California’s first-ever chief data officer, Silicon Valley tech executive Zac Townsend, has been on the job about four months. After being appointed to the position last summer, Townsend went on a weeks-long tour of state government for what he calls “user research” to familiarize himself with what agencies and departments want him to work on, and to set some priorities that could lead to results.
Townsend’s path to California started on the East Coast at Brown University, where he studied applied math and public policy. He quickly made a mark working on data-driven projects for an anti-human trafficking organization in Rhode Island, the NYPD, the MacArthur Foundation and New York City’s child welfare agency. Then he was recruited to do innovation-focused projects for then-Newark, N.J., Mayor Cory Booker. In time, Townsend would move to the Bay Area to co-found a financial services startup, before being named the state’s data chief, working in the Government Operations Agency (GovOps).
In an interview with Techwire earlier this year, Townsend shared his early impressions about California’s data program, the issues he’d like to work on, the roots of his interest in government, and what role he thinks vendors could play in the state’s data efforts. The interview is abridged and edited for clarity.
Q: Your mix of tech expertise and government experience is somewhat unusual. Where do you think your enthusiasm for public policy started?
A: Part of my passion for government service comes from the fact I was raised by a single father, and he worked as a postman. But then, after that, he was unemployed, and there was this moment in my teenage years where we were applying to a bunch of government programs trying to get government assistance. Just as a child who had lived with technology and had used the Internet and had spent money on broadband and all that stuff, it was bewildering to me that there were all these different offices and you couldn’t go to one place and you had to fill out the same form over and over. … I sort of developed this odd passion for making government programs and government services work, and work better.
Q: What types of projects did you work on for Newark, and are there parallels in California state government?
A: It was very much similar to the work we do here at GovOps — thinking about what are modern techniques of business, management and technology and how do you apply those methods to the public sector.
For Mayor Cory Booker I did everything from think about an incubation space for Audible, which is in Newark, to doing broadband negotiations for the schools, to something we hope to do here – using data to make policy decisions. An example was we had a certain amount of budget to tear down vacant buildings in Newark. We got all the data on the vacant buildings that had fires, and we built a model, and then we ranked those that were most likely to have fires. Another project I did there was to make it easier to get a birth certificate, because every day I walked into Newark City Hall, and there was a gigantic, long line for people to get their birth certificate. It was a simple thing, but the office was using outdated technology, and with little more than an Access database, we reduced wait times significantly.
So I think all these things are important, from little ways that people interact with the government to big policy Implications or allocating resources, and data — where these other modern tools can
play a part in making these decisions.
Q: It’s still early on and you’re relatively new. But what were your first tasks as the state’s chief data officer?
A: I don’t mean this as a buzz phrase, but I’m trying to do user research: Who are the people that are going to use the data that we have, and how can I facilitate? Ultimately the job responsibilities, as written, are to work on and improve all data collection, data analysis, data sharing, data transparency and data standards for the state of California. And to be candid, in two and a half years, it’s not as if I will be able to wave a magic wand and get all of that done to a perfect degree across California – one of the largest operating enterprises in the world. So my first task is to listen to people, to hear what they’re saying and what they want, and then formulate a plan of the actionable deliverables I hope to have in my time here.
I think it underscores that [GovOps Secretary] Marybel [Batjer] was looking for someone who was familiar with and literate in modern data science techniques and technology. She wanted someone who could speak that language, understood the private-sector attitude of how to use data to make decisions and think about performance and to improve all that we do. On the other hand, she was looking for someone who had some experience in government and was realistic about the pace, how to make change and building coalitions.
Q: That said, have you thought about what you would like to focus on?
A: There are four broad areas I have been thinking about. The first is continuing the great work GovOps has done around open data. We’ve launched a new open data portal, we’re moving into having a pilot that focuses on data sets around climate change and green [government], broadening that scope to having a full-service open data portal and a one-stop shop. The only thing I’m thinking about there, is to focus our efforts on high-value data sets. Sometimes, you will see other states and they pride themselves on the [high number] of data sets they release, and it’s really just Excel files or what’s most readily available on someone’s computers. I want to focus on what people want, what they need, what’s the most interesting – and then try to get that data released.
