Measurement is a hot topic these days in the social sector, as increasing numbers of funders want to know exactly how their money is being used, and as nonprofits undertake rigorous evaluations to prove their programs work and attract funding for growth. But one of the most important uses of measurement is too often overlooked, and that is measurement for the purpose of learning and improving performance, or performance measurement.
The benefits of performance measurement can be substantial. Among the upsides: Organizations that measure to learn often find that they’re able to do more for their beneficiaries with less money; they’re also better able to adapt their programs to changing circumstances faster and more effectively. The problem is that the idea of actually doing performance measurement scares many nonprofit leaders. The process seems daunting, the rewards distant and unclear. As Isaac Castillo, director of Learning and Evaluation at the Washington, DC-based Latin American Youth Center (LAYC) put it, “all but a handful of nonprofits are overwhelmed by the concept of measurement as a means of improvement.”
Performance measurement needn’t be overwhelming. In fact, in the course of our work, we have seen a diverse set of organizations, of all sizes and at all stages of development, measure what they do on a regular basis and apply what they’ve learned to deliver ever greater results for their beneficiaries. Those organizations include a corporate foundation (Goldman Sachs’ 10,000 Women), a global NGO (Camfed International), a domestic, multi-service NGO (Latin American Youth Center) and a public-private collaborative (Strive). This article distills their experiences and others’, demonstrating clear benefits of performance measurement and suggesting a framework for thinking about it, informed by lessons learned in the field.
Brief note on terminology
There is, unfortunately, little consistency in how the social sector defines measurement-related terms. Throughout this paper, we use the following:
Performance measurement (also performance monitoring): Continuous tracking of data, typically by an organization’s own staff through an internal data system, for the primary purposes of accountability, learning, and improvement. The most powerful systems integrate program, financial, and organizational data, though our primary focus in this article is measurement of programs (i.e., models, approaches, interventions).
Evaluation: Discrete assessments or studies to answer critical questions about an organization’s program(s). Evaluations can be done internally or externally, to support learning and improvement, but also to demonstrate evidence and influence the field.
Impact evaluation: One type of evaluation that studies whether a change in outcomes can be attributed to an organization’s program. Outcomes proven through an impact evaluation are impacts.
Randomized control trial: One type of impact evaluation in which eligible participants are assigned at random to receive or not receive an organization’s program. While widely considered the most rigorous form of impact evaluation, it can only be used for programs that lend themselves to random assignment.
Benefits of Performance Measurement
Consider just a handful of the clear benefits of performance measurement:
Better results for beneficiaries
Goldman Sachs’ 10,000 Women is a big, multi-year investment to provide underserved women around the world with the business education and support services (mentoring, networking, advising) they need to grow their businesses and, by doing so, drive greater economic growth. Ayodeji Megbope, owner of catering business No Leftovers Nigeria Limited, successfully completed the program’s 180-hour curriculum at Pan African University in 2008, learning about business topics that ranged from operations to sales to people management. After graduation, as Megbope’s business began to grow, the 10,000 Women performance measurement system, which tracks usage of financial statements and methods of bookkeeping, among other indicators, revealed that she would benefit from additional counsel with regard to accounting. So Peter Bamkole, program director at Pan African University, recommended that Megbope work with an accountant, who helped deepen her ability to project profits and cash needs. Armed with that additional knowledge, Megbope negotiated a more favorable billing schedule with her suppliers, opened her first bank account, and created financial statements that allowed her to access capital. As a result, Megbope rapidly grew her business from $1,000 in monthly revenue before the program to $16,000 today, and she has hired seven additional full-time employees.
Lower costs to learn
Camfed is an international NGO that has financed access to education and provided business training and leadership development for more than 1 million girls and young women in rural Africa. According to Executive Director Ann Cotton, Camfed’s performance measurement system, which uses community members (including the young women in its own program) to help collect data, is “significantly less expensive and more beneficial than flying in evaluators” to do infrequent assessments. The reason is that regular measurement not only helps the organization track academic and social outcomes, but also helps staff detect in real-time when a young woman is not receiving Camfed’s services (often the result of community interference, a situation that Camfed can generally quickly rectify). What’s more, community data collectors (in addition to building their own skill sets) are improving the accuracy of the organization’s measurement because they “speak the local language and are sensitive to the undercurrents and power dynamics.” According to Camfed Director of Impact Laurie Zivetz, the organization has lowered its measurement-related costs in recent years by equipping data collectors with hand-held computers and, in some cases, mobile phones. In communities where mobile phones are being used, information about payments that Camfed has made on behalf of its clients, which used to be collected quarterly, is now being collected weekly or even daily.
