Voices from the Field: 10 Places Where Collective Impact Gets It Wrong

This article is reprinted with permission. Its original form can be found at the Global Journal of Community Psychology Practice.

In 2011, Kania and Kramer published a five-page article in the Stanford Social Innovation Review entitled “Collective Impact” (2011). The article was a well-written summary of their views of large scale social change efforts in communities. They suggested five conditions of collective impact:

1. common agenda
2. shared measurement
3. mutually reinforcing activities
4. continuous communication
5. backbone support

In the original article, and those that followed, Kania and Kramer were explicitly and implicitly critical of much of what came before them. In one chart (Hanleybrown, Kania, & Kramer 2012), they compare Isolated Impact with Collective Impact as if those were the only two options, omitting the numerous examples of community-wide coalitions that moved beyond Isolated Impact but were not explicitly labeled Collective Impact (for one example see the exhaustive literature on Healthy Communities, Norris, 2013).

As a result of that short publication and extensive marketing by Kania and Kramer’s consulting firm FSG, and a few follow-up articles (Hanleybrown, Kania, & Kramer, 2012; Kania and Kramer, 2013) they created a remarkable revolution in government and foundation approaches to community coalition building and collaboration. Many of these funding organizations are now declaring that they are using a “Collective Impact” approach. The upside of this is that attention is once again brought to the need to promote multi-sector collaboration in communities.

The downside of this is that Collective Impact is based on only a few case studies that the authors themselves were not involved in creating and implementing but rather observed after their development. The articles included neither research nor reference to learning from all the previous research, studies, and community experiences in the field. Observing successful coalitions provides the observer one basis for learning about community coalitions, but being involved in successfully and unsuccessfully developing coalitions provides a deeper and more nuanced understanding of coalitions that apparently was not available for Kania and Kramer. Thus, not surprisingly, Collective Impact gets much about collaboration wrong, regarding both the goals and processes of community change collaboration. In light of the uncritical, widespread adoption and funding of Collective Impact by government agencies and foundations, it is necessary to examine and assess Collective Impact much more critically and thoughtfully.

In this article, I articulate ten important issues and concerns that Collective Impact fails to adequately acknowledge, understand, and address. These failings have serious consequences for the engaged communities. I welcome the community of activists and scholars who are engaged in coalitions, partnerships, and collaboratives to react, disagree and/or to add to the list of concerns.

1) Collective Impact does not address the essential requirement for meaningfully engaging those in the community most affected by the issues.

Collective Impact does not set a priority of engaging those most affected by the issue in their collaborative impact processes. The grassroots communities most affected are not necessarily consulted or meaningfully share in Collective Impact decision-making. The result is to ignore and denigrate critical community knowledge, ownership, and support for sustainability. This can further result in creating solutions that may not be appropriate or compatible with the population being served. This is not surprising because Kania and Kramer come from a top-down business consulting model. Collective Impact never explicitly states that you need to engage the people most affected by the issue(s) driving the coalition. Unfortunately, Collective Impact’s approach is not unusual; in general, collaboration processes used by coalitions of all kinds do not meaningfully involve grassroots community members or other stakeholders directly affected by their work (Himmelman, 2001). This is a serious omission. Coalitions without grassroots voices are very likely to create solutions that do not meet the needs of the people most affected by them and treat people disrespectfully in their community change process.

Without engaging those most directly affected, Collective Impact can develop neither an adequate understanding of the root causes of the issues nor an appropriate vision for a transformed community. Instead the process will likely reinforce the dominance of those with privilege and continue to support the existing nonprofit helping sector that works without creating changes based on meaningful community input and involvement.

2) A corollary of the above is that Collective Impact emerges from a top-down business consulting experience and is thus not a community development model.

The model of Collective Impact is mainly about engaging the most powerful organizations and partners in a community and getting them to agree on a common agenda. They explicitly state that Collective Impact is about bringing “CEO-level cross-sector leaders together” (Hanleybrown 2012). In reality, what community coalitions need is to engage both the most powerful and least powerful people in a community, finding ways for them to talk and work together to address the community’s priorities for action and the impediments to change in institutions and organizations serving the community. This is the heart and soul of community development coalition work and seems absent in Collective Impact.

