As I’ve observed in past columns, innovation may be one of the most overused buzzwords in modern discourse. That doesn’t mean that large and small organizations in every sector don’t need to find ways to achieve the real thing. Beyond the hype, after all, is a simple goal: improve processes, products, services, and outcomes through a change in how something is manufactured, delivered, planned, or implemented.

Small tweaks in large organizations, from corporations to colleges to government agencies, can produce significant, even disruptive, changes downstream when scaled throughout an organization. Rapid experiments in startups, small businesses, and “skunk works” can lead to iterative changes that get products or services that customers and citizens want to market quickly. The promise of such approaches and hunger for genuine innovation has led generations of people to try new methods to making what they do better, along with generations of risk-averse superiors to protect the status quo, for fear of increased risk, embarrassment, or simply trying something new when an existing approach works.

In a new report from the IBM Center for The Business of Government, Rachel Burstein and Alissa Black explore a recent trend in the US: the creation of “innovation offices,” entities within government specifically tasked with catalyzing and enabling improvements. Many such centers have been set up in cities, states, and federal agencies, with varying results, for reasons the authors explore.

As I read the report on my recent travels, I particularly enjoyed Burstein and Black’s reflection on the historical dynamics that have led the government to experiment with and then adopt innovation centers, albeit on a different kind of scale:

“With pressures from the existence of more nimble startups and global competition in the late 20th and early 21st century, many large corporations rethought their R&D groups, sometimes spinning them off and sometimes differentiating between product development and systems work. In some cases, corporations established chief innovation officer posts, not so much as a replacement for R&D groups, but to signal a commitment to new product and systems development that was more agile, timely, and integrated than R&D groups sometimes were. At the same time, the World War II and Cold War eras’ huge investments in new technologies and coordination between public and private sectors were largely things of the past.

It is into this orbit that government innovation offices have emerged, particularly in the last five years. Some are designed as R&D groups akin to those within business, but the vast majority do not have nearly the same levels of monetary investment as corporate entities of the past. Partly this has to do with changes in the content of research, as the shift to research on computing and service-related technologies over large-scale machinery and industrial projects has reduced costs. But it also reflects political and economic realities that are much more acute in government agencies than in corporations. Transparency rules and taxpayer concern about government spending sometimes complicate large scale R&D programs housed wholly within government entities or government research with external partners. In addition, at the state and local levels, very few resources or models exist for R&D.

At the same time, many recognize the need to develop new solutions to persistent problems, transforming the way that government operates and serves the public. Whether they have innovation offices or not, many government leaders have adopted the language of innovation–including concepts of disruption, open innovation, user-centered design, and the lean startup–from business as a way to address these challenges.”

Burstein and Black emerged from their research critical of some approaches but optimistic overall. “After dozens of conversations with practitioners, we believe that government innovation offices and chief innovation officer posts have the potential to be more than just the hype surrounding them,” they wrote. “In many cases, these offices are doing extraordinary work and are staffed by visionary leaders. To thrive long term, though, government innovation offices must be structured, staffed, and resourced appropriately and thoughtfully, with careful attention to meeting critical needs and solving big challenges.”

The relatively short report (a 36-page PDF, not including foreword and appendices) should be a worthwhile read for government officials, its intended audience, but may also be of interest to academic leaders, business executives, and other policy makers in large institutions considering adopting the approaches to catalyzing improvement described within.

Full disclosure: I was one of 25 people interviewed by the researchers, so the outcome was informed on some level by our phone conversation, including my skepticism about the sustainability of many of the approaches I’d documented and analyzed over the years. Given that concern, I was heartened to see a useful swing at a definition for innovation to ground the discussion that followed:

“Innovation has come to mean many things–something good, novel, risky, creative, technology-driven, and so on. Government innovation also conjures a variety of meanings for a host of audiences. A cash-strapped city manager might identify a new town website as his community’s most important innovation in years. A technology entrepreneur might cite a public/private partnership to encourage the repurposing of technology for state use. A parent of school-age children might point to real-time updates on school bus locations as a stellar example of government innovation. A government staffer may describe a training program for employees to develop and incubate new projects as innovative. A Cabinet secretary may cite a new open data portal as an important innovation, changing the way that government does business.

