More on this book
Community
Kindle Notes & Highlights
Read between
February 17 - February 22, 2024
The fear of having exercised too much power, and being criticized for it, is ever present for many public servants. The result is a compulsion to consult every imaginable stakeholder, except the ones who matter most: the people who will use the service.
The United States Department of Agriculture tried this when it was about to issue a rule regarding the minimum peanut content in peanut butter. Advocates wanted it to be at least 90 percent peanuts, manufacturers wanted to require only 87 percent peanuts, and adjudicating that 3 percent difference under the formal rulemaking process took the Food and Drug Administration twelve years of the 1960s and 1970s. The case went almost all the way to the Supreme Court, and the oral hearing alone took twenty weeks and produced a 7,736-page transcript.12 (The advocates ultimately prevailed.)
In our attempts to keep government small, it’s not the disease that has hurt us, it’s the cure.
The environmental movement got really good at stopping bad things from happening but less adept at enabling, and speeding up, good things. Now the habit of suing has itself become a great threat to our environment.
“I understand that it’s complicated,” she told the MITRE team. “But it needs to make sense to a person.”
Conversely, even what seems like a very simple interaction can become burdensome when legal and policy thinking trump service design.
When you come across services like this, you may assume that they’ve been poorly designed. The reality is often that they haven’t been designed at all.
As with Medicare doctors, the more information you ask for, the more you reward administrative capacity, and the harder it is to help those who need it most.
User Needs, Not Government Needs
Over the course of those stories, which can represent generations, they’ve learned that help isn’t always easy to find. We can’t fix this until we understand that in government, we’re not starting a new relationship, we’re repairing a deeply broken one.2
We don’t lack for examples of government needs, but it can be harder to grasp what the GDS meant by user needs. Defenders of the waterfall process will tell you that gathering requirements, the first stage of waterfall project management, is when the team understands the needs of the users. But software requirements are not user needs. Writing down everything everyone thinks the system might need to do per official policy and per multiple conflicting compliance regimens does not help you understand your users. It doesn’t even really help you understand the bureaucracy. It just gives you a list
...more
“You know what high-precision accuracy requires?” Dave asks. “A really high-burden, complex form that starts to resemble a full application.”
Much of the time, it’s not even the eligibility that keeps a client from getting the benefit—it’s the need to persist through all the forms, interviews, documentation, and phone calls.
being 100 percent thorough and technically accurate is always safer, even if it’s worse for the people the service is supposed to help. But to Dave the tradeoff is worth it. The upside is a better experience for the vast majority of clients, which ultimately results in higher enrollment. A detailed, heavyweight screener may meet an agency’s requirements, but it doesn’t meet the users’ needs.
Project management is the art of getting things done. Product management is deciding what to do in the first place—and also, as in the case of the benefit screeners, deciding what not to do.
Mike Byrne, the guy who built the broadband map for the FCC, estimates that most government tech projects could cost 10 percent of what they do and still provide 85 percent of the functionality.5 I hereby dub this “Byrne’s Law.”
(Equity policies can also make it hard to choose a phased-rollout approach. Government services are not supposed to treat different groups differently, and pilot programs designed to start small before launching big are often scuttled on those grounds. As we saw with the EDD, these well-intentioned policies can result in harm to the very people the policy is supposed to protect.)
We talked but did not communicate.
Classssssssssic
Perhaps they were right, but in that moment I thought about his 7 a.m. meetings and his eighty-three pages and I saw Henry as that kid in class who did the long division out to twenty-five decimal places and completed every extra-credit assignment early. Henry’s superpower was that he was thorough. Now Congress was the teacher—a distant, harsh teacher. And Henry was determined to get an A. No, an A+.
The policy team was at the joke end of a big game of telephone.
In the first year of the program, no more than a handful of doctors even asked about that option. It was easy enough to support those few without any technology at all.
The best solution is almost always a no tech solution designed (but not built) by tech ppl
Making room for product management, and the iterative, user-centric practices that come with it, is long, hard work.
Product management is an active, not passive, take on fundamental questions of representation and voice. It doesn’t rely on simply issuing an open invitation to whoever wants to participate in the discussion. It starts from the recognition that most of the people who need to be heard won’t show up—their relationship to government went sour long ago, sometimes generations ago, as Genevieve Gaudet observed.
The mechanisms we’ve built for public input have inadvertently allowed for colossal capture by government needs, special interests, and even well-meaning but often misguided advocates. Product management done right is an anti-capture mechanism.
paperwork favors the powerful.
“It was absolutely phenomenal. But it wasn’t until that moment in front of that whiteboard that they realized how complicated the policy was becoming.”