The second thing is there’s a lot of data the state has that won’t be made publicly available [because] it’s personally identifiable and at the individual level. But there’s a lot of work, particularly analytic work, that can be done with that data. Something that happened before I was here, is strong academic partnerships where individual departments work with professors and universities on research questions. … Those sorts of partnerships are the type of thing I want to further and deepen.
The third thing is, how do we use data in our operations? How do we use data to manage things and have a data-driven process? I think what ultimately will happen, at this early date, is there will be a few departments who are very interested in using data-driven management, and we’re going to do a demonstration project or two: Here’s what having a department dashboard would look like, for example. What I’m realizing is that our big challenges aren’t so much that we need new, fancy machine-learning tools or a lot of programming. It’s less that and it’s more looking at the mission statement of a program and its essential purpose. How would we measure the right outcome for a government program?
The fourth thing is data standards. And that’s both data standards within the state of how we store data, and what the metadata looks like, but also there are a lot of reporting mandates on counties. Some counties and cities play well together, and some don’t. So there’s definitely an overarching question: Is there a way to standardize data around – “Insert topic area” – so that cities and counties and the state could more readily share information?
Q: What’s your initial sense of how mature California’s data program is?
A: I actually grade California quite highly compared to other states. There’s a very robust open data program. I would say that once the GovOps portal is federated, I think it will show that we have nearly, if not the biggest, collection of data publicly available. Data-driven management work and actionable intelligence work also is pretty far along in its thinking. Truthfully states are just not very good at it. Ultimately I think we’re entering a third wave of innovation. I think there’s a lot of innovative work that happened in cities, and that work is spreading. I think it is states’ time. It’s our moment.
Q: What are your thoughts on how the private sector might be able to help California with its many data projects?
A: I’ll be honest, Silicon Valley is certainly becoming a little more political, for better or worse, in the sense that they’re realizing that governments matter and I think that has a wide range of implications. You can see the dialog changing in the Bay Area around housing, housing policy, around Airbnb and Uber. In general, I think there has been a recognition that government matters, that regulation matters, and that being engaged — whether in the private sector, or in my case as many other people who’ve chosen to serve going into government.
As far as engaging on the data front itself, I’ve been having some conversations. People reach out to me or GovOps Secretary Batjer before I was here, about what sort of data the state government could release [to the private sector] that could be helpful. One example is Caltrans: It’s not the most beautiful API [application program interface] in the world, but Caltrans releases a ton of data every day on traffic. And Waze and Uber and Google and Apple all use that data very day. So I do think we want to engage those partners, we want to think about what data the state has that might be useful. All those companies are mapping companies, and the state has a lot of data on roads. And that’s historically how we’ve engaged them. But I think across the spectrum I’m interested and open to finding partnerships or data that we have that could be useful to the private sector.
As for vendors, to be candid, it’s just not up to me. There are certain things we hope to accomplish, like we have an open data portal. There was a bid, there was a [procurement] process. But I think the most important thing here, and whether it’s [California’s] child welfare work or any number of things the state is doing, I think what you’re seeing is that, whether it’s California or others, governments and vendors have gone in lockstep together and gotten further and further away from the most modern technologies. And that makes a lot of sense because, governments want to be conservative and want things that work — and the vendors become more conservative because they’re not selling you agile solutions, or agile solutions development. And you’re in this weird symbiotic relationship where you just spiral toward having big IT systems that are built in a certain way.
I think what you’re seeing now, at all levels of government, is you’re breaking out of that. Maybe vendors that I’ve talked to are relieved. It’s not like these vendors don’t have huge arms selling to the private sector that aren’t doing modern, innovative and interesting work. For some of them they’ve had to keep around legacy technology and infrastructure just because that’s what their government clients want to buy. So I really saw more enthusiasm than apprehension about new solutions [integrating data] that have outcome measures, a data dashboard for a department director or program manager to see what’s happening operationally in real time. Yeah, I think those are all conversations that are happening, but again, those are conversations driven by Amy Tong, our state CIO, and the Department of Technology. But I’ve only seen enthusiasm about my role and the type of work we’re hoping to do in the future.
