other nonprofits serving young people, have also found ways to deliver the same or better services without increasing their costs.
So productivity gains can happen—nonprofits can increase their efficiencies. But they don’t just happen. Productive nonprofits standardize their best practices, invest in essential people and processes, manage their costs, and measure their progress. Adopting these practices is not easy, because of the nature of nonprofits’ work and prevailing funding practices. Nevertheless, Teach for America, Jumpstart, and Year Up all managed to overcome these obstacles and multiply their impact.
Not surprisingly, these findings largely dovetail with those of researchers who have examined productivity in other sectors. But although there are scores of studies related to productivity in the forprofit sector, we have not been able to identify comparable studies in the nonprofit sector. We hope this small study sparks many more.
Two Paths to Productivity
Measuring productivity can be a tricky matter, even in the for-profit sector. Companies often have to go beyond tracking cost per widget produced. A classic example is that of the tire manufacturer. When a consumer buys a tire, he or she typically takes into account not just the cost of the tire but also the number of miles it will travel. Technological advancements may increase the cost of each tire, but decrease the cost of each mile traveled on it—and the latter is the metric that matters most to the consumer. The tire company that focuses on cost per tire is likely to be bested by one that grasps the true unit of value to the consumer and focuses on tire cost per mile traveled.
This same concept of value applies to the work of nonprofits. It’s not cost per output (such as a youth served) that provides a window into productivity, but rather cost per outcome (a youth who achieves the results targeted by the organization). Here’s where things get more complicated for nonprofits, though. Although it’s relatively easy to calculate tire cost per mile (i.e., the tire’s total cost divided by the number of miles it will go), it’s far harder to calculate dollars required for a nonprofit to achieve an outcome, for reasons we will explore below.
Let’s back up and define some terms. Outputs are the amount of work a nonprofit does – in other words, the quantity of the programs or services it delivers. In the case of Year Up this would be the number of students participating in the program. Outcomes, in contrast, are the results of the nonprofit’s work – in other words, the benefits for participants during or after their involvement with the organization. At Year Up this would be the number of graduating students placed in livable-wage jobs. Finally, the success rate reflects the number of outputs that turn into outcomes—for example, the ratio of students in the program to those who go on to get jobs paying a livable wage.
As the equation below illustrates, there are two paths to increasing a nonprofit’s productivity: reduce the costs of producing each output, or increase the number of outputs that turn into outcomes—in other words, increase the success rate. In practice, as these nonprofits show, organizations typically work on both facets of the equation.
Jumpstart improved its productivity by decreasing its cost per output while holding its success rate steady. Jumpstart’s mission is to work toward the day when every child in America will enter school prepared to succeed. To this end, corps members work one-on-one with preschool students and their families to incorporate language and literacy into daily activities. The organization’s output is the total number of students with whom corps members work, and its outcome is the number of students who achieve greater gains in school readiness than a comparison group.
At Jumpstart, staff salaries are the greatest expense. In 2002 management noticed that the organization’s Boston site had twice as many staff members as its other sites, yet the other sites still turned out similar numbers of children ready for school. Armed with these data, Jumpstart could reduce the number of its Boston-region staff by almost half without worrying that it would be compromising the quality of its outcomes. In other words, Jumpstart reduced its cost per output by halving the number of salaries it paid at its Boston site while serving the same number of children equally well.
Year Up likewise increased its productivity by reducing its cost per output. But the organization first had to spend money to save money. As part of its intensive training program, Year Up pays its students a biweekly stipend for the hours they have worked at their internships. Historically, Year Up gathered time sheets manually and processed the checks internally—a time-intensive process. Seeing more growth and expense on the horizon, the organization invested in a centralized mechanism for collecting student data at all of its sites. An automated process then translates the data into a format accessible for a payroll outsourcing partner, providing a scalable solution to a potentially expensive challenge. And so the organization is able to train the same number of young people for less cost.
Teach for America, in contrast, improved its success rate. It did so by increasing the number of its recruits who turned into truly effective educators. As this example illustrates, efficiencies sometimes come from spending more money, not less: Spending more on each teacher increased the number of teachers who later proved successful in front of their classrooms. By spending more money on outputs, organizations may wind up spending less on outcomes because they increase their success rates.
