May 23, 2018

Bridge SF City Spotlight | City of San Francisco & Exygy

Roshen Sethna, Exygy’s Civic Partner, and Barry Roeder, at the City of San Francisco, share the story of creating DAHLIA, San Francisco’s affordable housing portal in 2018.

Meet Roshen and Barry

Sethna: 

Hi, everybody. My name is Roshen Sethna and this is my colleague Barry Roeder. We are here to talk to you today about our latest wave of digital innovation in government, and we're here to talk to you about how it's kind of doing two things. One, it is reinventing service delivery between agencies and citizens and, two, it is really starting to help transform government as a system and that's really exciting to us. Barry and I have had the privilege of working on this digital product here in San Francisco with the Mayor's Office of Housing and Community Development where Barry works. We have seen how it's improved service delivery and we've also seen how it's starting to transform government especially at the policy level. 

I'm going to first introduce myself and give Barry a chance to introduce himself. Like I mentioned, my name is Roshen. I'm a Partner at Exygy. Exygy is a digital innovation studio based here in San Francisco in the Mission. We are focused on building healthy communities and I lead our Civic Tech practice. Exygy has a really deep Civic Tech practice even before I was at the firm. We've worked with several local agencies and regional agencies including the Mayor's Office of Housing but also local transit organizations like BART and MTC, educational organizations such as the School District and the Committee of Information Technology, and a variety of agencies at the City and County of San Francisco. Personally I was really excited to start working on this housing project a few years ago because my background in education is policy, but right out of school I was a housing counselor in Houston. This was right after the financial crisis so there were a lot of rules that were changing with regards to housing especially homeownership. One of my jobs was to help first-time homebuyers in particular areas navigate these new rules. It was pretty tough for people to stay up to date on what was available to them and what they were eligible for. My role as a counselor was really important but also involved a lot of manual research and looking things up and double checking things. I was really excited to work on a product in San Francisco that could help people access those resources on their own and navigate them in a really easy way. Barry's going to talk a little bit more about the product that we built and introduce himself. 

Roeder: 

Thank you. Good morning, everybody. My name is Barry Roeder, and it's true, I work at the Mayor's Office of Housing and Community Development in San Francisco. We are the agency that's responsible for building and overseeing all of the affordable housing in San Francisco. We have multiple portfolios of housing. My role there is Strategic Initiatives and I'm focused particularly on digital strategy for the department. I've worked with Roshen as she indicated on one of the projects that we'll talk about today, which of course is our DAHLIA housing portal. The Dahlia housing portal is a San Francisco affordable housing portal at housing.sfgov.org. It is a product that we built to create a single-stop resource for people that are looking for and wanting to apply for affordable housing that didn't exist before. We've been fortunate recently to win a couple of awards. One is the Spur Good Government award and the other is a Digital Leader award. We're getting some traction and we actually have some news. We're expanding the platform regionally. Alameda County has signed on recently. It'll be the first regional platform of its kind which is pretty exciting. We'll talk a little bit more about that as we go forward.

Government Solves Public Problems with Policy

Roeder: 

Before we go further, though, I thought it would be useful to talk a little bit about the role of government. Some of you might know, depending on the circles that you travel in, it's pretty easy for technologists to be dismissive of legacy systems. It's clunky, it's old, throw it out. Let's have a hackathon over the weekend and we'll just replace everything with an app. Well you can imagine, you can think about services that you get from your jurisdiction wherever you live. There are certain things like cleaning the streets or showing up to vote or certain things that couldn't be replaced by an app. There are some really great things actually about legacy systems. Some of them, in fact, are still best practice. First of all, it's a democratic system obviously with an eye on equity. There's a focus on all the constituencies, not just who has the money for whatever app. It's the way we're set up, it's how we do business as a democratic society and that's represented in our legacy systems. So again, some of these systems that when they were introduced were initially best practice, the problem with them is that innovation comes slowly. Things don't move that quickly in terms of how we improve or how we respond to current needs with the systems that we have. We'll talk a little bit more about how we can make some adjustments to that. 

Current Government System

Sethna: 

To dive a little bit deeper into the current system because Barry and I believe that you need to understand the context in which you're working so that you can involve the system itself. You guys are probably really familiar with this, but one of the biggest levers that government obviously has is policy and oftentimes policies made at the top, that might be housing policy, transit policy health care policy, is then pushed down to the agency or department level. There might be a housing authority, a healthcare department, a transit authority that is carrying out a lot of the policy that is set by legislators or policy makers. The way they often do that is through programs and services that then you and I interact with as members of the public. One thing we've started to see in the last 10, 15, 20 years as people have started to say, "How can we better serve citizens? How can agencies better deliver services to citizens through those programs and services that they're offering?" In the digital age we've started to see that it's often done with digital products, so this is where digital products tend to sit. Whether they're websites, mobile applications, informational applications, they sit at this level. They have really helped agencies better deliver services to the public. One of the really big examples of this that you guys are all familiar with is the healthcare.gov redesign. We've seen digital innovation work really well at this level, but we think that these products actually have an opportunity to change more than just service delivery. That's what we're going to chat a bit about today. 

What is DAHLIA?

Sethna: 

One product we wanted to talk to you guys about is a product that Barry and I have worked on which is called DAHLIA. DAHLIA, as Barry mentioned, is a one-stop shop for affordable housing listings in San Francisco and it also offers a space for citizens to apply for that housing. DAHLIA stands for the Database of Affordable Housing Listings Information and Application. There will be a quiz at the end of this and it's also the city flower which is cute. We were able to centralize resources in one place, but it took the last two and a half years to robustly build out this site. When we started the project in 2015 and Barry's team had started it really earlier than that. 

The Problem

Sethna: 

When Exygy came into it, we were trying to understand with the city what was the landscape we were facing, what's the actual problem that we're really trying to solve. The Mayor's Office of Housing contracted with Exygy as well as a couple different agencies to help solve this issue. What we first did was we understood the landscape. We at the time understood that the late Mayor Ed Lee had some goals around housing. He was slated to build 30,000 new units by 2020 and 10,000 of those are slated to be permanently affordable. We all know the City has a housing crisis and there were efforts being made to increase the actual housing stock. However as this housing stock is coming online, one of the problems was how do you find it, how do you find all these units and various buildings and navigate them. The answer to that at the time was not very easy. 


Previous housing sites

This is what people faced when they were searching for housing in 2015. There were several different government agencies sites. You might go to the healthcare department, the housing department. Also, oftentimes nonprofits created their own resources, so you could go to those lists. It was really tough for people to understand what information is most up to date. They were seeing inconsistencies between different sites and they were also having to navigate text heavy sites to understand if they were eligible for this, is their household size too big or too small, is their income level just right. They're trying to navigate a lot of resources and we did have a lot of what we call "power users" that were able to navigate this system and knew it really well. Let's say you were one of those "power users" and you did navigate the system. You were able to find a housing listing or a resource that you were going to apply to. What happened then? At that point, you were met with an arduous application process. At the time the application process was all on paper, so no digital application. You had to usually go in person to pick up the application because each building had a different application to make sure to get the right application for the right building. And once you did get your hands on it, these applications were really lengthy. 15+ pages often required a lot of supplementary information: your tax returns, your pay stubs, things like that. It was a really cumbersome process to get the application together. It might take several days to do that. Let's say you were able to do that. You took several days, you're able to get through the application process, compile all the information you need for that listing, and you're ready to turn it in. At this point you would go to the building generally to turn it in and you'd be met with long lines to submit applications, which was not a really great experience for housing applicants because oftentimes all of the applicants would need that full application window to put their application together. Many times people were waiting until the last minute to apply up until that 5 PM deadline on that application day. This is not only a bad experience for the housing applicant but a really stressful experience for the leasing agents and the building owners who are trying to manage getting everyone's application in by that due date. It's a really stressful process all around obviously if you're a parent, if you have a couple jobs, if you don't have the time to stand in line for several hours. People were often applying to several listings. This experience would happen every time you needed to submit an application. Let's say you were able to get through this step and you actually turned in your PayPal application to the leasing agent. What was that post application process like? I might let Barry speak to this because he's particularly passionate about the post-application process.

Roeder: 

Roshen knows when we talk about DAHLIA that one of my favorite things is right here: the carnival drop, which I think Twitter or Google or most tech companies don't have one of these. You turned in your application and we literally pulled out a roll - think about the toilet paper unfolding - of the red carnival tickets and you got one and one went in the drum. We had to tape the door of this drum shut because it flies open when we spin it and we had to pull the tickets out. It's an absurd process at least for this day and age. I want to give credit and thanks to the folks that came up with this whenever they did decades ago. I'm sure it was best practice at that time. It's not best practice now. Quick note about the fact that we do place our units through a lottery process in San Francisco, that varies by jurisdiction in the Bay Area. There's a combination of weightless and lottery. Both of those things are functional in DAHLIA.  

Sethna: 

Just to add for transparency purposes, let's say a thousand people applied to a listing. The Mayor's Office of Housing staff would literally have to pull every single ticket out of that drum. Because if I'm an applicant, I want to know that I got pulled even if I got pulled at number 900 or 950. So again, not only a stressful process for the applicant but stressful for the Mayor's Office of Housing staff who are really managing these lotteries for each of the buildings in San Francisco. Once the applications were actually in and the lottery was run, oftentimes these lotteries were just posted and this part actually was digital. They were posted online and people could see their ranking in the lottery system but there's not a really great way after that to know when you're going to get a listing. That's kind of the the different problems we are facing both in the housing landscape and the problems that the citizens were facing that were applying. 

The Solution

Browse DAHLIA at housing.sfgov.org

Sethna: 

Our solution was DAHLIA. You can access DAHLIA at housing.sfgov.org. I can talk at length about the different features that we have built out over this site because we at this point have been working on it for over two and a half years. One of the ethos that we had when we built this site was to offer affordable housing seekers the same level of experience and dignity that market rate housing seekers would get when they're searching. An experience that you may be really used to on a Trulia or a Zillow. We wanted to offer that same level of experience. One of the things we did throughout the last two and a half years is every few weeks we user tested with users. Actual housing applicants past, present, future. We put mockups or actual code in front of them, ask them to do a certain task, see how they got through that task on the site. One of our early user testers was looking at the screen in front of her and said, "Am I on the right site? I'm not sure if this is the right site." We're like, "Yeah, that's right. This is DAHLIA. We'd like you to see if you can find a listing you're eligible for." She said, "I don't think this is the site I'm supposed to go to. This looks too nice. I don't think this is affordable housing." This is something we were also trying to do with DAHLIA: rebuild trust between government and citizens. Everyone in this room has stood in line at the DMV at some point for six hours. We're not used to a certain level of experience when we interact with government, so that was part of engaging users in this process. To have the government agencies say, "Yup, this is for you. These are resources that we want to serve you with." 

