page image

CHAPTER IV - Is Government Hackable?

If transformation and continuous innovation are difficult to accomplish in the corporate world, they are even harder to achieve in the public sector. The Roundtable explored the special challenges of working for change within government and tried to identify strategies that might make it possible.

Jennifer Pahlka founded the nonprofit Code for America in 2009 and served since then as its Executive Director, working primarily with cities, counties and states to develop new apps that improve the delivery of their services. In 2013/14, she spent a year as Deputy Chief Technology Officer for the federal government where her main accomplishment was the launching of the U.S. Digital Service in May 2014.

As someone committed to bringing the creativity and disruptive spirit of tech start-ups to the public sector, Pahlka soon learned that there were several fundamental characteristics of government, and of the federal government in particular, that worked against that spirit: First, government workers can always invoke the excuse that “seeking to bring about disruptive change is illegal.” Government agencies are driven by legislative mandates and constrained by budget allocations. In such a context, avoiding failure is a higher priority than pursuing improvements. The problem with this reasoning, Pahlka concluded, is that focusing on avoiding failure actually makes it more likely to happen.

In addition, there is no such thing as a product manager in government who is responsible for making a program work. When Pahlka reviewed existing IT-based applications inside an agency and identified real problems that could be fixed, she was told that the problems did not matter because no one was responsible for improving them. Ironically, government agencies are motivated by neither a drive toward scalable efficiency (doing more with less) nor toward scalable learning (getting better faster), but rather by the imperative to “do it by the book.” An example is the resistance within government IT shops to moving from traditional “waterfall” processes for software development—an approach that puts all planning up front and “amounts to a pledge not to learn anything while doing the actual work”—to a more modern “agile” approach that is based on successive iterations of development, testing and refinement to deliver the best possible product. Although the private sector has embraced agile development, waterfall development processes are deeply embedded in government IT operations, and there are simply no mechanisms to fund agile approaches.

When she took the job as the country’s Deputy Chief Technology Officer, Pahlka’s goal was “to change the culture of government.” She ended up believing that it was not possible to change culture, but that it was possible to change individuals’ behavior, incentives and sources of inspiration. One of the core beliefs of Code for America was that “every system is hackable” (in the good sense). But during her time in Washington, she concluded that the system of government was not hackable.

Pahlka’s major accomplishment while in Washington was the launch of the U.S. Digital Service (USDS), which represents a deliberate effort to introduce a very different mindset into the operation of government. As is often the case with major innovations, it owes its existence to a crisis: the initial failure of and its rescue by a small team of tech experts recruited from the private sector. The singular success of that effort encouraged President Obama to support the establishment of a small group of professionals based in the White House that could “parachute” into large agencies to work on solving tough IT problems. The USDS was deliberately designed to operate according to a set of principles that stood in dramatic contrast to typical government approaches: “make it simple and intuitive…use agile and iterative processes…default to open.” Among the early successes of the USDS, which gave it greater credibility and increased support, were projects for the Veterans Administration (which, interestingly, was an early adopter of agile software development processes) and the Department of Defense.

Digital Service Plays
  1. Understand what people need
  2. Address the whole experience, from start to finish
  3. Make it simple and intuitive
  4. Build the service using agile and iterative practices
  5. Structure budgets and contracts to support delivery
  6. Assign one leader and hold that person accountable
  1. Bring in experienced teams
  2. Choose a modern technology stack
  3. Deploy in a flexible hosting environment
  4. Automate testing and deployments
  5. Manage security and privacy through reusable processes
  6. Use data to drive decisions
  7. Default to open

Pahlka strongly believes in the potential of data, properly used, as a tool to identify and document problems and motivate change. However, although governments are typically “swimming” in data, much of it is irrelevant or not in a usable form. She cited the example of the food stamp program in California (known as CalFresh), which has widely promoted the use of MyBenefits, a web-based app intended to help potential beneficiaries determine their eligibility and apply for coverage. The reality, however, was that the application was almost impossible to use by ordinary people. As a result, California had one of the lowest rates of program participation in the country even though the state was spending a considerable amount of money to promote it. Pahlka found that there was no useful data available on the performance of the MyBenefits system in terms of abandonment rates or other relevant metrics. She discovered that one simple way to capture and communicate the real problems with the system was to simply record someone who was trying and failing to use the system to get benefits, and then show it to high level officials who might be able to do something about it.

Pahlka acknowledged that it does not make sense to try to make everyone in government innovative. But systems can be “bimodal,” and allow innovation to happen even while continuing to carry out all the mandated functions. (This same lesson could also apply to banks that want to be more innovative but need to maintain reliable systems that do not change rapidly.)

Sometimes the problems to be solved can be very simple: when she worked with the Seattle Police Department to make information about “superutilizers” of the system more easily available, the solution involved providing officers in the field with mobile phones that increased access to the necessary information. The biggest problem in implementing this innovation was helping the new users learn how to open the OtterBox cases that protected their phones!