Coalitions across the country have years of experience in bringing a wide range of the community stakeholders to the table, not just the most powerful. Often this was not the case. Early in the history of substance abuse prevention work, partnerships made the top down mistake. At the start (in 1989), the Robert Wood Johnson Foundation’s Fighting Back substance abuse prevention coalitions required having the most powerful people in the community at the table—the mayor, the police chief and the school superintendent (Wielawski, 2004). As the role and effects of the community environment surrounding substance abuse issue became clearer, we began to see that we needed all sectors of the community and the youth themselves at the table. At that point the coalitions began to evolve and become more effective. Unfortunately Collective Impact seems stuck in this old, less effective model with CEO leadership as central to their process.

3) Collective Impact does not include policy change and systems change as essential and intentional outcomes of the partnership’s work. 

Many coalitions in the U.S. are focused on creating public health outcomes (prevention of substance abuse, obesity, opioid addiction, health disparities, etc.). In recent years, led by the CDC (Frieden, 2010), these coalitions have moved in the direction of policy and systems changes as their most powerful and desired outcomes. Certainly in public health coalitions (which comprise many of the coalitions in the U.S.), following the CDC’s lead and addressing policy change and systems change has become the gold standard of outcomes. Systems change is now recognized as a key priority and best practice in community change partnerships so this is a serious omission in Collective Impact.

If we are not changing policies in order to change systems, we are continuing to do fragmented, isolated work. For years, community coalitions addressed specific focused issues without asking about the ecological and historical factors that impact the outcomes. Smoking cessation coalitions taught us all this lesson dramatically as they went beyond smoking prevention education for young people to a focus on implementing anti-smoking policies in systems across the community—restaurants, schools, work sites, public buildings. And it worked! So now we better understand that policies are at the heart of the work of community coalitions. But where is the policy and system change in Collective Impact?

4) Collective Impact misses the social justice core that exists in many coalitions.

Increasingly coalitions are applying “root cause” analyses to understanding their community issues. As they do this and understand the concept and ramifications of social determinants of health, critical social justice issues become clear and urgent, such as: income inequality, systemic and structural racism, sexism, homophobia, etc. Collaborative efforts then must mobilize to address these issues which can be difficult to do in top-down collaboratives; those with the most power and privilege dominate and control top-down coalitions and often have interest in maintaining their privilege and the status quo. Collective Impact is a great tool for those who already have power, but it is less suitable and more challenging for those with relatively little power who are working to improve the lives of people and their communities.

For example, alternative partnership models such as the REACH (Racial and Ethnic Approaches to Community Health) coalitions funded by the CDC are aimed to address systemic racism and create systems level change. The Boston REACH coalitions that emerged from the Boston Public Health Commission (Baril, et. al., 2011) were all required to do root cause analyses of their community’s issues. This led to understanding the racial health disparities in their communities in the context of social determinants of health (housing, economic inequality, education, etc) and the institutional racism that is part of each of these determinants and their related systems. With this approach, addressing structural racism became not just a possibility but a necessity.

5) Collective Impact, as described in Kania and Kramer’s initial article, is not based on professional and practitioner literature or the experience of the thousands of coalitions that preceded their 2011 article. 

When dealing with an issue as complex as collective actions taken by the multiple sectors of a community, we need to be continually learning from those who came before us and from the communities themselves. When I first began working with coalitions almost 40 years ago, even then I found valuable resources from a wide range of fields, including community psychology, civic engagement, racial justice, public health, political science, and organizational development among others. Since then, the literature, experience, and tools for coalition building have grown exponentially and are utilized extensively by coalitions in a wide variety of circumstances.