A couple of things that we did on this site. One, we streamlined the actual housing listing to the right. We made it a really standardized format so now all of the listings have consistent data and information. We also turned it into a storytelling format from the text heavy sites to the new product. We surface information people cared about the most. People told us during user testing, "Just tell me the minimum income I need to get into this unit and tell me what the rent is. Tell me those two pieces of information and then give me more information if I want to dig in." We designed the information hierarchy of this page very intentionally based on what people were telling us they needed when they're browsing. We made it a storytelling format. The second thing we did was we standardized and digitized the application process. So that's on the mobile screen. We turned it into a step-by-step really easy to read application process. We call it our short form application process because one of the things we did is that we talked with a lot of leasing agents and housing developers. We asked them, "Hey, at this point people are really just applying to be in a lottery to have a chance to get a unit. What is the basic amount of information you need from people? Why are you asking them for their tax returns? Do you really need that to run a lottery?" We worked with them to understand what the minimum fields were and we're able to reduce what we were asking of the user to make it comparable to what they're applying for. If I'm spending a few minutes applying and my chances are maybe not so great because the lottery's big that time, at least I'm only spending a few minutes applying instead of spending several days. So again, we are treating the user with this level of dignity. 

The site is also translated into the city's main languages. It's accessible so folks that need to use screen readers, for example, can read through it. This product set up a foundation as these listings come online, they can get inserted into this standardized format that we're using now. 


DAHLIA impact by the numbers

‍A couple of statistics on how the site's been doing. It now takes 15 minutes to apply online. It's shorter if you even use one of your past applications to pre-fill a new application. That took that several day process and reduced it down to 15 minutes. When the online application feature went live which was in November of 2016, we saw that about 90% were using online applications. We were actually able to increase that 85% to 90% by doing several things like translating applications into different languages and things like that. We're trying to move folks online because that also helps the leasing agents with paper data entry. It reduces the amount of paper. We've seen 80,000 since November of 2016 and about 130 listings have gone through the site. Something that we are super proud of and we are excited to see grow is that about 100 households have been placed in units per month. Of course as listings come online and as larger listings come online, we're seeing this grow and we're really excited to be able to track how many folks we're placing over time. We've had a lot of success with this product on the service delivery side and the ability to measure how many people were placing. But, it wasn't just dropping design and code in people's laps. We weren't saying, "Here's a site. Great. Use it. Moving on." There were a lot of things that happened in the background like change management processes at the City that have made this product successful. 

Our Learnings

Sethna: 

One of the things we wanted to share with you guys is some of the learnings from what has made DAHLIA effective so that at your agencies and organizations, if you're embarking on building a product, you can learn from the stuff we learned.

Roeder: 

That's right. I think this is the point where I say, "Don't try this at home." The Mayor's Office of Housing and Community Development tried this at home a couple times before this and it wasn't successful. There's a lot that we've learned along the way that we thought we wanted to share with you about key elements in creating a successful public sector service delivery product. The first of them is the idea Roshen talked about asking the user. 

There's a story that I love to tell. Jay was talking earlier about and of course this conference today is about building bridges between the different sectors. We were fortunate through the Mayor's Office of Civic Innovation at the time to be the recipient of one of the first civic bridge projects that placed some very sharp Googlers with the Mayor's Office of Housing and Community Development for four months to help us look at the problem and how we would approach this. One of the things that we asked them I remember is, "Wow, not only are these Googlers but these are like really smart Googlers. So what's the secret of how you guys ask the user? What's the Google secret?" 

Lesson 1: Ask the user

The idea is to sit regularly with people that actually have to use this. It doesn't mean let's have a community conversation one afternoon in a conference room and think about what are you looking for, or let's build something and then send out a survey monkey and ask if they like it. On a scale of one to ten. No instead, let's sit with people all the way through and talk about what their needs are. There's a whole ecosystem around this project. There are housing developers, leasing agents, property managers on one side and housing counselors and other housing advocates on the other side. They've all got valuable things to bring to this, so having them at the table is really important. 

Lesson 2: Start small, test, and iterate

The next piece is something that probably a lot of you in the room are familiar with but the idea of developing this in an Agile fashion. The idea of Agile; the story I like to tell about that is that for those of you that went to college and are of a certain age, you probably turned in a binder for your semester project and it had the tabs in all the right spots and everything was perfectly tight. You didn't dare turn that thing in without everything perfect. The problem with that though is that you did that in your dorm room or whatever in a vacuum. You were creating that without input from the people that had to buy it because your professor or whoever. You didn't have to start some conversation about, "Wait a minute. Should I make some adjustments here?" Think instead about taking that binder and take the two or three most salient pages out of the middle, rip them out, sit down with people that actually are going to buy your product, make sure that those are right, stand it on its feet as a proof of concept. Or in agile speak, they call MVP, the minimum viable product, and then iterate from there. That was something that was really helpful to us here. Here it shows you what a typical agile cycle would be like. This actually could be a cycle for any project except that instead of doing this once over the course of 18 months or two years, which some public sector IT procurement processes look like, and then they come back with their 60 million dollar product that doesn't work, they do this over two years and then do it again and then do it again and then do it again. 

Lesson 3: Build the right team

Of course fundamental to any of this is to make sure that you build by having a good, solid team. You want to make sure that you get your change agents. Of course that's critical to any process, especially if you're endeavoring to do something differently. Fostering a collaborative culture and secure air cover. I'll give an example that's related to what we talked about a second ago with regard to agile. Imagine going to the controller's office who's used to sending out your request for proposal for a certain amount with specs. You're saying, "Well, we don't totally know our specs yet and we don't totally know what the total project cost is going to be. We're going to just do it every couple weeks. We'll look at how much we spend." That's an interesting conversation to have, but what our CFO and our department says is really great. Benjamin Klosty makes the case, "Yeah, but you know what? Every two weeks you're getting something. You're getting delivered something. And at any point if you don't like what you're not getting, you shut it off. Consult best practices. We talked about the benefit that we received from our working with Google and I can't underscore the idea that if there's something you want to do, look around and see what's been done, of course. 

Lesson 4: Find the right partners

We've been very fortunate in our work with Exygy to find someone who cares about the problem. It makes a big difference to have somebody, particularly in an agile process, that's thinking with the same mindset about how to solve this problem because none of us know what the outside is going to look like. Before then, though, is the idea of building modularly. We talked about that and I mentioned that the procurement process has to change. Maintaining open communication: What I want to say about that is that rather than having a kickoff meeting, meeting again in 18 months, and seeing how it turns out, we have meetings - we're probably in six or seven meetings a week together in one fashion or another. Whether it's the design or engineering or project management piece, it's a constant communication partnership working on the problem.

Sethna:

On the procurement side, it's important to be able to set up procurement so that you can work in an agile format. Typically a lot of public sector agencies will work by sending out a large spec sheet to the vendor and saying, "Hey, come back in a year. Show us a product. Cool." At that time, user needs have changed, business rules have changed, and the product doesn't meet the need anymore. If you're able to set up a procurement process like Barry and Benjamin and others were at MOCD in a way that's saying, "Hey, here's some hypothesis of what we want to build because we've talked to users and done some discovery work, but we really want you to continue testing, test some more, and continue to build it out." This really helps. Barry's team was able to release this product in 90 days. Three months from us writing the first line of code, we released something and all it really was was a site that had one type of rental listing listed. Nothing more - no online application, no translation in other languages, no accounts. We started to see how people interacted with the site and then started to layer on features and we were able to course correct because of that. Then MOCD in that scenario, is able to fully utilize Exygy's skills in design, engineering, and product as opposed to dictating what has already been decided to build. You're not getting the bang for your buck from the public sector side to do that. 


System diagram

Roeder:

In summary, ask the user. Can't emphasize that enough. Start small, test, and iterate, build the right team, and find the right partners. Let's check in again with the diagram that Roshen introduced to us and see where we are. We talked about introducing products between agencies like MOCD and our constituents, the public, to improve service delivery primarily. But what we found though in this process of sitting with people and listening to their needs and not just applicants but all our partners, is that we were building a feedback loop. We were building new communication with this constituency that helps us do our job better overall even outside of the context of the product itself. 

Using Data To Transform Policy

Sethna: To dig a little bit deeper into that, this feedback loop story is really around data. The data that we've seen has started to help us improve services but it's also starting to help improve policy. By policy, I mean policy at a variety of levels. I'm going to give a couple of stories around how DAHLIA did this. 

Use data to increase program effectiveness

One story is at the program effectiveness level. Like Barry mentioned, the City of San Francisco runs their housing on a lottery system. The leasing agent will run a lottery and then use those ranked lottery numbers to lease up the building. In that lottery process, we have something called preferences. It's basically certain citizens who get preferential treatment in the lottery because they might have been evicted from a certain neighborhood at a time or currently live in the same neighborhood as the listing. We want to keep them in that neighborhood and retain them. There's different preferences for people in the housing lottery process. One thing we were seeing is that people were really not taking advantage of these preferences that actually gave them a huge leg up in the lottery, and one preference they were not taking advantage of is called the Neighborhood Resident House preference. This is the preference where if you live in the same neighborhood as a listing, we'll give you preferential treatment to be able to retain you in that neighborhood. Only about 40 folks were actually claiming this preference. What we did was we simply redesigned this page. We added a banner at the top and said, "Hey, you qualify for this thing and this is what it is. We explained it to them because nobody understands what Neighborhood Resident Housing preference means. We actually changed some of the content. We worked with Barry's office to say, "Hey, can we just call it Live in the Neighborhood preference? Is that okay?" They said, "Yeah, that's okay as long as somewhere else on the page we put Neighborhood Resident Browsing preference because we have to have that on page." We worked within the business rules of the organization to make things more user friendly and we saw the claim rate of this preference jump to about 88%, more than double the claim rate. Really simple design and code changes make it so you're able to see if people are claiming services that they are eligible for, and if they're not, how can you drive them to those services and programs and encourage them to claim it and describe it to them in a more human way. 