The converse is also true. Reducing the cost per output can increase the cost per outcome if the measures taken to cut costs hurt the organization’s success rates. Consider, for example, a youth development organization that assigns youths to case workers. If its leaders decide to try increasing caseloads as a cost-saving measure—say by having case managers work with six youths each instead of its past standard practice of working with four—they could very well be disappointed. The larger caseloads could decrease the effectiveness of workers’ interactions, with the resulting decline in the organization’s success rate more than trumping the salary cost savings. Such unintended consequences are one of the reasons that it is unwise to focus exclusively on either reducing costs or increasing success rates. Organizations have to consider the impact of their efforts on both variables to achieve the desired results.
How to Increase Productivity
Increased productivity doesn’t just happen. Jumpstart, Teach for America, and Year Up observed four management basics: standardize best practices; invest in staff and critical activities; manage costs aggressively; and measure progress.
Standardize Best Practices
All three organizations discovered and spread their most cost-effective practices throughout their organizations.  Jumpstart, for example, has standardized every aspect of its program and made these practices available in hard copy guidebooks as well as on its intranet. Corps members can sign into the system at any time and get lesson plans and curricula. Teach for America has created a centralized information system to capture and share best practices. This sort of standardization helps to keep costs down by increasing efficiency: Staff spend less time reinventing the wheel and more time executing effective practices. It also ensures that all sites are using the practices that have been shown to lead to the best outcomes, which can have a positive effect on the organization’s success rate.
Invest in Staff and Critical Activities
All three organizations invested more in the activities most important to the success of their programs and the sustainability of their organizations. For example, in 2000 Teach for America embarked on an ambitious five-year growth plan. The organization knew that in order to attract the best college graduates it would have to invest more in on-campus recruiting. It also added staff positions to its teacher training and support teams, invested in better data management and communications systems across regions, and made a commitment to pay higher salaries to attract and retain better talent. Teach for America’s leaders point to highly qualified and well-resourced teachers as propelling the organization’s healthy success rate, and strong teacher retention as offsetting recruiting and training investments.
Jumpstart also invested in building its management capacity, noting that when it comes to salaries, “paying more costs less.” Although this may sound counterintuitive, the logic holds: Better compensation reduces turnover, increases the average experience level of the staff, and builds institutional knowledge.
Manage Costs Aggressively
All three organizations paid a great deal of attention to managing their costs, particularly with respect to major costs like salaries. At Year Up, overhead and training stipends—for both the participating students and the companies that provide them with internships—are the organization’s two biggest costs. To manage overhead, Year Up constantly monitors its administrative and personnel ratios to keep them as low as possible, even as class sizes grow. Training stipends receive similar scrutiny. Year Up analyzes and experiments with stipend levels at each of its four sites to determine the minimum level necessary to secure target student and company participation levels, thus ensuring that its cost per output remains low.
All three organizations established performance measurement systems to give them timely feedback about their costs, outputs, and outcomes. Year Up, for example, has created a set of PowerPoint performance dashboards for reporting regularly on six critical areas of their work: recruiting and retaining talented staff, cultivating a solid student pipeline, teaching marketable skills, providing high-quality service to corporate partners, positioning students for ongoing success, and creating a sustainable program infrastructure. The management team then sets performance benchmarks and rates its progress against each benchmark.
Take the critical area “positioning students for ongoing success,” for example. Year Up breaks this area down into three metrics: average starting wage, graduate employment, and college attendance by class. The associated dashboard charts the data in these three subareas and then rates the performance of each relative to the benchmark goals, using a green, yellow, or red traffic light icon.
Why Nonprofits Don't Calculate Productivity
The three organizations in our study all tracked costs and measured outcomes. Yet none had brought its data together to calculate cost per outcome, even though all appreciated the value of this metric. One reason for this oversight is that as each organization accumulated experience, it refined its definition of a successful outcome, complicating the process of tracking the metric. Consider Jumpstart, which uses a tool called the “School Success Checklist” to assess its students’ language, literacy, initiative, and social skills. As Jumpstart learned more about its programs and participants, it changed the checklist. In the 2001-02 school year Jumpstart increased the number of items from 10 to 15, and in 2003 it revised the ratings and subscales for each item yet again. To participate in our study, Jumpstart staff had to dig back into its historical data and restate them in terms of its new school success checklist.
