US government programs aren’t always the most user-friendly1.
During the pandemic, many people - we all know some - lost their jobs and filed for unemployment insurance. By summer of 2020, hundreds of thousands of Californians had applied for benefits without receiving them. When they reached out to the state agency responsible, the Employee Development Department (EDD), a dismal 1 out of 1,000 calls was addressed, typically by novice staff unable to assist.
Here’s another example: If you wanted to file for SNAP (money from the government to purchase food - aka food stamps), you’d need to fill out an application with 200+ questions. If approved, you had no app to check your card balance, causing potential anxiety during shopping since the only way to determine your balance was by reviewing a receipt post-purchase or through a lengthy hotline call.
In Recoding America, Jennifer Pahlka delves into these issues and their solutions. After working in the private sector, she founded Code for America and served as the US’s Deputy Chief Technology Officer, offering her a unique perspective.
What makes it hard for the government to deliver effective solutions?
A wide gap between the law-makers and the law-doers. Government revolves around three P's: policy (rule-making), politics (influencing policy), and (im)plementation (actualizing bills into tangible changes). In DC, policy and politics are seen as important, exciting, and prestigious. People care and think much less about implementation. Lawmakers might put certain specifications on a law and not realize how hard they are to implement: for example, making a Spanish language Obamacare site would have been very time consuming at launch (“You can either have one site in English that works or two sites that don’t.”) Frequently, policy executors opt for the strictest interpretation to earn "extra credit" (e.g., asking food stamp / SNAP recipients if they own a burial plot given it’s technically an asset).
Some government employees are terrified to deviate from existing processes, even if they are inefficient or ineffective. Often, they are judged on process and opposed to outcome. If the outcome isn’t great, it’s understandable, but it’s when they don’t follow process that they can get into trouble with oversight committees and the like.
Government policy mandates that the government not do anything in-house that it can effectively outsource. This makes government at time hostage to contractors and makes it take longer to get anything done, as it must go through a bid process, etc.
Kludge: there is rarely one system, but instead multiple levels of systems that have been tacked on and become fragile and complex. At the US department of Veterans Affairs, for example, the diagram for one benefit system alone was displayed on a wall 20 feet long and 8 feet tall. Let’s take the EDD for example:
Original software layer: a green-screen database written in the ancient language COBOL that ran on mainframes.
Eventually, people started writing macros - small bits of code - to automate processes. Employees user interfaces now included these macros, which saved time, but also increased fragility, given there were only a few people who knew how to write macros and how the EDD’s specific macros operated.
Three added systems in the 2000s duplicated the original, slowing things down
With kludge, each solution only makes the problem worse in the long-term. I wrote more about this in my review of the book Sludge, which touched on many similar themes to Recoding America.
Thoughts and reactions
I absolutely loved this book. I find the general topic of “how to get things done” to be interesting, and the fact that she went deep on one organization was much more helpful than books at a higher level (I enjoyed this more than How Big Things Get Done).
While obviously some of the lessons are most applicable to government, I found the general themes relevant to work and life in general.
I also came away with a deep amount of respect for her and the people she profiles in the book, whether it’s the long-term employee who changes things from within or the former Google software engineer who foregoes a salary and lives off food stamps to see what the experience is really like. Initiating change is challenging, especially within the government. And while government employees sometimes get a bad rap, there are plenty of employees here who work extremely hard to get things done within an inefficient system. There aren’t many villains, but instead people doing the best they can in a system much larger than themselves.
Some themes that I appreciated:
Are you a project manager or a product manager? Project managers drive projects forward by coordinating across multiple stakeholders, keeping track of deliverables, and “getting things done.” They are great at implementing. Product managers, though, involve designing the actual product and making sure it fits customers needs. According to Pahlka, there are too many project managers in government and not enough product managers. More broadly, I thought this was a helpful framework and probably good for everyone to think about at some point – “am I doing what I’m doing simply to get this done, or have I thought about the why?” Which brings me to my next point..
Who exactly is the customer? A structural issue that she identifies is that the “customers” of many government of organizations are not the citizens who would benefit from the program (e.g., people applying for food stamps or insurance through Obamacare), but instead Congress, who passes the law and oversees its implementation. As mentioned above, agency employees might be tempted to get “extra credit” by implementing a law in as thorough a way as possible, even if it makes it a much harder experience for the end-user An important “life lesson” is that you are often not the customer:
Healthcare: Your provider is largely focused on getting reimbursed by your insurance company and keeping them happy.
Technology: Most websites / social media platforms are not designed with your wellbeing in mind, but instead to maximize engagement and monetization (customer: advertiser).
Apartments / real estate: Property management companies work for the owners, not the tenants.
Typically, if you don't pay, you're likely not the customer.
The danger of “do no harm” and “two jobs.” As mentioned above as well, many government employees are judged on process and opposed to outcome. This echoes Ray Dalio’s sentiment that everyone has two roles: their actual job and managing perceptions of their performance.
The difficulty of being plugged in as a senior person. Senior lawmakers or executives may overlook on-the-ground policy implementation challenges. One story in the book centers on the VA, whose website to apply for benefits didn’t work. One veteran had tried to apply dozens of times only to get the same error message. At some point, he agreed to be recorded and interviewed while applying. This video made its way to one of the VA’s leaders, who was shocked and embarrassed that this is happening and was supportive of the necessary changes. As a senior leader, you’re supposed to “manage by walking around,” but this can be harder than it seems. Your direct reports are trying to “manage up” and control your expectations. More junior employees might be afraid to speak their minds.
Technology v. resources v. Process: There’s an idea that government (and many companies) run on old mainframes and their problems could be solved if they modernized their systems and software. While the first part of the sentence is definitely true, she argues that fixing processes is much more important than fixing technologies. We often think “moving to the cloud” will solve everything, but if the underlying process for employees or end-user is just as bad, not much will change. A 200+ question application for food stamps is hard no matter where the technology is hosted. And similarly, politicians will sometimes make a big deal about hiring more people, that that’s not always the right solution: in the case of unemployment insurance, it slowed things down as these new employees had to be trained and couldn’t add much value immediately.
People, whether in government or the private sector, typically act based on their priorities and constraints, not out of malice. E.g., making their boss happy, following process so they don’t get called out by an oversight committee, being as loyal as possible to the letter of the law. And lawmakers, in insisting websites work in a variety of languages, are trying to create equality as opposed to IT nightmares.
Everything has tradeoffs. Every feature you add to a product adds some complexity and takes away time from other priorities the team has. Sometimes it’s worth it, sometimes it’s not. But you can also be flexible and come up with different ways to generate “features.” For example, Obamacare only applies to US citizens, but there are 17 unique immigration statuses that it also applies to. Even though these cover a tiny percentage of people, configuring the website to validate and handle all of these edge cases was incredibly time-consuming. They could've been directed to a call center, allowing the website to function effectively for most users. She references “Byrne’s Law”, which suggests that most government tech projects could be executed at 10% of the current cost while still retaining 85% of the functionality.
The Amazon effect. I don’t think the DMV has ever been known for providing seamless digital experiences, but we likely perceive it as worse now given consumer apps are so convenient and easy to use. At the same time, should we hold everything to the bar of a trillion-dollar company with virtually unlimited resources?
The book (and this piece) are focused on the United States government; I don’t have a perspective on other countries’ governments.