Use data to improve business processes

Another example is a little bit about the programmatic level - going to the organizational level. Barry's colleagues at the Mayor's Office of Housing are currently as we speak rewriting the housing policy manual that the Mayor's Office of Housing uses. It's a lot of business processes around how buildings are listed, what the building needs to do to report to the Mayor's Office of Housing -  all the business rules around managing the housing process. A lot of conversations in that rewrite came from the product. One really small example of this is that on a listing the minimum income you need to get into that listing is generally a multiplier of rent. If rent is 500 bucks a month, then the minimum income requirement might be two times that or three times that or 3.2 times that. The problem was that each building was using a different multiplier. When our engineers asked while building the site what the multiplier rule was and what multiplier buildings usually use, we didn't actually know because every building was doing their own thing. The Mayor's Office of Housing is taking the opportunity to streamline rules like that and really create a more equitable process between buildings and the consistency that didn't exist before. This is happening in a lot of different business process levels. That's business processes. 

Use data to transform housing policy

The third and final layer is starting to get to this legislative level. We are starting to see a lot of data come through DAHLIA which I'll talk a little bit more about. The Mayor's Office of Housing can now understand where they should invest in housing, what they should build, if their two bedrooms are meeting the needs of the population more than three bedrooms and studios, if people like to live in certain neighborhoods and why that is, if housing is close to transit or other jobs. They can start to analyze what's coming through the site to better understand where they should invest their time and money and their programs.

User Demographics

Overview of the demographics of DAHLIA users.

Sethna: 

A little bit more just about the demographics we're seeing in DAHLIA. A lot of folks are applying, not surprisingly, from a lot of major Bay Area cities to live in San Francisco. We're seeing over 89 languages represented on the site, a really wide age range of 25 to 65, we have a lot of senior buildings on the site, we see seniors apply, and about 58% of users right now are accessing the site on mobile. It's really common for folks to access it on mobile as mobile is generally a primary access point for the internet for a lot of users. A lot of people access the site through their housing counselor or at a public computer or at their local non-profit. We're trying to understand who is seeing the site, where are they coming from, what's the need in San Francisco. The reason we can start to analyze this at a little bit more of a population level is because we really are getting a lot of data coming through. At this point, we're about at 101.4 million site visits and we can start to see the trend over years. What was happening in 2015, 2016? How do we see, as there are more listings coming online, different things that people gravitate towards? There's really this volume of data coming in at the Mayor's Office of Housing in a way that wasn't before. Barry's going to talk a little bit about what the Mayor's Office of Housing team is saying about this. 

Roeder:

Sometimes when we talk about this, I'd like to say, "And you thought we were talking about a service delivery product. We thought we were, too." One of my colleagues refers to it as this data exhaust. There's so much stuff that we get from Google Analytics and we use another service called Heap Analytics. One of the things I talked about earlier is that one of the government's roles is to have an eye on equity. So as Maria Benjamin, the director of our housing placement programs, says, "We finally have real data to help us identify and remove barriers to housing placement for specific segments of our population." 

Let's look at our housing funnel for a second. Up at the top level, we have now almost 1.5 million site visits. Say, of those 1.5 million site visits, eight percent are X demographic and we have 80,000+ applications. Wait a minute, only six percent of the applications are that demographic. This is interesting in the almost 2,500 units that we place to date through DAHLIA, it drops to 5%. What's going on? What's the problem? Is there an issue in the app that's dissuasive to people creating an application? Is there something in our process? That's something that obviously is a great concern to us and this gives us an opportunity to really look at what's happening with that. Even beyond that though, Roshen made reference to this. In a steering committee meeting about a year and a half ago, Kate Hartley who's our Executive Director said simply and eloquently and succinctly, "This will change what we build."

The way we used to think about that is, okay, we have some money to build some housing and there's a little bit of, "Well, what do you guys think? Ten one bedrooms because that's what we did last time? There was no source of data to understand what we should actually be building. Now we can say, "Wait, stop. If you're going to build a 50-unit building in the Mission, in that area, people of this household size, and that income, that race ethnicity with that operating system on their computer, are looking for that type of unit." Whoa. We've never had that before. We've been in regional conversations, smaller jurisdictions in the Bay Area, where someone from the City Council, maybe it's an affluent City Council member says, "Yeah, I was driving through and I saw a Latino family on the street and I thought we should build some more of those units like that." Well, that's a little bit ad hoc, a decision like that. It's really exciting to have this information that will help us answer the question about what kind of housing crisis we have, who particularly is struggling, and how we focus the relatively scarce resources that we have on this issue. 

Creating Feedback Loop

Sethna:

To go back to our diagram because we love diagrams. We've talked about where products are, we've talked about this feedback loop between agencies and the public the product is starting to create, but we're really talking about here is changing business rules. Policy is a larger feedback loop between policymakers and the public. Barry mentioned this earlier, but he leads a steering committee at the City. They check in on DAHLIA really regularly, and the steering committee is composed of a housing policy advisor for the Mayor, Mayor's Chief of Staff, a lot of folks that either influence or are involved with policy making in the City. They're able to start to have these touch points to ask what's coming through DAHLIA, what are we seeing? The City can even start to say we have this inclusionary housing policy. How is that working out for people? Are people actually accessing this housing? How is it working? Whose needs is it meeting? Whose is not meeting? This conversation that's being started at the steering committee in other places is really exciting to us because we get to use this really pretty simple digital product to start these policy level conversations at the City and County. If this kind of feedback loop process looks familiar to you, especially a lot of folks who use agile are in the startup community, that's probably because it is. 

Roeder: 

That's right. That's that diagram we pulled from the earlier slide. The idea of being able to make pivots on the fly in your development. We were talking about that in the context of building a product. Now we're talking about it in the context of establishing policy. Because we have those feedback loops and more frequent interaction between the public and policy makers, we can do a better job of aligning policy to the needs of constituents.

Sethna:

I think we're really particularly excited about this cycle of feedback loops and Barry's really passionate about the image of City Hall.

Roeder: 

I love the picture of City Hall because this is San Francisco City Hall for those of you that don't know. There's this beautiful, old, I'll say, legacy government building that now has this fancy digital LED lighting system. The idea is that we can take the strengths of a legacy system like we've discussed earlier: the service delivery infrastructure, democratic process, eye on equity. We can leverage that with this groundbreaking agile data-driven process that allows us to shorten innovation cycles. We can imagine one corner of this building started with CAT5 cables wrapped up in an old stone column. We're really excited about the possibilities that that methodology offers.

Sethna: 

We'll be around all week. We're really excited to chat with every one of you. Like Barry mentioned, we're actually talking to a lot of other jurisdictions about DAHLIA itself to implement it regionally and think through what a regional solution and a regional kind of database of affordable housing looks like first here in the Bay and then in other areas. Please come talk to us about that. Thanks!

May 23, 2018

Bridge SF City Spotlight | City of San Francisco & Exygy

Roshen Sethna, Exygy’s Civic Partner, and Barry Roeder, at the City of San Francisco, share the story of creating DAHLIA, San Francisco’s affordable housing portal in 2018.

Meet Roshen and Barry

Sethna: 

Hi, everybody. My name is Roshen Sethna and this is my colleague Barry Roeder. We are here to talk to you today about our latest wave of digital innovation in government, and we're here to talk to you about how it's kind of doing two things. One, it is reinventing service delivery between agencies and citizens and, two, it is really starting to help transform government as a system and that's really exciting to us. Barry and I have had the privilege of working on this digital product here in San Francisco with the Mayor's Office of Housing and Community Development where Barry works. We have seen how it's improved service delivery and we've also seen how it's starting to transform government especially at the policy level. 

I'm going to first introduce myself and give Barry a chance to introduce himself. Like I mentioned, my name is Roshen. I'm a Partner at Exygy. Exygy is a digital innovation studio based here in San Francisco in the Mission. We are focused on building healthy communities and I lead our Civic Tech practice. Exygy has a really deep Civic Tech practice even before I was at the firm. We've worked with several local agencies and regional agencies including the Mayor's Office of Housing but also local transit organizations like BART and MTC, educational organizations such as the School District and the Committee of Information Technology, and a variety of agencies at the City and County of San Francisco. Personally I was really excited to start working on this housing project a few years ago because my background in education is policy, but right out of school I was a housing counselor in Houston. This was right after the financial crisis so there were a lot of rules that were changing with regards to housing especially homeownership. One of my jobs was to help first-time homebuyers in particular areas navigate these new rules. It was pretty tough for people to stay up to date on what was available to them and what they were eligible for. My role as a counselor was really important but also involved a lot of manual research and looking things up and double checking things. I was really excited to work on a product in San Francisco that could help people access those resources on their own and navigate them in a really easy way. Barry's going to talk a little bit more about the product that we built and introduce himself. 

Roeder: 

Thank you. Good morning, everybody. My name is Barry Roeder, and it's true, I work at the Mayor's Office of Housing and Community Development in San Francisco. We are the agency that's responsible for building and overseeing all of the affordable housing in San Francisco. We have multiple portfolios of housing. My role there is Strategic Initiatives and I'm focused particularly on digital strategy for the department. I've worked with Roshen as she indicated on one of the projects that we'll talk about today, which of course is our DAHLIA housing portal. The Dahlia housing portal is a San Francisco affordable housing portal at housing.sfgov.org. It is a product that we built to create a single-stop resource for people that are looking for and wanting to apply for affordable housing that didn't exist before. We've been fortunate recently to win a couple of awards. One is the Spur Good Government award and the other is a Digital Leader award. We're getting some traction and we actually have some news. We're expanding the platform regionally. Alameda County has signed on recently. It'll be the first regional platform of its kind which is pretty exciting. We'll talk a little bit more about that as we go forward.

Government Solves Public Problems with Policy

Roeder: 

Before we go further, though, I thought it would be useful to talk a little bit about the role of government. Some of you might know, depending on the circles that you travel in, it's pretty easy for technologists to be dismissive of legacy systems. It's clunky, it's old, throw it out. Let's have a hackathon over the weekend and we'll just replace everything with an app. Well you can imagine, you can think about services that you get from your jurisdiction wherever you live. There are certain things like cleaning the streets or showing up to vote or certain things that couldn't be replaced by an app. There are some really great things actually about legacy systems. Some of them, in fact, are still best practice. First of all, it's a democratic system obviously with an eye on equity. There's a focus on all the constituencies, not just who has the money for whatever app. It's the way we're set up, it's how we do business as a democratic society and that's represented in our legacy systems. So again, some of these systems that when they were introduced were initially best practice, the problem with them is that innovation comes slowly. Things don't move that quickly in terms of how we improve or how we respond to current needs with the systems that we have. We'll talk a little bit more about how we can make some adjustments to that. 