Another challenge to tracking cost per outcome is the lag between the time when an organization incurs program costs and the time when it can measure outcomes. For example, Jumpstart cannot possibly know whether students have improved until the end of the school year. In contrast, the organization can readily calculate cost per output (specifically, cost per tutor hour), a more immediately useful measure that helps Jumpstart’s managers understand what is happening right now. As a result, the internal pressure to monitor costs naturally points people toward cost per output rather than cost per outcome.
The time lag between when an organization incurs costs and when it realizes outcomes is even more pronounced when it makes a large investment in its program or infrastructure, because these costs may all show up in one year. The benefits, in contrast, may take years to appear. For example, Teach for America’s cost per outcome actually increased for several years while its investments were taking hold, only to decrease in the longer term.
Matching expenses to the outputs and outcomes of one specific cohort of youth is also difficult. For example, Teach for America’s costs for a given fiscal year include the cost to support teachers who are already in the classroom and the costs to recruit and train new teachers, who will not teach until the following year. To calculate Teach for America’s cost per outcome for this study, we had to break down and reassemble its cost, output, and outcome data.
Without broader-based research, we cannot say how widespread reductions in cost per outcome are in the nonprofit sector. Nevertheless, on the basis of Bridgespan’s work over the years with more than 100 nonprofits, we suspect that nonprofits do not regularly or fully realize such reductions. The reasons are many, but several additional barriers (besides the ones already mentioned) stand out in particular.
One of the impediments to achieving increased productivity is the lack of funding for nonprogram expenses. Building information systems to track outputs, outcomes, and costs requires money. The staff time needed to analyze the data is also not cheap. Many nonprofits simply do not have the money to make these investments, and foundations and other donors are often unwilling to provide the funding. As a result, many nonprofits have only a rudimentary understanding of what their true costs per outcome and output are.
A second reason why nonprofits do not pay more attention to productivity is that they face little pressure to do so. Some of the organizations in our study cited an internal desire to decrease their cost per outcome, but none of them mentioned any external pressure to do so. Businesses face unrelenting external pressure from customers and competitors to reduce prices while maintaining quality. Companies are also under pressure from investors to increase their earnings per share, which they can often achieve by increasing productivity.
When funders and other stakeholders do focus on nonprofits’ costs, they most often scrutinize cost per output, not cost per outcome. Year Up, for example, usually reports the cost per participating student, not the cost per student placed in a living-wage job. Similarly, Teach for America usually reports the cost to recruit, train, and place a teacher, not the cost to recruit, train, and place a successful teacher. Some funders even provide a fixed amount of funding per output. This focus on cost per output can be counterproductive if it is not married to a focus on maintaining or improving outcomes.
The difficulty of achieving scale presents another challenge. In the for-profit sector, building economies of scale is a well-understood and oft-practiced way to increase efficiency. Although scale does not automatically decrease cost per outcome, it often allows an organization to spread fixed costs over more activities and to accumulate experience faster. And so growth can help an organization reduce costs and learn more quickly. The three organizations we studied are larger than average, but none is among the largest in the sector. Given the fragmented nature of the nonprofit sector, with the vast majority of organizations having annual budgets under $1 million, it may be hard for the average nonprofit to achieve economies of scale.
Yet another barrier is the failure to sustain activities long enough to decrease their costs or improve their success rates. Foundations have a reputation for being more interested in supporting the next exciting idea than in continuing to support a proven idea or program. Accordingly, nonprofits that depend primarily on foundation support may find it difficult to sustain a given program – and therefore to accumulate the kind of experience that can lead to savings. The fact that two of the three organizations we studied received more funding from public sources than from foundations may not be entirely coincidental.
Jumpstart, Teach for America, and Year Up all give real-world proof that maintaining quality while reducing cost is not only a theoretical goal but also a practical possibility. We hope that this research will inspire more nonprofits to track and manage their cost per outcome. We also hope that it will encourage more funders to shift their focus from cost per output to cost per outcome and to provide nonprofits with the long-term unrestricted support that will enable them to do the same.
