Predicting Homelessness Before It Happens: Kevin Finn on Prevention, Data & Impact

What if you could predict who is going to be homeless in the next sixty days? What would you do about it? Flywheel Social Enterprise Hub goes above and beyond to inspire and support social enterprise through focused innovation. Our innovation lab is a powerful platform for bringing together civic leaders, funders, and founders with a shared mission to create focused innovation on a priority community challenge. Hi.

I'm Joan Kaup, coproducer of On The Fly!, where we shine a spotlight on social impact entrepreneurs and the folks who support them. Today, I have the pleasure of talking with Kevin Finn, the president and CEO of Strategies to End Homelessness in Cincinnati, Ohio. Flywheel pulled together innovative thought leaders to tackle the challenge of identifying families before they become homeless. We're going to talk about that and more. Hi, Kevin.

Thank you for joining us on On The Fly!.

Thanks for having me.

It's a pleasure. So just just for context and just so our listeners have more of a sense of who you are and where you're coming from, boil down in a few minutes your journey. How did you get to be president and founder of strategies to end homelessness?

Like, I could probably spend our whole time just answering that question, but let me see how quickly I can summarize. So I went on a mission trip to Peru when I was in high school, and it was very much an experience of third world poverty. And it changed my life, and really I decided that I wanted to work on poverty related issues for my career as a result of that experience. And it took me about ten years to figure it out, but after about ten years I figured out that to me homelessness is the most extreme version of poverty that you see, you know, in the richest country in the world, where we are. So, you know, I decided that working on homelessness was what I wanted to do.

My first job working with the homeless was as a street outreach worker working with homeless teenagers beginning in 1998. And what I figured out from that job was that the homeless service agencies were doing a lot of great work but they weren't necessarily coordinating their efforts and they didn't even necessarily talk to each other all that much. So in 2007, with a lot of help from a lot of people, I started Strategies to End Homelessness And, you know, our mission is to lead a coordinated community effort to end homelessness in Greater Cincinnati. So the emphasis is really on coordinating the system because what we don't want to have is, you know, some organizations doing things that are sort of redundant meanwhile there gaps are in the homeless services system that no one is working to fill. So, I mean, our job is to make sure that the entire system is coordinated, reduce redundancy, increase efficiency, and make sure that there aren't gaps in services that are still needed.

But that's sort of been my, you know, journey to ending up where I am now.

And that coordinated effort is working. I mean, I hear only good positive things about it. How many organizations are you working with right now?

So we we coordinate the work of about 30 different organizations who are all doing different things related to homelessness or preventing homelessness locally. Another role that we have grown into over the last ten years or so is that we actually manage a lot of the funding related to homeless services. So you know the big funder of homeless services is the Department of Housing and Urban Development. So we coordinate the HUD funding that comes into Hamilton County related to homeless services but we also manage some state, county, and city of Cincinnati funding. So we pass through funding to about 20.

We're gonna we work with about 30.

Okay.

We pass through funding to about 20. So our role is different depending on which organization you're you're with, which of our partner organizations somebody might be with.

And to manage that funding and accrue all that funding requires a lot of statistics and a lot of record keeping. Is this something that was in place or did you have to work to make that happen?

Well, in Hamilton County, actually, you know, HUD requires communities to have what's called a homeless management information system or HMIS. Hamilton County was actually the first community in the nation to have all of its homeless service organizations participating in the same HMIS system which means we have comprehensive data on who's homeless in Hamilton County at least going back, you know, twelve, thirteen years. Some communities don't have that for last year. Northern Kentucky, you know, some surround some areas very close to Cincinnati couldn't tell you how many homeless people they had last year, and we can tell you down to the person all the way back to 2013 how many people were homeless in Hamilton County. In 2017, we started working with eighty four fifty one, the data science company in Downtown Cincinnati, to start looking at that data to make sure we were learning as much from that data as we possibly could, identifying ways to you know who was most likely to be homeless, who's most likely to return to homelessness again, all that sort of thing.

Really in the last nine years or so it's really been a point of emphasis for us to make sure that we were getting as much out of that data as we could. But what we quickly learned was, you know, that data on people when their homelessness is a little bit of a flash in the pan. Like we originally thought that would tell us everything we wanted to know, but it really only tells us about from when a person became homeless to when they exited homelessness, and it didn't tell us anything about where were they before they were homeless, where do they go when they leave homelessness. So in more recent years we've gotten into collecting data from a variety of sources that tells us more about who is at risk of becoming homeless, who's at risk of returning to homelessness, that sort of thing, so that we can impact the issue in a broader way than we could if we were just looking at data on people while they're homeless.

And I believe that was the crux of the hackathon of the innovative lab that Flywheel pulled together. So it was to address how do we stabilize housing, how do we predict homelessness. How did that come about, and how did that work?