Current Government System

Sethna: 

To dive a little bit deeper into the current system because Barry and I believe that you need to understand the context in which you're working so that you can involve the system itself. You guys are probably really familiar with this, but one of the biggest levers that government obviously has is policy and oftentimes policies made at the top, that might be housing policy, transit policy health care policy, is then pushed down to the agency or department level. There might be a housing authority, a healthcare department, a transit authority that is carrying out a lot of the policy that is set by legislators or policy makers. The way they often do that is through programs and services that then you and I interact with as members of the public. One thing we've started to see in the last 10, 15, 20 years as people have started to say, "How can we better serve citizens? How can agencies better deliver services to citizens through those programs and services that they're offering?" In the digital age we've started to see that it's often done with digital products, so this is where digital products tend to sit. Whether they're websites, mobile applications, informational applications, they sit at this level. They have really helped agencies better deliver services to the public. One of the really big examples of this that you guys are all familiar with is the healthcare.gov redesign. We've seen digital innovation work really well at this level, but we think that these products actually have an opportunity to change more than just service delivery. That's what we're going to chat a bit about today. 

What is DAHLIA?

Sethna: 

One product we wanted to talk to you guys about is a product that Barry and I have worked on which is called DAHLIA. DAHLIA, as Barry mentioned, is a one-stop shop for affordable housing listings in San Francisco and it also offers a space for citizens to apply for that housing. DAHLIA stands for the Database of Affordable Housing Listings Information and Application. There will be a quiz at the end of this and it's also the city flower which is cute. We were able to centralize resources in one place, but it took the last two and a half years to robustly build out this site. When we started the project in 2015 and Barry's team had started it really earlier than that. 

The Problem

Sethna: 

When Exygy came into it, we were trying to understand with the city what was the landscape we were facing, what's the actual problem that we're really trying to solve. The Mayor's Office of Housing contracted with Exygy as well as a couple different agencies to help solve this issue. What we first did was we understood the landscape. We at the time understood that the late Mayor Ed Lee had some goals around housing. He was slated to build 30,000 new units by 2020 and 10,000 of those are slated to be permanently affordable. We all know the City has a housing crisis and there were efforts being made to increase the actual housing stock. However as this housing stock is coming online, one of the problems was how do you find it, how do you find all these units and various buildings and navigate them. The answer to that at the time was not very easy. 


Previous housing sites

This is what people faced when they were searching for housing in 2015. There were several different government agencies sites. You might go to the healthcare department, the housing department. Also, oftentimes nonprofits created their own resources, so you could go to those lists. It was really tough for people to understand what information is most up to date. They were seeing inconsistencies between different sites and they were also having to navigate text heavy sites to understand if they were eligible for this, is their household size too big or too small, is their income level just right. They're trying to navigate a lot of resources and we did have a lot of what we call "power users" that were able to navigate this system and knew it really well. Let's say you were one of those "power users" and you did navigate the system. You were able to find a housing listing or a resource that you were going to apply to. What happened then? At that point, you were met with an arduous application process. At the time the application process was all on paper, so no digital application. You had to usually go in person to pick up the application because each building had a different application to make sure to get the right application for the right building. And once you did get your hands on it, these applications were really lengthy. 15+ pages often required a lot of supplementary information: your tax returns, your pay stubs, things like that. It was a really cumbersome process to get the application together. It might take several days to do that. Let's say you were able to do that. You took several days, you're able to get through the application process, compile all the information you need for that listing, and you're ready to turn it in. At this point you would go to the building generally to turn it in and you'd be met with long lines to submit applications, which was not a really great experience for housing applicants because oftentimes all of the applicants would need that full application window to put their application together. Many times people were waiting until the last minute to apply up until that 5 PM deadline on that application day. This is not only a bad experience for the housing applicant but a really stressful experience for the leasing agents and the building owners who are trying to manage getting everyone's application in by that due date. It's a really stressful process all around obviously if you're a parent, if you have a couple jobs, if you don't have the time to stand in line for several hours. People were often applying to several listings. This experience would happen every time you needed to submit an application. Let's say you were able to get through this step and you actually turned in your PayPal application to the leasing agent. What was that post application process like? I might let Barry speak to this because he's particularly passionate about the post-application process.

Roeder: 

Roshen knows when we talk about DAHLIA that one of my favorite things is right here: the carnival drop, which I think Twitter or Google or most tech companies don't have one of these. You turned in your application and we literally pulled out a roll - think about the toilet paper unfolding - of the red carnival tickets and you got one and one went in the drum. We had to tape the door of this drum shut because it flies open when we spin it and we had to pull the tickets out. It's an absurd process at least for this day and age. I want to give credit and thanks to the folks that came up with this whenever they did decades ago. I'm sure it was best practice at that time. It's not best practice now. Quick note about the fact that we do place our units through a lottery process in San Francisco, that varies by jurisdiction in the Bay Area. There's a combination of weightless and lottery. Both of those things are functional in DAHLIA.  

Sethna: 

Just to add for transparency purposes, let's say a thousand people applied to a listing. The Mayor's Office of Housing staff would literally have to pull every single ticket out of that drum. Because if I'm an applicant, I want to know that I got pulled even if I got pulled at number 900 or 950. So again, not only a stressful process for the applicant but stressful for the Mayor's Office of Housing staff who are really managing these lotteries for each of the buildings in San Francisco. Once the applications were actually in and the lottery was run, oftentimes these lotteries were just posted and this part actually was digital. They were posted online and people could see their ranking in the lottery system but there's not a really great way after that to know when you're going to get a listing. That's kind of the the different problems we are facing both in the housing landscape and the problems that the citizens were facing that were applying. 

The Solution

Browse DAHLIA at housing.sfgov.org

Sethna: 

Our solution was DAHLIA. You can access DAHLIA at housing.sfgov.org. I can talk at length about the different features that we have built out over this site because we at this point have been working on it for over two and a half years. One of the ethos that we had when we built this site was to offer affordable housing seekers the same level of experience and dignity that market rate housing seekers would get when they're searching. An experience that you may be really used to on a Trulia or a Zillow. We wanted to offer that same level of experience. One of the things we did throughout the last two and a half years is every few weeks we user tested with users. Actual housing applicants past, present, future. We put mockups or actual code in front of them, ask them to do a certain task, see how they got through that task on the site. One of our early user testers was looking at the screen in front of her and said, "Am I on the right site? I'm not sure if this is the right site." We're like, "Yeah, that's right. This is DAHLIA. We'd like you to see if you can find a listing you're eligible for." She said, "I don't think this is the site I'm supposed to go to. This looks too nice. I don't think this is affordable housing." This is something we were also trying to do with DAHLIA: rebuild trust between government and citizens. Everyone in this room has stood in line at the DMV at some point for six hours. We're not used to a certain level of experience when we interact with government, so that was part of engaging users in this process. To have the government agencies say, "Yup, this is for you. These are resources that we want to serve you with." 

A couple of things that we did on this site. One, we streamlined the actual housing listing to the right. We made it a really standardized format so now all of the listings have consistent data and information. We also turned it into a storytelling format from the text heavy sites to the new product. We surface information people cared about the most. People told us during user testing, "Just tell me the minimum income I need to get into this unit and tell me what the rent is. Tell me those two pieces of information and then give me more information if I want to dig in." We designed the information hierarchy of this page very intentionally based on what people were telling us they needed when they're browsing. We made it a storytelling format. The second thing we did was we standardized and digitized the application process. So that's on the mobile screen. We turned it into a step-by-step really easy to read application process. We call it our short form application process because one of the things we did is that we talked with a lot of leasing agents and housing developers. We asked them, "Hey, at this point people are really just applying to be in a lottery to have a chance to get a unit. What is the basic amount of information you need from people? Why are you asking them for their tax returns? Do you really need that to run a lottery?" We worked with them to understand what the minimum fields were and we're able to reduce what we were asking of the user to make it comparable to what they're applying for. If I'm spending a few minutes applying and my chances are maybe not so great because the lottery's big that time, at least I'm only spending a few minutes applying instead of spending several days. So again, we are treating the user with this level of dignity. 

The site is also translated into the city's main languages. It's accessible so folks that need to use screen readers, for example, can read through it. This product set up a foundation as these listings come online, they can get inserted into this standardized format that we're using now. 


DAHLIA impact by the numbers

‍A couple of statistics on how the site's been doing. It now takes 15 minutes to apply online. It's shorter if you even use one of your past applications to pre-fill a new application. That took that several day process and reduced it down to 15 minutes. When the online application feature went live which was in November of 2016, we saw that about 90% were using online applications. We were actually able to increase that 85% to 90% by doing several things like translating applications into different languages and things like that. We're trying to move folks online because that also helps the leasing agents with paper data entry. It reduces the amount of paper. We've seen 80,000 since November of 2016 and about 130 listings have gone through the site. Something that we are super proud of and we are excited to see grow is that about 100 households have been placed in units per month. Of course as listings come online and as larger listings come online, we're seeing this grow and we're really excited to be able to track how many folks we're placing over time. We've had a lot of success with this product on the service delivery side and the ability to measure how many people were placing. But, it wasn't just dropping design and code in people's laps. We weren't saying, "Here's a site. Great. Use it. Moving on." There were a lot of things that happened in the background like change management processes at the City that have made this product successful. 

Our Learnings

Sethna: 

One of the things we wanted to share with you guys is some of the learnings from what has made DAHLIA effective so that at your agencies and organizations, if you're embarking on building a product, you can learn from the stuff we learned.

Roeder: 

That's right. I think this is the point where I say, "Don't try this at home." The Mayor's Office of Housing and Community Development tried this at home a couple times before this and it wasn't successful. There's a lot that we've learned along the way that we thought we wanted to share with you about key elements in creating a successful public sector service delivery product. The first of them is the idea Roshen talked about asking the user. 