We recognize that changing the rules of the funder-grantee game is never risk-free. Wrongly applied, external pressure on nonprofits to focus on cost per outcome could simply become yet another reporting burden on capacity-strapped organizations. Foundations may also be tempted to compare two organizations’ cost per outcome without taking into account important differences between them, such as the populations they serve or the kinds of overhead that they have to bear. Equally, it could lead nonprofits to shift their costs, to serve less needy populations, or to make unsustainable cost cuts.
Done right, however, reducing cost per outcome will lead to more bang for the nonprofit buck – a greater impact across the whole range of issues that nonprofits grapple with on society’s behalf. And impact, after all, is the true bottom line of nonprofit work.
The authors thank the nonprofit executives of the profiled organizations for sharing their experiences and data. We also thank our colleagues from the Bridgespan Group—Jeff Bradach, Meghan Gouldin, Jennifer Lee, Gail Perreault, and Nan Stone—for their thoughtful contributions.
Our Research Methods We addressed two general questions in our study: Have any nonprofits maintained, or even improved, their outcomes while reducing costs? If so, how did they do it? We first identified nonprofits with a reputation for achieving results and managing costs. To keep the analysis simple, we focused on single-program, direct-service nonprofits in the fields of youth development and education. Three organizations emerged from our preliminary investigations – Jumpstart, Teach for America, and Year Up. For each organization we collected and analyzed a minimum of five years worth of cost, output, and outcome data. We also conducted detailed interviews with the organization’s executive director, chief financial officer, director of research and evaluation, and other key people. Finally, we compared our findings to secondary research on efficiency studies in the public sector and in for-profit service companies.
In light of the composition and small size of our sample and the preliminary nature of the research, our findings are, at best, suggestive. Nevertheless, they do shed light on the central question of whether nonprofits can maintain, or even improve, the quality of their outcomes while reducing costs. –A.N. & R.S.
 Notable studies in public sector and for-profit service industries, besides those mentioned elsewhere in this article, include Vernon Altman, Marty Kaplan, and Ravi Vijayaraghavan, “The Challenge of Cutting Costs,” Tele.com (May 2003); Theodore Levitt, “The Industrialization of Service,” Harvard Business Review (September 1976); Theodore Levitt, “Marketing Intangible Products and Product Intangibles,” Harvard Business Review (May 1981); Brent Keltner, David Finegold, Geoff Mason, and Karin Wagner, “Market Segmentation Strategies and Service Sector Productivity,” California Management Review (Summer 1999); and Ashish Nanda, “Scale and Scope in Professional Service Firms,” Harvard Business School Publishing (April 2006).
 Among studies illuminating effective cost management in the public sector is John Drew and Nina Bhatia, “Applying Lean Production to the Public Sector,” The McKinsey Quarterly (June 2006).
 For further research on the value of introducing technology to transaction-level processes in a service organization, see Michael Van Biema and Bruce Greenwald, “Managing Our Way to Higher Service Productivity,” Harvard Business Review (July 1997).
 For additional reading on spreading best practices in the service sector, see Theodore Levitt, “Production Line Approach to Service,” Harvard Business Review (September 1972).
 The service-profit chain provides a comparable framework for understanding the link between employee satisfaction, profit, and growth. See James Heskett, Thomas Jones, Gary Loveman, W. Earl Sasser Jr., and Leonard Schlesinger, “Putting the Service-Profit Chain to Work,” Harvard Business Review (March/April 1994).
 For further research on the measurable impact of using benchmarks to improve performance and reduce employee turnover, see Fred Reichheld and Christine Detrick, “Loyalty: A Prescription for Cutting Costs,” Marketing Management (September/ October 2003).
 For additional research discussing the challenges in integrating performance metrics that tie to the organization’s strategic goals, see Keith Leslie and Catherine Tilley, “Organizing for Effectiveness in the Public Sector,” The McKinsey Quarterly (2004).
 For additional reading on economies of scale and the experience curve, see Pankaj Ghemawat, “Building Strategy on the Experience Curve,” Harvard Business Review (March/April 1985); and Bruce Henderson, “The Experience Curve Reviewed,” Perspectives (1973).
 For a more complete discussion of funding sources, see William Foster and Gail Fine, “How Nonprofits Get Really Big,” Stanford Social Innovation Review (spring 2007).
 System costs are the total costs required to achieve the target outcome. Teach for America, for example, does not pay the cost of its teachers’ salaries; these are covered by the school systems where its teachers serve.