One Acre Fund [i] is a ~$3 million organization that provides a comprehensive bundle of seed, fertilizer, training, marketing, and insurance to more than 30,000 one-acre farmers in East Africa. At any one time, One Acre Fund has 20-30 experiments underway to explore new opportunities to increase the profit on each acre planted, including testing different fertilizer configurations, adjusting the ratios of training staff to farmers, and revising repayment schemes. The organization’s leaders regularly gather data about these experiments, analyzing and interpreting it to help them test, prototype, and quickly refine new approaches. As a result, One Acre Fund is able to implement the best ideas and improvements in time for the next crop season.
Improved resource allocation
Since 10,000 Women works with more than 70 academic and nonprofit partners, its leaders find that making decisions about where and how to allocate resources is a constant challenge. Which countries should it invest in? Which partnerships? Which programmatic activities? The initiative’s performance measurement system informs these decisions. For example, demographic and historical business growth data gathered from the first round of participants in 2009 showed that program partners were having difficulty finding women who were both underserved and had growth potential. Goldman Sachs used this information to reallocate resources, increasing its investment in local marketing and outreach activities, thereby enabling partners to more easily locate women who meet the program’s selection criteria.
Lower program costs
Boston-based Jumpstart provides individualized mentoring and tutoring to pre-school children in low-income communities. In 2002, senior leaders noticed that the organization’s Boston site had twice as many staff members as its other sites, yet the organization’s measurement system showed that Boston turned out similar numbers of children ready for school (Jumpstart’s primary measure of success). Armed with this ‘cost-per-outcome’ data, Jumpstart moved to bring staffing costs in Boston in line with other sites, while still maintaining the quality of its offerings, and the outcomes it achieved.[ii]
A Framework for Performance Measurement, and Lessons Learned
Many social sector leaders tell us that they find the concept of measurement—and the intricacies of developing a measurement system they can feel confident about—confusing. One way to cut through that confusion is to think about learning and improvement as a continuous lifecycle, and to think about measurement as a tool that can help an organization move through that lifecycle. Social sector organizations that use measurement in this way generally follow some variation of a process first popularized in the corporate world, which has been using proven improvement tools such as Six Sigma for decades. In the chart below[iii], we portray the lifecycle with four parts: Define, Measure, Learn, and Improve.
As an organization’s leaders get increasingly clear about the results they aspire to achieve, and about what they need to do to achieve those results (Define), they are better able to figure out what information will tell them how they’re doing (Measure), understand what works and what doesn’t (Learn), and explicitly apply what they’ve learned to better their results (Improve). Keeping this four-part lifecycle in mind not only helps leaders decide what to measure and why, but also helps them decide what not to measure, an important factor in keeping measurement from becoming overwhelming and ineffective.
In our experience, many nonprofit leaders have the hardest time with the first part of the process, defining the impact they intend to achieve and articulating exactly what needs to happen (and why) in order to achieve that impact. But each part presents its own challenges. The experiences of exemplars in the social sector offer the following five lessons that suggest how to get a productive performance measurement system up and running.
Lesson One: Begin with the end in mind[vi]
Before even thinking about metrics or systems, it’s important to get exceedingly clear about the results the organization will hold itself accountable for, also known as its intended impact.[v]
Want to start a discussion with your organization or grantee about how to use performance measurement to continuously improve? For each lesson, ask the corresponding questions:
- Begin with the end in mind. Can we articulate, in one sentence, who or what ends our organization serves and what change we seek to create, when?
- Anchor measurement in your theory of change. Can we tie each and every metric we collect to at least one decision that tests whether our theory of change is working?
- Create a culture of measurement. Does the leadership of our organization (including the board) use data to make decisions? Have we created the right forums for staff at all levels to wrestle with the data and use it to drive improvements, all in real-time?
- Ensure all contributors benefit. Have we identified each player that contributes to or uses the measurement system and how they gain value from it?
- Get better at measurement over time. Has our organization mapped out how our measurement approach will get better as we mature and evolve?