There's a story that I love to tell. Jay was talking earlier about and of course this conference today is about building bridges between the different sectors. We were fortunate through the Mayor's Office of Civic Innovation at the time to be the recipient of one of the first civic bridge projects that placed some very sharp Googlers with the Mayor's Office of Housing and Community Development for four months to help us look at the problem and how we would approach this. One of the things that we asked them I remember is, "Wow, not only are these Googlers but these are like really smart Googlers. So what's the secret of how you guys ask the user? What's the Google secret?" 

Lesson 1: Ask the user

The idea is to sit regularly with people that actually have to use this. It doesn't mean let's have a community conversation one afternoon in a conference room and think about what are you looking for, or let's build something and then send out a survey monkey and ask if they like it. On a scale of one to ten. No instead, let's sit with people all the way through and talk about what their needs are. There's a whole ecosystem around this project. There are housing developers, leasing agents, property managers on one side and housing counselors and other housing advocates on the other side. They've all got valuable things to bring to this, so having them at the table is really important. 

Lesson 2: Start small, test, and iterate

The next piece is something that probably a lot of you in the room are familiar with but the idea of developing this in an Agile fashion. The idea of Agile; the story I like to tell about that is that for those of you that went to college and are of a certain age, you probably turned in a binder for your semester project and it had the tabs in all the right spots and everything was perfectly tight. You didn't dare turn that thing in without everything perfect. The problem with that though is that you did that in your dorm room or whatever in a vacuum. You were creating that without input from the people that had to buy it because your professor or whoever. You didn't have to start some conversation about, "Wait a minute. Should I make some adjustments here?" Think instead about taking that binder and take the two or three most salient pages out of the middle, rip them out, sit down with people that actually are going to buy your product, make sure that those are right, stand it on its feet as a proof of concept. Or in agile speak, they call MVP, the minimum viable product, and then iterate from there. That was something that was really helpful to us here. Here it shows you what a typical agile cycle would be like. This actually could be a cycle for any project except that instead of doing this once over the course of 18 months or two years, which some public sector IT procurement processes look like, and then they come back with their 60 million dollar product that doesn't work, they do this over two years and then do it again and then do it again and then do it again. 

Lesson 3: Build the right team

Of course fundamental to any of this is to make sure that you build by having a good, solid team. You want to make sure that you get your change agents. Of course that's critical to any process, especially if you're endeavoring to do something differently. Fostering a collaborative culture and secure air cover. I'll give an example that's related to what we talked about a second ago with regard to agile. Imagine going to the controller's office who's used to sending out your request for proposal for a certain amount with specs. You're saying, "Well, we don't totally know our specs yet and we don't totally know what the total project cost is going to be. We're going to just do it every couple weeks. We'll look at how much we spend." That's an interesting conversation to have, but what our CFO and our department says is really great. Benjamin Klosty makes the case, "Yeah, but you know what? Every two weeks you're getting something. You're getting delivered something. And at any point if you don't like what you're not getting, you shut it off. Consult best practices. We talked about the benefit that we received from our working with Google and I can't underscore the idea that if there's something you want to do, look around and see what's been done, of course. 

Lesson 4: Find the right partners

We've been very fortunate in our work with Exygy to find someone who cares about the problem. It makes a big difference to have somebody, particularly in an agile process, that's thinking with the same mindset about how to solve this problem because none of us know what the outside is going to look like. Before then, though, is the idea of building modularly. We talked about that and I mentioned that the procurement process has to change. Maintaining open communication: What I want to say about that is that rather than having a kickoff meeting, meeting again in 18 months, and seeing how it turns out, we have meetings - we're probably in six or seven meetings a week together in one fashion or another. Whether it's the design or engineering or project management piece, it's a constant communication partnership working on the problem.

Sethna:

On the procurement side, it's important to be able to set up procurement so that you can work in an agile format. Typically a lot of public sector agencies will work by sending out a large spec sheet to the vendor and saying, "Hey, come back in a year. Show us a product. Cool." At that time, user needs have changed, business rules have changed, and the product doesn't meet the need anymore. If you're able to set up a procurement process like Barry and Benjamin and others were at MOCD in a way that's saying, "Hey, here's some hypothesis of what we want to build because we've talked to users and done some discovery work, but we really want you to continue testing, test some more, and continue to build it out." This really helps. Barry's team was able to release this product in 90 days. Three months from us writing the first line of code, we released something and all it really was was a site that had one type of rental listing listed. Nothing more - no online application, no translation in other languages, no accounts. We started to see how people interacted with the site and then started to layer on features and we were able to course correct because of that. Then MOCD in that scenario, is able to fully utilize Exygy's skills in design, engineering, and product as opposed to dictating what has already been decided to build. You're not getting the bang for your buck from the public sector side to do that. 


System diagram

Roeder:

In summary, ask the user. Can't emphasize that enough. Start small, test, and iterate, build the right team, and find the right partners. Let's check in again with the diagram that Roshen introduced to us and see where we are. We talked about introducing products between agencies like MOCD and our constituents, the public, to improve service delivery primarily. But what we found though in this process of sitting with people and listening to their needs and not just applicants but all our partners, is that we were building a feedback loop. We were building new communication with this constituency that helps us do our job better overall even outside of the context of the product itself. 

Using Data To Transform Policy

Sethna: To dig a little bit deeper into that, this feedback loop story is really around data. The data that we've seen has started to help us improve services but it's also starting to help improve policy. By policy, I mean policy at a variety of levels. I'm going to give a couple of stories around how DAHLIA did this. 

Use data to increase program effectiveness

One story is at the program effectiveness level. Like Barry mentioned, the City of San Francisco runs their housing on a lottery system. The leasing agent will run a lottery and then use those ranked lottery numbers to lease up the building. In that lottery process, we have something called preferences. It's basically certain citizens who get preferential treatment in the lottery because they might have been evicted from a certain neighborhood at a time or currently live in the same neighborhood as the listing. We want to keep them in that neighborhood and retain them. There's different preferences for people in the housing lottery process. One thing we were seeing is that people were really not taking advantage of these preferences that actually gave them a huge leg up in the lottery, and one preference they were not taking advantage of is called the Neighborhood Resident House preference. This is the preference where if you live in the same neighborhood as a listing, we'll give you preferential treatment to be able to retain you in that neighborhood. Only about 40 folks were actually claiming this preference. What we did was we simply redesigned this page. We added a banner at the top and said, "Hey, you qualify for this thing and this is what it is. We explained it to them because nobody understands what Neighborhood Resident Housing preference means. We actually changed some of the content. We worked with Barry's office to say, "Hey, can we just call it Live in the Neighborhood preference? Is that okay?" They said, "Yeah, that's okay as long as somewhere else on the page we put Neighborhood Resident Browsing preference because we have to have that on page." We worked within the business rules of the organization to make things more user friendly and we saw the claim rate of this preference jump to about 88%, more than double the claim rate. Really simple design and code changes make it so you're able to see if people are claiming services that they are eligible for, and if they're not, how can you drive them to those services and programs and encourage them to claim it and describe it to them in a more human way. 

Use data to improve business processes

Another example is a little bit about the programmatic level - going to the organizational level. Barry's colleagues at the Mayor's Office of Housing are currently as we speak rewriting the housing policy manual that the Mayor's Office of Housing uses. It's a lot of business processes around how buildings are listed, what the building needs to do to report to the Mayor's Office of Housing -  all the business rules around managing the housing process. A lot of conversations in that rewrite came from the product. One really small example of this is that on a listing the minimum income you need to get into that listing is generally a multiplier of rent. If rent is 500 bucks a month, then the minimum income requirement might be two times that or three times that or 3.2 times that. The problem was that each building was using a different multiplier. When our engineers asked while building the site what the multiplier rule was and what multiplier buildings usually use, we didn't actually know because every building was doing their own thing. The Mayor's Office of Housing is taking the opportunity to streamline rules like that and really create a more equitable process between buildings and the consistency that didn't exist before. This is happening in a lot of different business process levels. That's business processes. 

Use data to transform housing policy

The third and final layer is starting to get to this legislative level. We are starting to see a lot of data come through DAHLIA which I'll talk a little bit more about. The Mayor's Office of Housing can now understand where they should invest in housing, what they should build, if their two bedrooms are meeting the needs of the population more than three bedrooms and studios, if people like to live in certain neighborhoods and why that is, if housing is close to transit or other jobs. They can start to analyze what's coming through the site to better understand where they should invest their time and money and their programs.

User Demographics

Overview of the demographics of DAHLIA users.

Sethna: 

A little bit more just about the demographics we're seeing in DAHLIA. A lot of folks are applying, not surprisingly, from a lot of major Bay Area cities to live in San Francisco. We're seeing over 89 languages represented on the site, a really wide age range of 25 to 65, we have a lot of senior buildings on the site, we see seniors apply, and about 58% of users right now are accessing the site on mobile. It's really common for folks to access it on mobile as mobile is generally a primary access point for the internet for a lot of users. A lot of people access the site through their housing counselor or at a public computer or at their local non-profit. We're trying to understand who is seeing the site, where are they coming from, what's the need in San Francisco. The reason we can start to analyze this at a little bit more of a population level is because we really are getting a lot of data coming through. At this point, we're about at 101.4 million site visits and we can start to see the trend over years. What was happening in 2015, 2016? How do we see, as there are more listings coming online, different things that people gravitate towards? There's really this volume of data coming in at the Mayor's Office of Housing in a way that wasn't before. Barry's going to talk a little bit about what the Mayor's Office of Housing team is saying about this. 

Roeder:

Sometimes when we talk about this, I'd like to say, "And you thought we were talking about a service delivery product. We thought we were, too." One of my colleagues refers to it as this data exhaust. There's so much stuff that we get from Google Analytics and we use another service called Heap Analytics. One of the things I talked about earlier is that one of the government's roles is to have an eye on equity. So as Maria Benjamin, the director of our housing placement programs, says, "We finally have real data to help us identify and remove barriers to housing placement for specific segments of our population." 

Let's look at our housing funnel for a second. Up at the top level, we have now almost 1.5 million site visits. Say, of those 1.5 million site visits, eight percent are X demographic and we have 80,000+ applications. Wait a minute, only six percent of the applications are that demographic. This is interesting in the almost 2,500 units that we place to date through DAHLIA, it drops to 5%. What's going on? What's the problem? Is there an issue in the app that's dissuasive to people creating an application? Is there something in our process? That's something that obviously is a great concern to us and this gives us an opportunity to really look at what's happening with that. Even beyond that though, Roshen made reference to this. In a steering committee meeting about a year and a half ago, Kate Hartley who's our Executive Director said simply and eloquently and succinctly, "This will change what we build."