Here are a small sample of comprehensive community-wide collaboration resources that are not cited or maybe even known by Kania and Kramer:

• Among the most acclaimed and utilized is Fran Butterfosses’ comprehensive, Coalitions and Partnerships in Community Health (2007) that articulates her and Michelle Kegler’s Community Coalition Action Theory (Butterfoss and Kegler, 2002).
• Others’ significant scholarly writing about partnerships in Public Health include:
  • Michelle Kegler(1998)
  • Meredith Minkler (2012)
  • Nina Wallerstein(2005)
• In Community Psychology, this work has a long history in the work of:
  • Seymour Sarason (1979)
  • David Chavis(1992, 2001)
  • Steve Fawcett (2000)
  • Bill Berkowitz (2000)
  • Pennie Foster Fishman (2010, 2011)
  • Vince Francisco (2000)
  • And my own writings (Wolff 1995, 2001, 2003, 2010)
• There is an extensive literature and experience in the field of Healthy Communities, including two recent volumes of the National Civic Review focused on the topic (Norris, 2013) and important writing about healthy communities by others such as Twiss (2000), Kurland (2001).
• There are also extensive related contributions from other fields:
  • Political Science: Himmelman (2001)
  • Collaborative Leadership: Chrislip and Larsen (1994)
  • Community Development: Potapchuk (1999)
  • Community Organizing: Kaye (1996)

This rich multi-disciplinary literature teaches us that the process of communities working together to create collaborative change is very complex and is impacted by multiple variables. The literature also identifies processes, methods, and models that have led to the creation of successful collaborations that create changes in programs, practices, and policies in communities. Collectively, we already know a great deal about the tools necessary to do this work. One of the most comprehensive and internationally acclaimed examples is the Community Tool Box. The Community Tool Box provides over 7000 pages of free downloadable material on community health and development using collaborations and partnerships (Fawcett, et. al., 2000).

Collective Impact flounders by failing to learn from all these wonderful contributions in the literature and the field from all the above disciplines. How can Collective Impact propose converting a whole field with a five-page article that has virtually no references to the concepts and findings of others? And how can government agencies and foundations uncritically adopt such a model that mislabels observations about a few examples of community collaboration as valid research?

6) Collective Impact mislabels their study of a few case examples as “research.”

The Stanford article cites a few successful examples of community coalitions and draws their Collective Impact generalizations from them. This is a very limited sample and it seems that Kania and Kramer only observed these coalitions and drew conclusions rather than having actually been involved in the messy work of creating coalitions like the ones they note. It is actually stunning to realize that Kania and Kramer changed the world of coalition building simply by observing and distilling insights from a few successful coalitions, but never actually tried creating, implementing, and evaluating a coalition themselves.

In my own work with hundreds of coalitions, I have learned that there is much to be learned from the biggest best funded top down coalitions that succeed and those that fail, as well as from the smallest that also succeed and fail. I understand we draw our generalizations from the coalitions that we work with, and I have always done so myself. However, seeing that Collective Impact has become the gold standard of coalition building for government and foundations on such a limited sample and such limited actual experience is deeply disconcerting. It is fascinating to note that many government agencies (federal, state, local) and foundations are now calling for all of us to follow Collective Impact as the model if we wish to be effective and funded. Yet this is an intervention with absolutely no evidence-based research. Aren’t these the same government and foundation organizations that demand evidence-based research from us in all their program applications?

One has to wonder what makes funders so attracted to Collective Impact. Could it be that the simple five Collective Impact components allows funders to believe that coalition building can be simplified, and that they finally have the key to success for these messy multi-variable entities called coalitions? Or could it be that Collective Impact’s top down approach is most compatible with the collaboration change approach of foundations? Or could it be Collective Impact’s avoidance of addressing policy or advocacy which makes Collective Impact coalitions a safer and less controversial funding bet?

7) Collective Impact assumes that most coalitions are capable of finding the funds to have a well-funded backbone organization.