A good test of whether the leadership team has this clarity is to ask each member to answer the following question with a single sentence: Who or what ends does your organization serve and what change do you seek to create, when? Until an organization’s senior leaders have the same answer to that question, (agreeing about intended impact), there is no point to measuring anything. Many large, sophisticated organizations have found themselves “drowning in data[vi] precisely because they were not sufficiently thoughtful and disciplined at this essential first step, or because the leadership delegated “measurement” to subordinates without first providing clarity about intended results at the governing level.
Importantly, starting with the end in mind means something much more specific and rigorous than setting broad goals. Holding the organization accountable for certain outcomes means allocating the resources necessary to achieve those outcomes. If, in the eyes of the organization’s leaders, the outcome doesn’t merit that investment, then progress towards that outcome isn’t worth measuring.
Goldman Sachs was highly intentional about establishing the outcomes it intended to achieve through 10,000 Women when designing the initiative in 2008. As Dina Habib Powell, global head of corporate engagement, explained, “Economic research showed that investing in women would drive economic growth and subsequently lead to healthier, safer, better educated families and communities. But to have this ‘multiplier effect,’ we first had to be sure that we were creating growth. So our key measure of success was going to be that businesses were increasing revenue and creating jobs.” The initiative’s ‘end’ result, or intended impact, was ultimately defined as follows: “Over five years, 10,000 underserved women around the world will expand their businesses and create jobs.” Being this specific (about target population, time frame, and desired outcome) was a critical first step for Goldman in getting measurement right.
Lesson Two: Anchor measurement in your theory of change
With a clearly defined intended impact, the organization’s leaders should ask: How, exactly, does the organization achieve results? Answering this question means articulating the organization’s theory of change.[vii] A robust theory of change specifies the complete set of programs, activities, organizational capabilities, and relationships that are required to achieve the outcomes the organization will hold itself accountable for. Once this theory of change is explicit and agreed upon, it is much easier to decide what information to collect to reveal whether the organization is on track, and if not, what’s needed. Just as the frame of a house defines its functionality and strength, so does a sound theory of change provide the architecture for measurement systems that work.
As shown above, the 10,000 Women example illustrates how translating a theory of change into a logic model allows for metrics (indicators that can be tracked), tools (instruments used to gather data for each metric), and timing to be anchored very explicitly in what the organization is trying to achieve. Once the initiative’s leaders specified the tools, they then built an online data system to receive and analyze the data; hired local measurement staff at each partner university to collect, validate, and input the data; developed standardized reports (or dashboards) to display critical information for all users of the system; and established targets for key metrics.
As Marcele Carneiro Gama Viana, 10,000 Women project manager at the Brazilian university Fundação Dom Cabral (FDC), explained, with this measurement system in place, “We’ve been able to learn very quickly what is working and where our Scholars [participants] need more help. For example, we were very surprised to learn that 90 percent or more of the Scholars in our first cohort had not given formal feedback to their employees over the past 12 months. We knew we needed to revise our human resources module to ensure a focus on this critical issue. The measurement has helped us make quick adjustments like this so the next cohort can benefit right away.” Measuring along the full logic model, as 10,000 Women does, helped FDC learn about a change in business practice that was not occurring such that it could implement an improvement in the business knowledge it was imparting.
It’s an ideal scenario when anchoring measurement to a theory of change results in such a precisely defined and useful system. But this lesson can also be highly effective when it’s taken to heart in a very simple way. Consider how measurement unfolds at the community-based LAYC, which provides multiple services to a diverse group of more than 4,000 youth in Washington, DC and Maryland. LAYC’s Castillo told us about an annual process he goes through with each of his programs: “We gather line staff and managers together and ask them a few basic questions. What should your participants look like at the end of the program? What changes should you see along the way? What indicators would best tell you if you are seeing these changes? And finally, how will you collect this data? Sometimes the answers are the same as the prior year, sometimes different. Where needed, we modify the measurement strategy.”
A final note on anchoring measurement to your theory of change: If you’re not sure you’ve selected the right sorts of measures, and you’re worried about information overload, you can test the measures you’ve selected by seeing if they align with the kinds of decisions you make on a regular basis. The best performance measurement systems are designed to inform specific decisions; leaders avoid the temptation to collect lots of information just because it seems interesting. Ask, “What are the specific decisions I will need to make during the course of this program? What qualitative and quantitative information will I need to make those decisions?” Then make sure that the system provides that information in the time required. Correspondingly, ask what decision each metric will inform. If there’s no answer, drop it. Many performance measurement systems fall under their own weight, because staff collects lots of information that seemed useful, but informed nothing in particular.