The way we used to think about that is, okay, we have some money to build some housing and there's a little bit of, "Well, what do you guys think? Ten one bedrooms because that's what we did last time? There was no source of data to understand what we should actually be building. Now we can say, "Wait, stop. If you're going to build a 50-unit building in the Mission, in that area, people of this household size, and that income, that race ethnicity with that operating system on their computer, are looking for that type of unit." Whoa. We've never had that before. We've been in regional conversations, smaller jurisdictions in the Bay Area, where someone from the City Council, maybe it's an affluent City Council member says, "Yeah, I was driving through and I saw a Latino family on the street and I thought we should build some more of those units like that." Well, that's a little bit ad hoc, a decision like that. It's really exciting to have this information that will help us answer the question about what kind of housing crisis we have, who particularly is struggling, and how we focus the relatively scarce resources that we have on this issue. 

Creating Feedback Loop

Sethna:

To go back to our diagram because we love diagrams. We've talked about where products are, we've talked about this feedback loop between agencies and the public the product is starting to create, but we're really talking about here is changing business rules. Policy is a larger feedback loop between policymakers and the public. Barry mentioned this earlier, but he leads a steering committee at the City. They check in on DAHLIA really regularly, and the steering committee is composed of a housing policy advisor for the Mayor, Mayor's Chief of Staff, a lot of folks that either influence or are involved with policy making in the City. They're able to start to have these touch points to ask what's coming through DAHLIA, what are we seeing? The City can even start to say we have this inclusionary housing policy. How is that working out for people? Are people actually accessing this housing? How is it working? Whose needs is it meeting? Whose is not meeting? This conversation that's being started at the steering committee in other places is really exciting to us because we get to use this really pretty simple digital product to start these policy level conversations at the City and County. If this kind of feedback loop process looks familiar to you, especially a lot of folks who use agile are in the startup community, that's probably because it is. 

Roeder: 

That's right. That's that diagram we pulled from the earlier slide. The idea of being able to make pivots on the fly in your development. We were talking about that in the context of building a product. Now we're talking about it in the context of establishing policy. Because we have those feedback loops and more frequent interaction between the public and policy makers, we can do a better job of aligning policy to the needs of constituents.

Sethna:

I think we're really particularly excited about this cycle of feedback loops and Barry's really passionate about the image of City Hall.

Roeder: 

I love the picture of City Hall because this is San Francisco City Hall for those of you that don't know. There's this beautiful, old, I'll say, legacy government building that now has this fancy digital LED lighting system. The idea is that we can take the strengths of a legacy system like we've discussed earlier: the service delivery infrastructure, democratic process, eye on equity. We can leverage that with this groundbreaking agile data-driven process that allows us to shorten innovation cycles. We can imagine one corner of this building started with CAT5 cables wrapped up in an old stone column. We're really excited about the possibilities that that methodology offers.

Sethna: 

We'll be around all week. We're really excited to chat with every one of you. Like Barry mentioned, we're actually talking to a lot of other jurisdictions about DAHLIA itself to implement it regionally and think through what a regional solution and a regional kind of database of affordable housing looks like first here in the Bay and then in other areas. Please come talk to us about that. Thanks!

May 23, 2018

Bridge SF City Spotlight | City of San Francisco & Exygy

Roshen Sethna, Exygy’s Civic Partner, and Barry Roeder, at the City of San Francisco, share the story of creating DAHLIA, San Francisco’s affordable housing portal in 2018.

Meet Roshen and Barry

Sethna: 

Hi, everybody. My name is Roshen Sethna and this is my colleague Barry Roeder. We are here to talk to you today about our latest wave of digital innovation in government, and we're here to talk to you about how it's kind of doing two things. One, it is reinventing service delivery between agencies and citizens and, two, it is really starting to help transform government as a system and that's really exciting to us. Barry and I have had the privilege of working on this digital product here in San Francisco with the Mayor's Office of Housing and Community Development where Barry works. We have seen how it's improved service delivery and we've also seen how it's starting to transform government especially at the policy level. 

I'm going to first introduce myself and give Barry a chance to introduce himself. Like I mentioned, my name is Roshen. I'm a Partner at Exygy. Exygy is a digital innovation studio based here in San Francisco in the Mission. We are focused on building healthy communities and I lead our Civic Tech practice. Exygy has a really deep Civic Tech practice even before I was at the firm. We've worked with several local agencies and regional agencies including the Mayor's Office of Housing but also local transit organizations like BART and MTC, educational organizations such as the School District and the Committee of Information Technology, and a variety of agencies at the City and County of San Francisco. Personally I was really excited to start working on this housing project a few years ago because my background in education is policy, but right out of school I was a housing counselor in Houston. This was right after the financial crisis so there were a lot of rules that were changing with regards to housing especially homeownership. One of my jobs was to help first-time homebuyers in particular areas navigate these new rules. It was pretty tough for people to stay up to date on what was available to them and what they were eligible for. My role as a counselor was really important but also involved a lot of manual research and looking things up and double checking things. I was really excited to work on a product in San Francisco that could help people access those resources on their own and navigate them in a really easy way. Barry's going to talk a little bit more about the product that we built and introduce himself. 

Roeder: 

Thank you. Good morning, everybody. My name is Barry Roeder, and it's true, I work at the Mayor's Office of Housing and Community Development in San Francisco. We are the agency that's responsible for building and overseeing all of the affordable housing in San Francisco. We have multiple portfolios of housing. My role there is Strategic Initiatives and I'm focused particularly on digital strategy for the department. I've worked with Roshen as she indicated on one of the projects that we'll talk about today, which of course is our DAHLIA housing portal. The Dahlia housing portal is a San Francisco affordable housing portal at housing.sfgov.org. It is a product that we built to create a single-stop resource for people that are looking for and wanting to apply for affordable housing that didn't exist before. We've been fortunate recently to win a couple of awards. One is the Spur Good Government award and the other is a Digital Leader award. We're getting some traction and we actually have some news. We're expanding the platform regionally. Alameda County has signed on recently. It'll be the first regional platform of its kind which is pretty exciting. We'll talk a little bit more about that as we go forward.

Government Solves Public Problems with Policy

Roeder: 

Before we go further, though, I thought it would be useful to talk a little bit about the role of government. Some of you might know, depending on the circles that you travel in, it's pretty easy for technologists to be dismissive of legacy systems. It's clunky, it's old, throw it out. Let's have a hackathon over the weekend and we'll just replace everything with an app. Well you can imagine, you can think about services that you get from your jurisdiction wherever you live. There are certain things like cleaning the streets or showing up to vote or certain things that couldn't be replaced by an app. There are some really great things actually about legacy systems. Some of them, in fact, are still best practice. First of all, it's a democratic system obviously with an eye on equity. There's a focus on all the constituencies, not just who has the money for whatever app. It's the way we're set up, it's how we do business as a democratic society and that's represented in our legacy systems. So again, some of these systems that when they were introduced were initially best practice, the problem with them is that innovation comes slowly. Things don't move that quickly in terms of how we improve or how we respond to current needs with the systems that we have. We'll talk a little bit more about how we can make some adjustments to that. 

Current Government System

Sethna: 

To dive a little bit deeper into the current system because Barry and I believe that you need to understand the context in which you're working so that you can involve the system itself. You guys are probably really familiar with this, but one of the biggest levers that government obviously has is policy and oftentimes policies made at the top, that might be housing policy, transit policy health care policy, is then pushed down to the agency or department level. There might be a housing authority, a healthcare department, a transit authority that is carrying out a lot of the policy that is set by legislators or policy makers. The way they often do that is through programs and services that then you and I interact with as members of the public. One thing we've started to see in the last 10, 15, 20 years as people have started to say, "How can we better serve citizens? How can agencies better deliver services to citizens through those programs and services that they're offering?" In the digital age we've started to see that it's often done with digital products, so this is where digital products tend to sit. Whether they're websites, mobile applications, informational applications, they sit at this level. They have really helped agencies better deliver services to the public. One of the really big examples of this that you guys are all familiar with is the healthcare.gov redesign. We've seen digital innovation work really well at this level, but we think that these products actually have an opportunity to change more than just service delivery. That's what we're going to chat a bit about today. 

What is DAHLIA?

Sethna: 

One product we wanted to talk to you guys about is a product that Barry and I have worked on which is called DAHLIA. DAHLIA, as Barry mentioned, is a one-stop shop for affordable housing listings in San Francisco and it also offers a space for citizens to apply for that housing. DAHLIA stands for the Database of Affordable Housing Listings Information and Application. There will be a quiz at the end of this and it's also the city flower which is cute. We were able to centralize resources in one place, but it took the last two and a half years to robustly build out this site. When we started the project in 2015 and Barry's team had started it really earlier than that. 

The Problem

Sethna: 

When Exygy came into it, we were trying to understand with the city what was the landscape we were facing, what's the actual problem that we're really trying to solve. The Mayor's Office of Housing contracted with Exygy as well as a couple different agencies to help solve this issue. What we first did was we understood the landscape. We at the time understood that the late Mayor Ed Lee had some goals around housing. He was slated to build 30,000 new units by 2020 and 10,000 of those are slated to be permanently affordable. We all know the City has a housing crisis and there were efforts being made to increase the actual housing stock. However as this housing stock is coming online, one of the problems was how do you find it, how do you find all these units and various buildings and navigate them. The answer to that at the time was not very easy. 


Previous housing sites

This is what people faced when they were searching for housing in 2015. There were several different government agencies sites. You might go to the healthcare department, the housing department. Also, oftentimes nonprofits created their own resources, so you could go to those lists. It was really tough for people to understand what information is most up to date. They were seeing inconsistencies between different sites and they were also having to navigate text heavy sites to understand if they were eligible for this, is their household size too big or too small, is their income level just right. They're trying to navigate a lot of resources and we did have a lot of what we call "power users" that were able to navigate this system and knew it really well. Let's say you were one of those "power users" and you did navigate the system. You were able to find a housing listing or a resource that you were going to apply to. What happened then? At that point, you were met with an arduous application process. At the time the application process was all on paper, so no digital application. You had to usually go in person to pick up the application because each building had a different application to make sure to get the right application for the right building. And once you did get your hands on it, these applications were really lengthy. 15+ pages often required a lot of supplementary information: your tax returns, your pay stubs, things like that. It was a really cumbersome process to get the application together. It might take several days to do that. Let's say you were able to do that. You took several days, you're able to get through the application process, compile all the information you need for that listing, and you're ready to turn it in. At this point you would go to the building generally to turn it in and you'd be met with long lines to submit applications, which was not a really great experience for housing applicants because oftentimes all of the applicants would need that full application window to put their application together. Many times people were waiting until the last minute to apply up until that 5 PM deadline on that application day. This is not only a bad experience for the housing applicant but a really stressful experience for the leasing agents and the building owners who are trying to manage getting everyone's application in by that due date. It's a really stressful process all around obviously if you're a parent, if you have a couple jobs, if you don't have the time to stand in line for several hours. People were often applying to several listings. This experience would happen every time you needed to submit an application. Let's say you were able to get through this step and you actually turned in your PayPal application to the leasing agent. What was that post application process like? I might let Barry speak to this because he's particularly passionate about the post-application process.

