Measuring Impact in the Nonprofit Sector: What Works and What Misleads

The push for nonprofit impact measurement has transformed the sector over the past two decades. Funders require it. Watchdog organizations rate nonprofits on it. Sector media debate its methodologies. The underlying motivation is reasonable: with limited philanthropic resources, donors and funders want to know whether the organizations they support are actually making a difference. But the measurement landscape has produced a complex mix of genuinely useful information and misleading metrics, with significant consequences for how nonprofits operate and what they prioritize. The overhead ratio, which measures what share of a nonprofit's budget goes to administrative and fundraising expenses rather than program delivery, became the most visible impact metric in the sector largely because it is easy to calculate from publicly available Form 990 data. Charity watchdog organizations long used overhead ratios as primary indicators of organizational quality. The problem is that overhead ratio is a poor proxy for effectiveness. Low overhead can indicate a lean organization that directs resources efficiently to programs. It can also indicate an underinvested organization that lacks the infrastructure to deliver programs effectively, evaluate their outcomes, or develop new staff capacity. The overhead myth, a term coined by GuideStar, Better Business Bureau Wise Giving Alliance, and Charity Navigator in a joint 2013 letter to the nonprofit sector, acknowledged this problem explicitly. The organizations that had promoted overhead as the primary metric of nonprofit health publicly disavowed the approach and called for more sophisticated impact measurement. Progress in moving the sector's measurement culture has been real but uneven. Individual donors in particular still frequently use overhead as a heuristic, partly because more sophisticated alternatives are harder to understand and access. Output metrics, which count the quantity of activities or services delivered, are often more available than outcome metrics, which measure changes in the conditions of people served. Counting meals served, people trained, or youth enrolled in programs is relatively straightforward. Determining whether meals served improved nutrition and health, whether training led to employment, or whether youth enrollment led to educational gains requires more sophisticated data collection, comparison groups, and evaluation design. Theory of change frameworks attempt to map the logical connections between program activities, their immediate outputs, intermediate outcomes, and long-term impact. A well-developed theory of change makes explicit the assumptions embedded in a program's design, which enables both better evaluation design and clearer organizational communication about what the program is trying to achieve and why. Theories of change have become standard in the sector, though quality varies enormously from sophisticated logical analysis to retroactively written narratives that do not actually guide measurement. Randomized controlled trials are often discussed as the gold standard of nonprofit impact evaluation, but they are expensive, methodologically demanding, and inappropriate for many types of programs. They require sufficient sample sizes, randomizable populations, and counterfactual conditions that are not always present. Organizations delivering universal services or working in geographies where a control group cannot ethically be denied services face particular challenges. Developmental evaluation approaches, which emphasize learning and adaptation rather than summative judgment, may be more appropriate for innovative programs still in early development. Social return on investment attempts to translate program outcomes into monetary values that can be compared to program costs, producing a ratio of social value created per dollar invested. The methodology has intuitive appeal but significant methodological challenges, including the difficulty of assigning monetary values to social outcomes, the problem of attribution in complex social systems, and sensitivity to assumptions that can dramatically change the calculated ratio. SROI analyses with weak methodological foundations have contributed to skepticism about the approach. Participatory evaluation approaches, which involve program participants in defining success measures and evaluating outcomes, challenge the assumption that external funders and evaluators should define what counts as impact. These approaches recognize that community members often have clearer perspectives on what meaningful change looks like in their lives than outside evaluators, and that evaluation processes themselves can build community capacity when designed with participation in mind. The risk of measurement gaming deserves frank acknowledgment. When organizational survival depends on demonstrating particular metrics, organizations face pressure to define and measure those metrics in ways that show favorable results. Cherry-picking participants, adjusting outcome definitions, gaming data collection, and emphasizing successes while minimizing failures are all responses to evaluation pressure that undermine the actual purpose of measurement. Creating evaluation systems that reward honest assessment and learning from failure, rather than only rewarding positive results, is a design challenge that funders and the field have not fully solved. Good impact measurement starts from genuine organizational questions about what is working and why, not from funder compliance requirements. Organizations that measure because they genuinely want to learn and improve what they do produce more honest, more useful, and ultimately more valuable information than those that measure primarily to satisfy external accountability requirements. Building internal cultures of learning, reflection, and honest assessment is the foundation on which meaningful measurement rests.