Kania and Kramer’s call for coalitions to have a Backbone Organization is welcome. Finding money for the staffing of coalitions has always been very difficult. Most funders want to fund the coalition’s change mission, goals, and programs, but very few grant-makers want to fund coalition staffing and operating costs. It is great to see an emphasis for the requirement of support for these essential core elements of coalitions.

Unfortunately, here again Collective Impact gets it wrong by asking for too much from the backbone organization. Collective Impact experts push for a well- funded backbone organization with multiple functions that require considerable resources and staff. These functions include, “providing overall strategic direction, facilitating dialogue between partners, managing data collection and analysis, handling communications, coordinating community outreach and mobilizing funding” (Hanleybrown, 2012). By giving all those responsibilities to the backbone organization, Collective Impact inevitably creates a top-down organization versus a truly collaborative one where leadership and responsibility is dispersed. The Collective Impact concept of a backbone organization is predicated on coalitions with extensive resources. However, in the hundreds of coalitions I have created, consulted with, or trained, very few can even afford paid leadership much less a $100,000 backbone organization.

8) Collective Impact also misses a key role of the backbone organization—building leadership. 

In well-run coalitions, the key role for the backbone organization needs to be to build coalition leadership as opposed to being the coalition leadership. This is based on the shared value of instituting collaborative leadership as well as democratic governance and decision-making for a coalition.

Collective Impact barely discusses the idea that leadership in a collaboration is different from ordinary organizational leadership. Again, there is excellent literature that provides a guide to democratic and collaborative governance. Almost twenty years before Collective Impact, David Chrislip and Carl Larsen’s Collaborative Leadership (1994) helped distinguish the unique characteristics and practices of collaborative leadership in coalitions, including the skills and functions of a collaborative leader and how they differ from traditional hierarchical leadership.

Coalition leaders themselves often emerge from traditional top-down non-profit organizations and need to learn a new style of leadership that facilitates ownership and leadership by the members. We have seen powerful charismatic coalition leaders who can energize a coalition but then fail when they cannot organize the energy that they stir up and delegate the responsibility.

9) Community-wide, multi-sectoral collaboratives cannot be simplified into Collective Impact’s five required conditions.

Coalitions are complex, constantly changing, and influenced by multiple variables. Having worked with numerous coalitions, I cannot imagine any five conditions that could apply universally. In writing The Power of Collaborative Solutions in 2010, I identified six principles and effective tools for consideration rather than prescriptive conditions:

1. Engage a broad spectrum of the community
2. Encourage true collaboration as the form of exchange
3. Practice democracy
4. Employ an ecological approach that emphasizes the individual in his/her setting
5. Take action
6. Engage your spirituality as your compass for social change

For example, the first condition of Collective Impact is creating a common agenda, and this is highly desirable and necessary. When we assist community coalitions through visioning exercises, including root cause analysis, and provide guidance that helps members develop a shared common agenda, it is an important accomplishment. However, we need to acknowledge that in some communities the conflicting self-interests can be insurmountable and the common agenda is either not achievable or can require a long time to arrive at. Collective Impact can frustrate those led to believe that complex activities such as developing a common agenda (often called a mission statement) can be achieved simply and quickly. The difficulties in this kind of collaborative decision-making can be even more frustrating when Collective Impact does not supply the community stakeholders with the tools that we know work.

10) The early available research on Collective Impact is calling into question the contribution that it is making to coalition effectiveness. 

The Collective Impact Model and Its Potential for Health Promotion (Flood, et. al., 2015) is among the first published scholarly assessments of the strengths and weaknesses of the Collective Impact approach. They note the lack of resident involvement and the absence of policy and advocacy in the Collective Impact model, suggesting that: “Since many community coalitions are deeply concerned with advocacy and policy change this omission can be problematic.” The study indicates that seeking a common agenda “will not be successful if done through coercive compromise” and without a backbone organization that has a “point of view” and a “broader mission, vision and values” (Bell in Flood et. al., 2015). The study also notes that the Collective Impact model does not provide detailed advice (and tools) to help coalitions create the needed continuous communication or common agendas. In its conclusion, the study states, “As our case study application suggests, Collective Impact appears to have utility as a conceptual framework in health promotion but one that may be usefully be augmented by some ‘tried and true’ insights and strategies from CCAT (Community Coalition Action Theory; Butterfoss and Kegler 2009)” (Flood, et. al., 2015). Additional thoughtful and insightful Collective Impact critiques are emerging in blogs and online from Mark Holmgren (2015), Vu Le (2015), and others.