Roeder: 

Roshen knows when we talk about DAHLIA that one of my favorite things is right here: the carnival drop, which I think Twitter or Google or most tech companies don't have one of these. You turned in your application and we literally pulled out a roll - think about the toilet paper unfolding - of the red carnival tickets and you got one and one went in the drum. We had to tape the door of this drum shut because it flies open when we spin it and we had to pull the tickets out. It's an absurd process at least for this day and age. I want to give credit and thanks to the folks that came up with this whenever they did decades ago. I'm sure it was best practice at that time. It's not best practice now. Quick note about the fact that we do place our units through a lottery process in San Francisco, that varies by jurisdiction in the Bay Area. There's a combination of weightless and lottery. Both of those things are functional in DAHLIA.  

Sethna: 

Just to add for transparency purposes, let's say a thousand people applied to a listing. The Mayor's Office of Housing staff would literally have to pull every single ticket out of that drum. Because if I'm an applicant, I want to know that I got pulled even if I got pulled at number 900 or 950. So again, not only a stressful process for the applicant but stressful for the Mayor's Office of Housing staff who are really managing these lotteries for each of the buildings in San Francisco. Once the applications were actually in and the lottery was run, oftentimes these lotteries were just posted and this part actually was digital. They were posted online and people could see their ranking in the lottery system but there's not a really great way after that to know when you're going to get a listing. That's kind of the the different problems we are facing both in the housing landscape and the problems that the citizens were facing that were applying. 

The Solution

Browse DAHLIA at housing.sfgov.org

Sethna: 

Our solution was DAHLIA. You can access DAHLIA at housing.sfgov.org. I can talk at length about the different features that we have built out over this site because we at this point have been working on it for over two and a half years. One of the ethos that we had when we built this site was to offer affordable housing seekers the same level of experience and dignity that market rate housing seekers would get when they're searching. An experience that you may be really used to on a Trulia or a Zillow. We wanted to offer that same level of experience. One of the things we did throughout the last two and a half years is every few weeks we user tested with users. Actual housing applicants past, present, future. We put mockups or actual code in front of them, ask them to do a certain task, see how they got through that task on the site. One of our early user testers was looking at the screen in front of her and said, "Am I on the right site? I'm not sure if this is the right site." We're like, "Yeah, that's right. This is DAHLIA. We'd like you to see if you can find a listing you're eligible for." She said, "I don't think this is the site I'm supposed to go to. This looks too nice. I don't think this is affordable housing." This is something we were also trying to do with DAHLIA: rebuild trust between government and citizens. Everyone in this room has stood in line at the DMV at some point for six hours. We're not used to a certain level of experience when we interact with government, so that was part of engaging users in this process. To have the government agencies say, "Yup, this is for you. These are resources that we want to serve you with." 

A couple of things that we did on this site. One, we streamlined the actual housing listing to the right. We made it a really standardized format so now all of the listings have consistent data and information. We also turned it into a storytelling format from the text heavy sites to the new product. We surface information people cared about the most. People told us during user testing, "Just tell me the minimum income I need to get into this unit and tell me what the rent is. Tell me those two pieces of information and then give me more information if I want to dig in." We designed the information hierarchy of this page very intentionally based on what people were telling us they needed when they're browsing. We made it a storytelling format. The second thing we did was we standardized and digitized the application process. So that's on the mobile screen. We turned it into a step-by-step really easy to read application process. We call it our short form application process because one of the things we did is that we talked with a lot of leasing agents and housing developers. We asked them, "Hey, at this point people are really just applying to be in a lottery to have a chance to get a unit. What is the basic amount of information you need from people? Why are you asking them for their tax returns? Do you really need that to run a lottery?" We worked with them to understand what the minimum fields were and we're able to reduce what we were asking of the user to make it comparable to what they're applying for. If I'm spending a few minutes applying and my chances are maybe not so great because the lottery's big that time, at least I'm only spending a few minutes applying instead of spending several days. So again, we are treating the user with this level of dignity. 

The site is also translated into the city's main languages. It's accessible so folks that need to use screen readers, for example, can read through it. This product set up a foundation as these listings come online, they can get inserted into this standardized format that we're using now. 


DAHLIA impact by the numbers

‍A couple of statistics on how the site's been doing. It now takes 15 minutes to apply online. It's shorter if you even use one of your past applications to pre-fill a new application. That took that several day process and reduced it down to 15 minutes. When the online application feature went live which was in November of 2016, we saw that about 90% were using online applications. We were actually able to increase that 85% to 90% by doing several things like translating applications into different languages and things like that. We're trying to move folks online because that also helps the leasing agents with paper data entry. It reduces the amount of paper. We've seen 80,000 since November of 2016 and about 130 listings have gone through the site. Something that we are super proud of and we are excited to see grow is that about 100 households have been placed in units per month. Of course as listings come online and as larger listings come online, we're seeing this grow and we're really excited to be able to track how many folks we're placing over time. We've had a lot of success with this product on the service delivery side and the ability to measure how many people were placing. But, it wasn't just dropping design and code in people's laps. We weren't saying, "Here's a site. Great. Use it. Moving on." There were a lot of things that happened in the background like change management processes at the City that have made this product successful. 

Our Learnings

Sethna: 

One of the things we wanted to share with you guys is some of the learnings from what has made DAHLIA effective so that at your agencies and organizations, if you're embarking on building a product, you can learn from the stuff we learned.

Roeder: 

That's right. I think this is the point where I say, "Don't try this at home." The Mayor's Office of Housing and Community Development tried this at home a couple times before this and it wasn't successful. There's a lot that we've learned along the way that we thought we wanted to share with you about key elements in creating a successful public sector service delivery product. The first of them is the idea Roshen talked about asking the user. 

There's a story that I love to tell. Jay was talking earlier about and of course this conference today is about building bridges between the different sectors. We were fortunate through the Mayor's Office of Civic Innovation at the time to be the recipient of one of the first civic bridge projects that placed some very sharp Googlers with the Mayor's Office of Housing and Community Development for four months to help us look at the problem and how we would approach this. One of the things that we asked them I remember is, "Wow, not only are these Googlers but these are like really smart Googlers. So what's the secret of how you guys ask the user? What's the Google secret?" 

Lesson 1: Ask the user

The idea is to sit regularly with people that actually have to use this. It doesn't mean let's have a community conversation one afternoon in a conference room and think about what are you looking for, or let's build something and then send out a survey monkey and ask if they like it. On a scale of one to ten. No instead, let's sit with people all the way through and talk about what their needs are. There's a whole ecosystem around this project. There are housing developers, leasing agents, property managers on one side and housing counselors and other housing advocates on the other side. They've all got valuable things to bring to this, so having them at the table is really important. 

Lesson 2: Start small, test, and iterate

The next piece is something that probably a lot of you in the room are familiar with but the idea of developing this in an Agile fashion. The idea of Agile; the story I like to tell about that is that for those of you that went to college and are of a certain age, you probably turned in a binder for your semester project and it had the tabs in all the right spots and everything was perfectly tight. You didn't dare turn that thing in without everything perfect. The problem with that though is that you did that in your dorm room or whatever in a vacuum. You were creating that without input from the people that had to buy it because your professor or whoever. You didn't have to start some conversation about, "Wait a minute. Should I make some adjustments here?" Think instead about taking that binder and take the two or three most salient pages out of the middle, rip them out, sit down with people that actually are going to buy your product, make sure that those are right, stand it on its feet as a proof of concept. Or in agile speak, they call MVP, the minimum viable product, and then iterate from there. That was something that was really helpful to us here. Here it shows you what a typical agile cycle would be like. This actually could be a cycle for any project except that instead of doing this once over the course of 18 months or two years, which some public sector IT procurement processes look like, and then they come back with their 60 million dollar product that doesn't work, they do this over two years and then do it again and then do it again and then do it again. 

Lesson 3: Build the right team

Of course fundamental to any of this is to make sure that you build by having a good, solid team. You want to make sure that you get your change agents. Of course that's critical to any process, especially if you're endeavoring to do something differently. Fostering a collaborative culture and secure air cover. I'll give an example that's related to what we talked about a second ago with regard to agile. Imagine going to the controller's office who's used to sending out your request for proposal for a certain amount with specs. You're saying, "Well, we don't totally know our specs yet and we don't totally know what the total project cost is going to be. We're going to just do it every couple weeks. We'll look at how much we spend." That's an interesting conversation to have, but what our CFO and our department says is really great. Benjamin Klosty makes the case, "Yeah, but you know what? Every two weeks you're getting something. You're getting delivered something. And at any point if you don't like what you're not getting, you shut it off. Consult best practices. We talked about the benefit that we received from our working with Google and I can't underscore the idea that if there's something you want to do, look around and see what's been done, of course. 

Lesson 4: Find the right partners

We've been very fortunate in our work with Exygy to find someone who cares about the problem. It makes a big difference to have somebody, particularly in an agile process, that's thinking with the same mindset about how to solve this problem because none of us know what the outside is going to look like. Before then, though, is the idea of building modularly. We talked about that and I mentioned that the procurement process has to change. Maintaining open communication: What I want to say about that is that rather than having a kickoff meeting, meeting again in 18 months, and seeing how it turns out, we have meetings - we're probably in six or seven meetings a week together in one fashion or another. Whether it's the design or engineering or project management piece, it's a constant communication partnership working on the problem.