Lesson Three: Create a culture of measurement
We agree with Brian—but how do you create a culture of measurement if you don't already have one? Our research suggests two critical factors: leadership commitment and intentional opportunities to learn and improve.
If a nonprofit’s CEO and board do not use data to make decisions, all attempts to instill a culture of measurement throughout the organization will languish. What’s more, the most successful measurement systems reflect a great deal of input from the CEO in their design. If measurement is to be embraced by an organization, its leadership must make measurement a top priority, then follow-through by providing staff with the time and resources and an open environment to share results, good or bad.
Successful organizations of all sizes also make measurement a structural necessity; their leadership teams create specific opportunities for the board and staff (at all levels) to learn from data, make decisions based on that data, and implement improvements. In other words, they bring to life the second half of the ‘define, measure, learn, improve’ cycle by explicitly creating forums that put useful information in front of the people who need to see it and create the space for that information to feed into decisions.
Consider the learning forums 10,000 Women has formalized, which include an online portal (to post documents and hold virtual discussions), quarterly learning calls, issue-specific taskforces, and leadership academies (annual in-person conferences). These learning forums provide an opportunity to reflect on, and interpret, information from the measurement system, resulting in recommended improvements to the program. Goldman Sachs then uses formal Quarterly Portfolio Reviews as an opportunity for the initiative’s leadership to decide on which recommendations to adopt, how to implement them, how they will be monitored over time, and whether they imply a change in the initiative’s intended impact or theory of change.
Lesson Four: Ensure that all contributors benefit
Most of us have experienced a “system solution” that was imposed upon us without our input. We have seen such systems fail, despite well-intentioned efforts by smart people. In the world of measurement, this problem of top-down design surfaces when data systems are designed by and for some stakeholders. The result is beneficiaries who fail to respond to surveys, staff who don’t fully input the data they are supposed to, grantees who provide incomplete information, and funders who collect but don’t read reports. Measurement only works when those who contribute to and use the system—as data providers, data collectors, analysts, overseers—directly benefit from the system.
LAYC ensures that its measurement system adds value to all. As one example of how beneficiaries gain, the organization prepares ‘progress reports’ from its measurement system so youth on probation can show their probation officers or judges they are 'flying straight.' To ensure its front-line staff benefit, the Learning and Evaluation team empowers the staff to design the reports they want from the system. As Castillo explained, “We tell our staff we want to do this for a single reason—to help them serve their clients better. We start small by showing them how the system can help them answer one important question they are facing. Then we have them present the results to their peers. Once they see how the system lets them do their jobs better, they actually demand more from us, rather than the other way around.”
Advice for smaller organizations
The measurement techniques and lessons described in this article very much apply even in organizations that may lack resources to hire a measurement expert or undertake expensive studies. Consider the following as guidelines:
Articulate your logic model. Blank templates are readily available, for free, on the web. Carve out two to three hours at your next staff meeting. Start with intended impact and theory of change, then get your logic model (inputs, outputs, and outcomes) down on paper.
Start small. Resist the urge to measure everything at once. Focus first on the one outcome your activities are most likely to achieve. Measure it before, during, and after your program. Also capture the amount of services each participant is getting, and perhaps the one or two most important characteristics of your participants that could affect how well they do on your program. Look at the data—do you see any relationships?
Drive to why. Organize a pizza dinner for your participants—try to get as representative a group as you can. Use the data to formulate some hypotheses about how you can improve your results—then test these hypotheses at the dinner. Try to leave with one or two tangible improvements to make.
Make the case. What would you be able to do with more resources? How would measurement improve, and more importantly, how would your impact improve? Chart out a vision and see if your board finds it compelling enough to shift resources to measurement. Then, see if any existing or potential funders might support your efforts.
To ensure that funders and senior leaders benefit, the Learning and Evaluation team worked with both parties to design dashboards that can automatically be populated from its measurement system and that enable data-driven decision making.