Such varied examples point to the difficulty of operating in the government innovation space. Without clear, common understanding of what constitutes government innovation, it is almost impossible to explore what those charged with encouraging innovation in government are currently doing, let alone what they should be doing and how they should be doing it.”

The report outlines the two standard missions of such innovation offices — producing external impacts upon a larger community or internal impacts within a government entity — and then breaks down the specific goals within those missions, from civic engagement or issue-oriented change to greater efficiencies or cultural shifts. These missions may be enabled (or hindered) by the selection of different structural or organizational models for the innovation office: laboratory, facilitator, advisor, technology build-out, liaison, or sponsored organization. Whether a specific approach will be successful depends upon both context and support, contingent upon a number of choices.

“We think that innovation offices/officers have a lot of potential for creating the desired internal and external outcomes, but only if they are given real responsibility within the organization and are asked to involve others in their projects,” Burstein and Black responded, in answer to my questions about the report’s conclusions.

“The danger of siloing innovation within a particular post is that innovation won’t extend to other departments. But, if the mayor, governor or agency head signals the importance of the mission by appointing the innovation officer to his cabinet, as in the State of Maryland for example, other departments are more easily convinced of the value of the project. Similarly, if the innovation officer involves other departments in his projects, as in Boston’s New Urban Mechanics’ partnership with the City’s Public Works and Transportation Departments, for example, that can go a long way toward infusing the entire organization with a commitment to developing innovative approaches. In other cases, an innovation office may sponsor innovation trainings for staff throughout the organization, as is the case in Nashville. On the other hand, if the innovation office is not staffed, structured or resourced appropriately, it may be useful to consider alternatives to an innovation office.”

While the body of the report provides specific successful examples of each of these models, I was left wanting to read more about where failures had occurred. The most significant criticism I have for the report is the absence of failures: it would be useful to know when the models chosen didn’t work, and why. Failure’s role in achieving innovation doesn’t escape mention entirely: On Page 31, in the table of “sample mission-aligned metrics,” the authors list “more willingness to take informed and reasonable risks, and learn from failure” under “internal change” and “Number of projects that are evaluated mid-course and changed or cancelled as a result” and “number of opportunities for employees to share what they are learning from innovation-related projects within their government entity and with the field” as sample measures. Two of the innovation centers studied, the Colorado Innovation Network and Montgomery County Innovation Program, specifically reference failure as well, citing the importance of sharing lessons and being transparent about such measures.

That said, these are glancing mentions. Two of the most significant barriers to innovation that I’ve observed in government over the years are human resources — inspiring people to public service and getting needed talent through the human resources process are significant challenges — and a tangled procurement system, the bugaboo that lies not only behind high-profile failures like Healthcare.gov but endemic failures throughout government IT.

The third barrier to innovation is a tolerance for failure. Innovation, by its nature, often comes through experimentation, which inevitably results in failures. Inventors, researchers, software developers, engineers, and scientists all know this dynamic well. Thomas Edison famously tried thousands of approaches to a light bulb before he found one that worked. The Apollo Program endured failed launches and tragic deaths.

One of the key success factors behind the success of the “New Urban Mechanics” in Boston lay in the ability of its leaders to insulate experiments against risk. In today’s polarized political environment and 24-hour news cycle, this isn’t a minor issue: any mistakes, much less programmatic failures, can and will be reported upon in the media and cited as evidence of incompetence by the opponents of elected officials. The dynamic leaves public officials in a difficult position: citizens want and deserve more efficient services and effective government programs, supported by cutting-edge technologies and management approaches, but they also distrust the government’s ability to deliver upon them. Some ideologues doubt that government can be innovative at all, despite centuries of evidence to the contrary. We’re still learning how open innovation should work in the public sector in this century, but it’s possible that smaller experiments, prototype grants, and iterative development, along with strong political leadership that emphasizes the need for such approaches and provides cover, may be successful.