Sethna:

On the procurement side, it's important to be able to set up procurement so that you can work in an agile format. Typically a lot of public sector agencies will work by sending out a large spec sheet to the vendor and saying, "Hey, come back in a year. Show us a product. Cool." At that time, user needs have changed, business rules have changed, and the product doesn't meet the need anymore. If you're able to set up a procurement process like Barry and Benjamin and others were at MOCD in a way that's saying, "Hey, here's some hypothesis of what we want to build because we've talked to users and done some discovery work, but we really want you to continue testing, test some more, and continue to build it out." This really helps. Barry's team was able to release this product in 90 days. Three months from us writing the first line of code, we released something and all it really was was a site that had one type of rental listing listed. Nothing more - no online application, no translation in other languages, no accounts. We started to see how people interacted with the site and then started to layer on features and we were able to course correct because of that. Then MOCD in that scenario, is able to fully utilize Exygy's skills in design, engineering, and product as opposed to dictating what has already been decided to build. You're not getting the bang for your buck from the public sector side to do that. 


System diagram

Roeder:

In summary, ask the user. Can't emphasize that enough. Start small, test, and iterate, build the right team, and find the right partners. Let's check in again with the diagram that Roshen introduced to us and see where we are. We talked about introducing products between agencies like MOCD and our constituents, the public, to improve service delivery primarily. But what we found though in this process of sitting with people and listening to their needs and not just applicants but all our partners, is that we were building a feedback loop. We were building new communication with this constituency that helps us do our job better overall even outside of the context of the product itself. 

Using Data To Transform Policy

Sethna: To dig a little bit deeper into that, this feedback loop story is really around data. The data that we've seen has started to help us improve services but it's also starting to help improve policy. By policy, I mean policy at a variety of levels. I'm going to give a couple of stories around how DAHLIA did this. 

Use data to increase program effectiveness

One story is at the program effectiveness level. Like Barry mentioned, the City of San Francisco runs their housing on a lottery system. The leasing agent will run a lottery and then use those ranked lottery numbers to lease up the building. In that lottery process, we have something called preferences. It's basically certain citizens who get preferential treatment in the lottery because they might have been evicted from a certain neighborhood at a time or currently live in the same neighborhood as the listing. We want to keep them in that neighborhood and retain them. There's different preferences for people in the housing lottery process. One thing we were seeing is that people were really not taking advantage of these preferences that actually gave them a huge leg up in the lottery, and one preference they were not taking advantage of is called the Neighborhood Resident House preference. This is the preference where if you live in the same neighborhood as a listing, we'll give you preferential treatment to be able to retain you in that neighborhood. Only about 40 folks were actually claiming this preference. What we did was we simply redesigned this page. We added a banner at the top and said, "Hey, you qualify for this thing and this is what it is. We explained it to them because nobody understands what Neighborhood Resident Housing preference means. We actually changed some of the content. We worked with Barry's office to say, "Hey, can we just call it Live in the Neighborhood preference? Is that okay?" They said, "Yeah, that's okay as long as somewhere else on the page we put Neighborhood Resident Browsing preference because we have to have that on page." We worked within the business rules of the organization to make things more user friendly and we saw the claim rate of this preference jump to about 88%, more than double the claim rate. Really simple design and code changes make it so you're able to see if people are claiming services that they are eligible for, and if they're not, how can you drive them to those services and programs and encourage them to claim it and describe it to them in a more human way. 

Use data to improve business processes

Another example is a little bit about the programmatic level - going to the organizational level. Barry's colleagues at the Mayor's Office of Housing are currently as we speak rewriting the housing policy manual that the Mayor's Office of Housing uses. It's a lot of business processes around how buildings are listed, what the building needs to do to report to the Mayor's Office of Housing -  all the business rules around managing the housing process. A lot of conversations in that rewrite came from the product. One really small example of this is that on a listing the minimum income you need to get into that listing is generally a multiplier of rent. If rent is 500 bucks a month, then the minimum income requirement might be two times that or three times that or 3.2 times that. The problem was that each building was using a different multiplier. When our engineers asked while building the site what the multiplier rule was and what multiplier buildings usually use, we didn't actually know because every building was doing their own thing. The Mayor's Office of Housing is taking the opportunity to streamline rules like that and really create a more equitable process between buildings and the consistency that didn't exist before. This is happening in a lot of different business process levels. That's business processes. 

Use data to transform housing policy

The third and final layer is starting to get to this legislative level. We are starting to see a lot of data come through DAHLIA which I'll talk a little bit more about. The Mayor's Office of Housing can now understand where they should invest in housing, what they should build, if their two bedrooms are meeting the needs of the population more than three bedrooms and studios, if people like to live in certain neighborhoods and why that is, if housing is close to transit or other jobs. They can start to analyze what's coming through the site to better understand where they should invest their time and money and their programs.

User Demographics

Overview of the demographics of DAHLIA users.

Sethna: 

A little bit more just about the demographics we're seeing in DAHLIA. A lot of folks are applying, not surprisingly, from a lot of major Bay Area cities to live in San Francisco. We're seeing over 89 languages represented on the site, a really wide age range of 25 to 65, we have a lot of senior buildings on the site, we see seniors apply, and about 58% of users right now are accessing the site on mobile. It's really common for folks to access it on mobile as mobile is generally a primary access point for the internet for a lot of users. A lot of people access the site through their housing counselor or at a public computer or at their local non-profit. We're trying to understand who is seeing the site, where are they coming from, what's the need in San Francisco. The reason we can start to analyze this at a little bit more of a population level is because we really are getting a lot of data coming through. At this point, we're about at 101.4 million site visits and we can start to see the trend over years. What was happening in 2015, 2016? How do we see, as there are more listings coming online, different things that people gravitate towards? There's really this volume of data coming in at the Mayor's Office of Housing in a way that wasn't before. Barry's going to talk a little bit about what the Mayor's Office of Housing team is saying about this. 

Roeder:

Sometimes when we talk about this, I'd like to say, "And you thought we were talking about a service delivery product. We thought we were, too." One of my colleagues refers to it as this data exhaust. There's so much stuff that we get from Google Analytics and we use another service called Heap Analytics. One of the things I talked about earlier is that one of the government's roles is to have an eye on equity. So as Maria Benjamin, the director of our housing placement programs, says, "We finally have real data to help us identify and remove barriers to housing placement for specific segments of our population." 

Let's look at our housing funnel for a second. Up at the top level, we have now almost 1.5 million site visits. Say, of those 1.5 million site visits, eight percent are X demographic and we have 80,000+ applications. Wait a minute, only six percent of the applications are that demographic. This is interesting in the almost 2,500 units that we place to date through DAHLIA, it drops to 5%. What's going on? What's the problem? Is there an issue in the app that's dissuasive to people creating an application? Is there something in our process? That's something that obviously is a great concern to us and this gives us an opportunity to really look at what's happening with that. Even beyond that though, Roshen made reference to this. In a steering committee meeting about a year and a half ago, Kate Hartley who's our Executive Director said simply and eloquently and succinctly, "This will change what we build."

The way we used to think about that is, okay, we have some money to build some housing and there's a little bit of, "Well, what do you guys think? Ten one bedrooms because that's what we did last time? There was no source of data to understand what we should actually be building. Now we can say, "Wait, stop. If you're going to build a 50-unit building in the Mission, in that area, people of this household size, and that income, that race ethnicity with that operating system on their computer, are looking for that type of unit." Whoa. We've never had that before. We've been in regional conversations, smaller jurisdictions in the Bay Area, where someone from the City Council, maybe it's an affluent City Council member says, "Yeah, I was driving through and I saw a Latino family on the street and I thought we should build some more of those units like that." Well, that's a little bit ad hoc, a decision like that. It's really exciting to have this information that will help us answer the question about what kind of housing crisis we have, who particularly is struggling, and how we focus the relatively scarce resources that we have on this issue. 

Creating Feedback Loop

Sethna:

To go back to our diagram because we love diagrams. We've talked about where products are, we've talked about this feedback loop between agencies and the public the product is starting to create, but we're really talking about here is changing business rules. Policy is a larger feedback loop between policymakers and the public. Barry mentioned this earlier, but he leads a steering committee at the City. They check in on DAHLIA really regularly, and the steering committee is composed of a housing policy advisor for the Mayor, Mayor's Chief of Staff, a lot of folks that either influence or are involved with policy making in the City. They're able to start to have these touch points to ask what's coming through DAHLIA, what are we seeing? The City can even start to say we have this inclusionary housing policy. How is that working out for people? Are people actually accessing this housing? How is it working? Whose needs is it meeting? Whose is not meeting? This conversation that's being started at the steering committee in other places is really exciting to us because we get to use this really pretty simple digital product to start these policy level conversations at the City and County. If this kind of feedback loop process looks familiar to you, especially a lot of folks who use agile are in the startup community, that's probably because it is. 

Roeder: 

That's right. That's that diagram we pulled from the earlier slide. The idea of being able to make pivots on the fly in your development. We were talking about that in the context of building a product. Now we're talking about it in the context of establishing policy. Because we have those feedback loops and more frequent interaction between the public and policy makers, we can do a better job of aligning policy to the needs of constituents.

Sethna:

I think we're really particularly excited about this cycle of feedback loops and Barry's really passionate about the image of City Hall.

Roeder: 

I love the picture of City Hall because this is San Francisco City Hall for those of you that don't know. There's this beautiful, old, I'll say, legacy government building that now has this fancy digital LED lighting system. The idea is that we can take the strengths of a legacy system like we've discussed earlier: the service delivery infrastructure, democratic process, eye on equity. We can leverage that with this groundbreaking agile data-driven process that allows us to shorten innovation cycles. We can imagine one corner of this building started with CAT5 cables wrapped up in an old stone column. We're really excited about the possibilities that that methodology offers.

Sethna: 

We'll be around all week. We're really excited to chat with every one of you. Like Barry mentioned, we're actually talking to a lot of other jurisdictions about DAHLIA itself to implement it regionally and think through what a regional solution and a regional kind of database of affordable housing looks like first here in the Bay and then in other areas. Please come talk to us about that. Thanks!

Share on Linked In

Exygy

Bloom is a product from Exygy. We partner with social impact organizations to design and build technology that improves lives.

Copyright ©2023. Exygy Inc.

Exygy

Bloom is a product from Exygy. We partner with social impact organizations to design and build technology that improves lives.

Copyright ©2023. Exygy Inc.

Exygy

Bloom is a product from Exygy. We partner with social impact organizations to design and build technology that improves lives.

Copyright ©2023. Exygy Inc.