Camfed International designs its measurement systems to ensure stakeholders at all levels—the family, the community, the district, and the country—benefit from the data. In doing so, Camfed has empowered its clients and influenced policy that affects the hundreds of thousands of women and girls it does not directly assist. Camfed Zambia’s Executive Director Barbara Chilangwa (the country’s former permanent secretary for Education) explained: “When we enter a new community, we first work with the planning officer to understand what local policy changes they want to make and how our data can support their efforts. We then align our measures to ensure Camfed data is directly comparable to government data and even ask the officials to help us design our instruments and train our enumerators. Then, as data comes in, they see the difference Camfed makes and they are armed with the data to institute policy changes.” For example, Camfed often helps local officials implement pre-existing (but often ignored) policies against corporal punishment by showing how child protection leads to improved attendance and performance in Camfed schools. Had Camfed not built its data systems with government needs in mind, it would not have such systemic impact at the country and local levels.
Lesson Five: Get better at measurement over time
With experience, organizations find that they can identify with increasing confidence the particular aspects of their programs that drive results, and the measures that give them the most valuable information. As a result, they’re increasingly able to reduce the time and expense of measurement, for example by pruning the number of metrics they track, or surveys they conduct, or by adjusting sample sizes.
Take the Strive Initiative in Cincinnati, a coalition of 300-plus nonprofits, school districts, foundations, and corporate funders that work in a coordinated fashion to address education issues from birth to college graduation and a meaningful career. Through adoption of Six Sigma and more than two years of facilitated bi-weekly meetings, the initiative developed common goals, evidence-based strategies, common outcome measures, and overarching metrics of regional impact that allow the coalition to work in a more coordinated and effective way.[ix] Through continuous testing and refining, the leadership team narrowed the indicators in the organization’s report card from over 150 at the start to just 10 today, with plans to reduce even further to just seven in the coming year. As Executive Director Jeff Edmondson explained, “People originally just threw out whatever indicators came to mind, but over time we were able to have an honest dialogue about which few would really tell us if we were moving the needle.” Learning what matters most enables nonprofit leaders to narrow their metrics to a “vital few” and do more with less.
The most challenging aspect of getting better at measurement over time is improving its rigor—that is, figuring out how to get to a more sophisticated understanding of your results. For a multi-service organization, improving rigor might mean pulling outcome data from its measurement system and cost data from its financial system to look at ‘cost per outcome’ for each activity offered. For an afterschool program, increased rigor might mean automatically pulling test score data from a government website into its own measurement system to get a basic sense of how its participants’ outcomes have changed relative to a larger peer set. Increased rigor might also mean conducting more in-depth analyses (such as regressions) that can begin to identify the nuanced characteristics of beneficiaries that make them particularly good candidates for a given program.
For some organizations, increased rigor may include evaluation studies to assess program design, strength and consistency of implementation, or the attribution of outcomes to the program (impact evaluation). When these organizations design their performance measurement systems to give as accurate a read of their outcomes as possible, they improve the odds that they will pick the right time to make the (typically) large investment of time and resources in an impact evaluation. Organizations at this stage often consult outside experts to assess feasibility and desirability of such evaluations.
The accompanying graphic, “How performance measurement should evolve over time[x],” shows the measurement activities an organization can add as it masters a given stage. The graphic hopefully provides a healthy rubric for CEOs, board members, and philanthropists to think through what they should be expecting of their measurement system as a program evolves. Becoming a ‘black belt’ isn’t easy: best-in-class organizations we have studied generally take a decade or longer to move all the way to the right.
How Can Donors Help?
Donors face a difficult dilemma regarding measurement. In holding themselves accountable for ensuring that their funds are making a difference, they understandably require reports from grantees. But what does it cost the grantee to prepare those reports? And what does that preparation do for the grantee, and the funder? It’s not unusual for nonprofits to invest weeks of staff time each quarter to prepare these reports. What’s heartbreaking is that most have little impact, on either the funder or the nonprofit. As Gail McClure, former vice president at the W.K. Kellogg Foundation recently commented, many funders “can stack evaluation reports to the ceiling and fill barns with them, but take very little action as a result.”
The good news is that funders that find themselves in this situation can move away from it relatively quickly. In fact, donors can play a critical and collaborative role in enabling nonprofit organizations to build the capability for performance measurement.
To begin, they can become thought partners. They can encourage grantees to get clear about their intended impact and to develop a rigorous theory of change. Asking pointed questions against each of the five ‘lessons learned’ (see earlier sidebar) will encourage a nonprofit’s leaders to turn the lens on their own organization in a productive way. Ultimately, the same metrics that help nonprofits get better over time are those that best inform the donor.