In the research interview with Black and Burstein, I observed that the appointment of chief innovation officers and the creation of innovation centers can risk other people in an organization to see innovation as something external, divorced from their job and everyday tasks. Should innovation or open government be a functional area or speciality, or are they outcomes, principles, or approaches to work that every person should be pursuing? If open government or innovation are defined around an existing office or role, like a chief information officer, its adoption elsewhere may be limited. Cultural changes, meaningful public participation, shift in work habits, and collaboration are not, after all, about technology, although they may be inhibited or enabled by it.

Burstein and Black explicitly acknowledge the potential risks of this dynamic in their report, including key factors that inform the decision to build such a center in the first place and a list of alternatives to such offices. These other options include, but are not limited to, “an innovation and leadership training program for selected staff, membership in organizations that promote knowledge sharing around government innovation, changes in recruiting practices to attract different skill sets to government, and public-private partnerships.”

When I asked the authors about other options, they offered more ideas: “There are lots of approaches that a governmental entity might take to facilitating innovative practices internally and externally,” they responded. “Among them: establishing an office of innovation (or appointing a chief innovation officer), offering innovation training to staff members throughout the organization, deploying innovation fellows to departments throughout the organization, sponsoring contests or challenges for staff (or the public) to offer innovative solutions, and changing marketing and hiring practices to attract more diverse candidates.”

One of these approaches, the use of so-called “innovation fellows,” has become popular in recent years, from the Code for America fellows in multiple American cities to Presidential Innovation Fellows, Consumer Financial Protection Bureau fellows, and Health and Human Services innovation fellows at the federal level to Code for Africa and Code for Europe fellows internationally. While such fellowships have been criticized for parachuting in technologists that may not understand the cultural, regulatory, or legal environment of a given government entity, many fellows have produced a growing number of notable successes, including the openFDA project I featured here months ago.

“It is beyond the scope of our report, but we believe that innovation fellows can work very well,” said Black and Burstein, in response to my query about the utility of this approach. “In our conversations with those at federal agencies, we heard about the energy and vitality that Presidential Innovation Fellows brought to their work, often infusing the rest of the staff with a spirit of innovation. On the other hand, the value of such fellowships is highly contingent as well. Fellows often leave after a certain amount of time, completing a single project. If the goal is to develop a particular solution, that may be adequate. On the other hand, if the idea is to create a more long-lasting impact on the culture of the organization, it may be helpful to have dedicated staff for this purpose.”

One of the dynamics I’ve watched over the years is the extent to which fellows choose to enter longer-term public service after their fellowship ends. The majority of Code for America fellows have not, choosing to join technology companies or spin their projects out into civic startups, a decision and dynamic that deserves consideration on its own. Several Presidential Innovation Fellows have joined 18F, the new “startup” with the US General Services Administration, that’s approaching building digital services using modern technology and lean management approaches.

One approach that Burstein and Black did explicitly warn against is the use of volunteers to propose, build, and deploy services or policies. While the intentions of volunteers may be good and their contributions genuine, focused upon improving their communities or the agencies that serve them, public officials that are depending upon volunteers showing up for hackathons or maintaining civic applications should be thoughtful about how such efforts will be sustained in the long term or integrated into government.

“One thing is for sure: financial resources are important,” said Burstein and Black. “A few local governments have volunteer fellows, often housed within an innovation office. And while these individuals may provide real value, overall this approach has significant problems. It is often difficult to compel volunteers to complete projects. At best, they are not integrated into City Hall. At worst, they are resented by salaried employees who may already be suspicious of an ‘innovation agenda’. This is why we believe that getting buy-in from elected officials and finding allies within government is so important, regardless of the innovation structure that is adopted.”