I would concur with the view that there are some helpful contributions in the writings of Kania and Kramer. They bring fresh eyes to the work of collaboration. They have certainly brought coalition building back to the forefront for grant-makers and many others with influence in the government and foundation/non-profit sectors. Now, we have to make sure that Collective Impact does not proceed without addressing the ten points noted above. Let’s work to improve Collective Impact so it can take its place alongside many other valuable models and resources designed to assist people and communities improve their well-being by engaging the grassroots communities themselves and creating a vision of transformative change.

In sum, I am hopeful that, if communities using Collective Impact and funders promoting it address the ten shortcomings discussed in this article, we will see improved applications of Collective Impact emerge:

• where those most affected by the issues lead the effort and share the decision making and the power;
• where the collaborative action is based on an understanding of the social, political, and social justice context in which the issues of the community are embedded, and addresses these issues head on; and
• where the Collective Impact work is more thoroughly based on the existing fields of coalition building and community development, learning from the acquired knowledge, experience, and available tools.

Let’s hope that we can muster the courage to challenge the Collective Impact juggernaut and bring our communities what they need and deserve. I know we have the desire to do this and now we need the will.

Summary Table: Ten Places Where Collective Impact Gets It Wrong

References

Baril, N., Patterson, M., Boen, C., Gowler, R., & Norman, N. (2011). The Grantmaking Model of a Local Health Department. Family and Community Health, 34 (18), 523-543.

Bell, J. Quoted in Flood,J, Minkler, M, Lavery,S., Estrada,J, & Falbe J (2015) in “The Collective Impact Model and Its Potential for Health Promotion: Overview and Case Study of a Healthy Retail Initiative in San Francisco ” Health Education and Behavior 42(5), 654-68.

Berkowitz, W. and Wolff,T. (2000). The Spirit of the Coalition, Washington D.C., American Public Health Association.

Butterfoss, F. (2007) Coalitions and Partnerships in Community Health, San Francisco, CA: Jossey Bass.

Butterfoss, F. and Kegler, M. (2002). Toward a comprehensive understanding of community coalitions: moving from practice to theory. In DiClementi, R.; Crosby, L.; & Kegler, M. (Eds) Emerging Theories in Health Promotion Practice and Research. San Francisco, CA: Jossey-Bass. 157-193

Chavis. D. (2001). Paradoxes and promise of community coalitions. American Journal of Community Psychology 29(2), 309-320.

Chavis ,D. Florin, P., & Felix, M.(1992). Nurturing grassroots initiatives for community development. The role of enabling systems. In T. Mizrahi and J. Morrison Community and social administration. Advances, trends, and emerging principles. Binghamton, NY: Haworth Press.

Chrislip, D., & Larsen, C. E. (1994). Collaborative Leadership: How Citizens and Civic Leaders Can Make a Difference. San Francisco, CA: Jossey-Bass.

Fawcett, S., Francisco, V., Schultz, J., Berkowitz, B., Wolff, T., & Nagy, G. (2000). The Community Tool Box: A web based resource for building healthy communities. Public Health Reports, 115 (2/3), 274–278.

Flood. J, Minkler. M., Lavery, S., Estrada, J., & Falbe, J. (2015). The Collective Impact Model and its potential for health promotion: Overview and case study of a healthy retail initiative in San Francisco” Health Education and Behavior 42(5):654-68.

Foster-Fishman, P.G., & Watson, E.R. (2010). Action research as systems change. In H.E. Fitzgerald, D.L. Zimmerman, C. Burack, & S. Seifer(eds.) Handbook of Engaged Scholarship: The Contemporary Landscape. Volume Two: Community-Campus Partnerships. East Lansing, MI: Michigan State University Press.