Building on that idea, donors can provide grantees with the resources they need to develop an internal measurement capacity. Performance measurement is most valuable when done internally by staff that experience firsthand the power of using data to make decisions. While impact evaluations are critical to fund for more established programs looking to replicate, funders should not reserve measurement-related funding until the time when an organization is ready for such an evaluation. And since measurement is often considered an overhead cost (and thus notoriously difficult to fund), donors who care should match their expectations with resource commitments.
Where possible, funders might also facilitate ‘shared measurement systems’ where grantees delivering similar activities and/or working with similar beneficiaries agree to share metrics, data systems, or both. Nonprofits may be surprised to learn how cost effective they are to operate and how valuable it can be to see data from other organizations. Strive Cincinnati’s Edmondson, for example, believes the cumulative costs of measurement—for the funders and service providers—are coming down considerably as they adopt the single platform. As he explained, “A complicated regression analysis that we needed 45 hours of paid consulting time to complete can now run with a press of the button.” More importantly, the rich information generated by combining academic, social service, and youth development data on a single child provides huge insights that drive better services. As Edmondson put it, “When you have the health provider, afterschool provider, and teacher all talking, kids learn much more effectively.[xi]
Ideally, as funders work more closely with grantees, they’ll create their own learning communities—trusting environments that encourage nonprofit leadership teams to share information with each other, benefit from practices that work well, and learn from failures. Over time, we hope to see funders holding nonprofits accountable for embracing a culture of learning and improvement, even if that means nonprofits don’t always deliver perfect results in their quarterly progress reports to funders.
Clients, Front and Center
The most effective performance measurement systems are laser-focused on beneficiaries. They tell you whether beneficiaries are seeing changes from your program, why or why not, and what you can do about it. They treat beneficiaries as clients, sharing back data to help clients bring about their own change, rather than extracting data to manage the change process internally. Best of all, these systems can be built and managed by any social sector organization, at any stage of development, using its own staff. As Ann Cotton, founder of Camfed International, summed up, “How can we improve our services and be accountable to our clients if we don’t measure?”
The authors gratefully thank Fay Twersky, former Director of Impact Planning and Improvement, Bill and Melinda Gates Foundation; Ann Cotton, Executive Director, Camfed International; Brian Trelstad, Chief Investment Officer, Acumen Fund; Jeff Edmondson, Executive Director, Strive; Isaac Castillo, Director of Learning & Evaluation, Latin American Youth Center; and Jackie Kaye, Director of Research and Evaluation, Wellspring Advisor for their excellent advice and thought partnership on the development of this article. We also would like to thank Regina Maruca for her assistance during the editorial process.
Sources Used for This Article
[i] One Acre Fund is a winner of the Skoll Award for Social Entrepreneurship, Draper Richards Fellowship and Financial Times Sustainable Banking Award. Matt Forti, co-author, is Founding Chair of the Board.
[ii] "More Bang for the Buck." Stanford Social Innovation Review, Spring 2008.
[iii] The focus of this article (in its examples) is on programmatic performance measurement; that is, performance measurement to improve the results for an organization’s target population. That said, the same lifecycle of continuous improvement would equally apply to performance measurement of people, finances, or other aspects of the organization.
[iv] Concept introduced by Stephen R. Covey in The 7 Habits of Highly Effective People. Free Press, September 15, 1990.
[vi] Concept originally introduced by Alana Conner Snibbe in an article of the same title in Stanford Social Innovation Review, Fall 2006.
[vii] For more about theory of change, please see The Bridgespan Group’s "Zeroing in on Impact," "Delivering on the Promise of Nonprofits," and Paul Brest’s "The Power of Theories of Change." Note that our focus (and examples used) in this article is on programmatic theory of change, but theory of change should also encompass the funding, organization, and partnership strategy to achieve those ends.
[viii] "Simple Measures for Social Enterprise." Innovations Magazine, Summer 2008.
[ix] "Breakthroughs in Shared Measurement and Social Impact." FSG Social Impact Advisors, July 2009.
[x] The graphic shown does not include how an organization might evolve its evaluation activities over time, which is beyond the scope of this article.
[xi] "Catalytic Philanthropy." Mark R. Kramer, Stanford Social Innovation Review, Fall 2009.