In making the case for simplicity, Pahlka argued that that the most critical applications should be text-based since text represents the lowest common denominator for users. If government is going to “go digital,” it is important to pay attention to who gets privileged and who may be excluded.

Several other Roundtable participants shared their experiences in attempting to bring innovation to the public sector. As Chief Technology Officer for the City of Los Angeles, Peter Marx had to deal with civil servants who were members of five different unions and had an average age in their 50s who were still running a mainframe in their IT operation. Even though mayors or other elected leaders are often strongly motivated to show progress and do good things, the permanent bureaucracy is held back by process and the need for stability. To make progress, it was necessary for Marx to bring in new people who were committed to making changes, and to get the city’s leadership to recognize and reward people for doing good things. The city has been able to measure progress though such things as customer surveys that show that there is more approval and fewer complaints from citizens. Over time, it is possible to change the narrative that “governments are bad and don’t do new things.”

Another obstacle is the lack of agreement on what defines success in government. Kelvin Westbrook noted that St. Louis County, where he lives, includes 90 separate municipalities and each may measure “success” differently. For example, a consultant to the City of St. Louis who was asked to study the big challenges that the city was facing identified as a top priority the ability to move quickly to prevent the city’s football team from moving. Providing better public services for the poor did not show up on the list of priorities.

According to Alaina Harkness, it is “astounding” how little we know about customer satisfaction in terms of government services and how little we know about the actual impact of the money spent by public agencies. The impact of the failure of for-profit companies to adapt to the Big Shift can be clearly seen in the steady decline in corporate return on assets (ROA). But there is no equivalent statistic to measure government performance. Perhaps the reason is that citizens just don’t care enough: consider the low level of voter participation in elections and the low levels of usage of 311 systems, particularly in low income neighborhoods where better services are most badly needed.

Are there true change agents in government, and if so, who are they? Like Marx, Harkness sees a big gap between the interest in change from political leaders who say they want to make things happen, and the bureaucratic culture of governmental operations that has a vested interest in keeping things running as they are. While bringing in outside consultants is often an effective strategy for initiating change in the private sector, that approach does not work very well in the public sector where established ways of working are deeply entrenched and outsiders are rarely heeded.

Resistance can also come from private companies that have developed strong relationships with government agencies and have built profitable businesses on providing services to them. When the House Oversight and Government Reform IT Subcommittee held a hearing on the performance of the U.S. Digital Service, it heard no protests at all from the agencies with which the USDS had worked, but it did get criticism from traditional IT contractors who complained that these upstarts were threatening their livelihood (even though a substantial majority of large government IT projects, particularly those involving outside contractors, end up in failure). The consultants’ message was, essentially, “Thanks for your innovation, geeks, we’ll take it from here.” (In a similar vein, Peter Marx described getting calls from people who complained that by moving the City of Los Angeles’s web sites to an open source platform that was lower cost and easier to use than proprietary software, he was taking money away from outside contractors.)

Peter Hirshberg also encountered challenges in trying to encourage innovation by the City of San Francisco over nearly a decade. His effort initially focused on expanding access to data that was being created or collected by the city and showing how it could be put to productive use. He began working at the “edge” of government rather than at the core, which would have been more difficult. He started by getting access to data on public transit services and making it available to artists and others who worked with it in a playful, creative way. He has also been experimenting with getting groups of students to interact with government data and figuring out new ways to use it as a form of civics education. The more playful these encounters are, he found, the more engaged people tend to be.

Based on these small successes, he is now attempting to work with core governmental systems. One thing that has changed is that the success of “digitally enabled” services like Uber has created a new sense of competition with traditional public transit services and sparked concern that busses are becoming less relevant to the community. Hirshberg is part of a public/private effort to look at “mobility as a service,” a vision that would integrate the services of transportation service providers (like Uber and Lyft), with traditional public transit, as well as new demand bus services as a single unified system to be optimized. In his experience, the biggest force for change in how government works is “the outside pushing in.”

Harkness also affirmed the potential value of opening government data to new audiences and new uses, but put more emphasis on getting better access to internal administrative data. For example, finding one individual who appears in multiple governmental systems could be a potential flag to identify a bigger problem, but there is no simple way to do this today.

New York City’s Municipal Dashboard

To track progress toward fulfilling public promises made by New York City Mayor Bill de Blasio, his administration created an app that displays nearly 1000 separate goals ranked according to priority, and indicates whether work toward meeting these commitments is “on track” (green), “off track” (red), or “at risk” (yellow). Several dozen staff members, including deputy mayors, chiefs of staff and other top city officials have access to the app via their smartphones and tablets. Although Mayor de Blasio himself does not use the app, he receives regular reports generated by the app. To date, the app is being used only by city staffers and has not been made available to the public.

It seems increasingly likely that tech—in the form of the proliferation of media channels and social networks that reinforce rather than challenge political perspectives—are part of the problem of citizen alienation from government and other institutions. (As Jennifer Pahlka put it, “Tech is destroying respect for authority.”) Now the question is whether tech can be part of the solution to make government more responsive to user needs and more accountable to its customers.

Title Goes Here
Close [X]