Foster-Fishman, P. & Watson, E. (2011). The ABLe Change Framework: A conceptual and methodological tool for promoting system change. American Journal of Community Psychology 49(3-4), 503-516.

Francisco, V, Fawcett, S. Schultz, J & Paine-Andrews .A. (2000), A model of health promotion and community development” In F. B. Balcazar, M Montero and J.R. Newbrough (eds) Health Promotion in the Americas: Theory and Practice, Washington DC.: Pan American Health Organization, pp. 17-34.

Frieden, T. (2010). A framework for public health action: The Health Impact Pyramid. American Journal of Public Health. April; 100(4): 590–595.

Hanleybrown, F., Kania, J., & Kramer, M. (2012). Channeling Change: Making Collective Impact Work, Stanford Social Innovation Review.

Himmelman, A.T. (2001). On coalitions and the transformation of power relations: collaborative betterment and collaborative empowerment. American Journal of Community Psychology 29(2), pp.277—285.

Holmgren, M. (2015). Part One: Community at the core of a theory of change. Collective Impact: Watch Out for the Pendulum Swing and Other Challenges.

Kania, J. and Kramer, M. (2011). Collective Impact. Stanford Social Innovation Review.

Kania, J. and Kramer, M. (2013). Embracing emergence: How Collective Impact addresses complexity. Stanford Social Innovation Review.

Kaye, G. and Wolff, T. (1996) From the Ground Up! A Workbook on Coalition Building and Community Development. Amherst, MA: AHEC/Community Partners., 1996

Kegler, M., Steckler, A., McElroy. K., and Malek, S. (1998). Factors that contribute to effective community health promotion coalitions: A study of 10 Project ASSIST coalitions in North Carolina. Health Education and Behavior 25(3), pp. 338-353.

Kurland, J. (2001). Coalition-Building: The promise of government. American Journal of Community Psychology 29(2) 285-291

Le, V. (n.d.) “Collective Impact”: Resistance is futile. Blue Avocado.

Minkler, M. and Wallerstein, N. (2005). Improving health through community organizing and community building: A health education perspective. in M. Minkler(Ed) Community Organizing and Community Building for Health and Welfare, New Brunswick, NJ: Rutgers University Press.

Minkler. M. (Ed.) (2012). Community Organizing and Community Building for Health and Welfare, New Brunswick, NJ: Rutgers University Press.

Norris, T. (2013). Healthy communities at twenty-five. National Civic Review, 102(4), pp. 4-9.

Potapchuk, W. (1999). Building an infrastructure of community collaboration. National Civic Review 88(3), pp. 165 – 170.

Sarason, S. (1979). The Challenge of the Resource Exchange Network. San Francisco CA.: Jossey Bass

Turner, S., Merchant, K., Kania, J., & Martin, E. (2012). Understanding the value of backbone organizations in Collective Impact: Part 1. Stanford Social Innovation Review.

Twiss, J. M., Duma, S., Look, V., Shaffer, G. S., & Watkins, A. C. (2000). Twelve years and counting: California’s experience with a statewide Healthy Cities and Community program. Public health reports, 115(2-3), 125-133.

Wielawski, I. M. (2004). The fighting back program. The Robert Wood Johnson Foundation Anthology: To Improve Health and Health Care, 7.

Wolff, T. (2001). Community coalition building—contemporary practice and research: introduction. American journal of community psychology, 29(2), 165-172.

Wolff, T. (1995). Healthy Communities Massachusetts: one vision of civic democracy. Municipal Advocate, 14(2), 22-24.

Wolff, T. (2003). The healthy communities movement: A time for transformation. National Civic Review, 92(2), 95-111.

Wolff, T. (2010). The Power of Collaborative Solutions: Six Principles and Effective Tools for Building Healthy Communities. San Francisco CA: Jossey Bass.