Many people assume the tools of the digital trade are bits and bytes. But walls, markers, flip charts, and sticky notes often play a more important role in digital work than screens and keyboards.
This is a best practice in the broader tech industry: teams have a better chance of shipping big, complex software projects if they first ship small, simple ones.
Yadira didn’t want her team to build an inferior product just because “Congress said to.”
“Agile processes promote sustainable development. The sponsors, developers, and users should be able to maintain a constant pace indefinitely.”
This reminds me of a comic by Randall Munroe from his much-loved series xkcd. In it, one stick figure is directing the work of another, presumably a programmer, who is sitting in front of a computer. “When a user takes a photo,” the standing figure says, “the app should check whether they’re in a national park.” The programmer replies, “Sure, easy GIS lookup. Gimme a few hours.” The other continues, “And check whether the photo is of a bird.” “I’ll need a research team and five years,”
A picture or link... Its 2024 comeon
Policy and delivery breathe increasingly different air.
A few weeks into working with the GDS team, he came to Tom with his inches-thick binder of documents and plopped it down on the table. “You know,” he said, “I’ve come to the realization that what I’m holding is really six hundred pages of untested assumptions. I’m just going to set it aside for a while.” Tom, who now works with the global transformation consultancy Public Digital, has now seen that realization occur many times. “Why,” he asks, “is so much policy educated guesswork with a feedback loop measured in years?” It doesn’t have to be.
I often hear that tech changes so fast it’s difficult to keep up. But what government needs to keep up with is not tech but people. In part because of technology, people’s behavior and expectations have changed. Simple apps that do a lot for us have made people less tolerant of complex, multistep paperwork processes. Our devices constantly steal our attention, making it harder to persist through administrative burden. Increased inequality has made the time and attention burden even more costly to those who are most likely to engage with government. We push people to online interactions, but
...more
For one thing, the pay disparities are not nearly as dramatic as one might imagine. An experienced technologist working at the USDS or 18F could make $172,500 a year.7 That’s less than most developers would get at a company like Google, but it’s not that far off from what you might get in a mid-tier role at a tech startup. The big difference is that, at a startup, you are hoping for a life-changing windfall when your company is acquired or goes public. But three-quarters of venture-backed startups fail outright; only a tiny minority of startup workers benefit from an exit that leaves them with
...more
This is partially why i frequently toy with the idea of doing a tour in usds. Im past this salary but if one startup hits id probably seriously consider it
But the developers and designers know that what they do will affect the lives of the people around them and that they’ll get to build systems in the public interest. It’s easy to complain about government but more satisfying to help fix it. As the USDS team is fond of saying, “Decisions are made by those who show up.”
The problem, it turned out, wasn’t with the programming of the site. It was that mail carriers had been compensating for incomplete data for decades. If you own a home and decide to subdivide it into apartments or add an in-law unit, you will go to your city’s building department to get the permits, but you might never officially inform the Postal Service. You’re not asked and you don’t think of it. It’s not a problem for your mail carrier, who gets one packet of all the mail for that address, sees three mailboxes, and simply divides up the letters appropriately. But it means that the USPS
...more
This highlight has been truncated due to consecutive passage length restrictions.
The combined USPS/USDS team got covidtests.gov ready in just three weeks.
“People are afraid that automating services would be harmful. From what I’ve seen, not automating is actually harmful. All these forms drive people crazy. They ask the same information over and over again.”
And because there are so few public servants who know how to work in that role, some of these vendors expect that training their bosses is part of the service they will provide.
As of 2011, the department had 2,258 separate business systems, including 709 just for human resource management and 335 more for financial management.
Vantage ftw
If you measure administrative burden, it will go down.
This fallacy is one of my critiques of the measure what matters worldview
“Software is made by people and for people,”
OUR ELECTED LEADERS keep thinking in terms of money, regulations, and oversight because those are the levers they have most immediately at hand.
Waterfall is a great fit for an institution without a core competency in digital, since it requires only complying with processes rather than exercising judgment. But it does not deliver software—or government—that works for people.
Neither the VA nor its oversight bodies knew that the veterans’ health care application didn’t work, but hundreds of thousands of veterans knew, because they had tried to use it.
There are a number of nonprofits that are dedicated to bringing a diverse set of technologists into the public sector, and they operate as a coalition under the banner of US of Tech.
That I’ve never heard of these people probably means they’re not doing a great job?
Few people read the tax code, or even part of it. What they know about tax policy they learn from filing their returns.
To most people, digital services are policy.
For people who do have to deal with such interactions, delivery matters a lot more than politics.