Well, the housing stability hackathon that we did with with Flywheel and Flywheel was awesome in helping pull that together was we were already working on a project to use predictive data analytics to identify the people at highest risk of becoming homeless. But there were gaps, and we knew there were gaps. So six months before the project was gonna go live was when we did the the hackathon with Flywheel. And the intention was to get a group of innovation minded people to look at what we were doing and offer improvements and and things like that. And one of the things that was specifically missing at that point was a way for people to self report that their housing was at risk and they thought they might get evicted or that sort of thing.

So the tenantguard.org website that came out of the hackathon, was an improvement suggested by the winning group in the hackathon. That that website is basically a survey that allows people to give us information, give us data that did not already exist about their own housing situations. And that data has proven to be very, very valuable to that project. And then it has and then the second purse function it serves is that we've had some people who we actually already knew were out there in the community, and we already knew they were at risk of homelessness, but we didn't know how to reach them. But when they find their way to the tenant guard website, they provide us updated contact information.

So now all of a sudden, we're able to make contact with them.

So I was there that day, and there were I think there were 40 applicants, and then it was boiled down to nine who actually worked with you on the data you presented, and they made presentations. And the one that rose to the top was we're gonna build a website where people can report themselves that they're at risk. And the website didn't exist. They built it from scratch for you all. Mhmm.

Is that right? And that was the eighty four fifty one team.

That was the team from eighty four fifty one.

You worked with them prior to the hackathon?

Not those particular eighty four fifty one employees. I think they heard about the hackathon and got involved in part because we had a prior relationship with eighty four fifty one. They had a prior relationship with Flywheel. But that group of of eighty four fifty one employees that came up with the idea, I'd never met them before. That was a new group of people.

So the team the team was led by Betsy Emchie, data scientist there at eighty four fifty one, which is an amazing company we have in town. Bjorn Borrell, Jacob Panesic, and Nick Ramos. Yes. So that was the team. Did they actually then ever work in your offices?

Or how how did you go from, oh, that's a good idea to, oh, now we've implemented it?

It was sort of a joint effort. My staff who do our website and different things because, know, ultimately, we needed to host the website and own it going forward with the Tenant Guard website. So we worked with my staff, worked with them over the course of several months, to get the Tenant Guard website up and running. So it was sort of a joint effort. And, you know, this was all happening and and, you know, the the hackathon was complete in February 2024, and the project went live in July 2024.

So it was in that window.

Fast.

Yeah. It was in that window. It it I don't I think that the website, Tenant Guard, I think, actually went live in August or September. It didn't quite start at the same time as the project, but very shortly thereafter. And then, you know, it has become a very valuable resource in terms of of the data sources we have available to us to use to predict who's going to become homeless.

It has become one of the most valuable data sets that we have plus the additional wrinkle of giving people the opportunity to update their contact information so that we can reach them because what sometimes was happening again was we would identify these families that we knew were out there in the community. We knew they were at risk, but all the contact information we had for them was out of date, so we couldn't connect with them. And Tenant Guard also, in addition to providing us additional data about that family's situation, gives them the opportunity to tell us how we can reach them.

I I have to tell you, they were not when they were saying that people will come forward and and share this information, I'm like, yeah. No. They won't. I mean, I understood the conversation. Like, if we get if we get from the utility company with the bills that are not being paid, if we know the children who are missing a lot of school or changing school to school in one year, if we you know, data that you could get from third parties and pull it all together and try to interpret all that.

Okay. I was following that. But the whole idea of and then we'll have this website, and people would just go to us, and they'll come to us and tell us they're at risk. And I thought, no. They won't, but they do.

They do. They do. You know, a big part of the website has also been promoting the website in places where people whose housing might be at risk are going so that they learn about the website. So there has been sort of a marketing component to getting the word out there about the website. So but when people do know about it, they do fill it out and provide us with very, very valuable information.

You know, the biggest frustration I have around that is just a question of capacity, and I think we're probably only able to actually offer to us assistance to probably one out of the one out of 10 of the people who go to the TenantGuard website and and fill it out. But that's just a capacity issue. You know, we I should absolutely mention this whole project, which we refer to as the housing stability collaborative, is using predictive data analytics to identify families who are most likely to get an eviction notice and then we proactively reach out to those families and offer them assistance. There is a service delivery component on the back end. This whole project has been started with funding that came from the city of Cincinnati.

The city set aside $2,100,000 a few years back for what they called the Impact Award, which was a first of its kind grant to do something innovative and new that would reduce evictions. And we were one of multiple organizations that applied for that Impact Award funding and were lucky enough to receive it. So, you know, the city of Cincinnati, I think, deserves a lot of credit for doing something very innovative in this space. But because of some limitations on the funding they provided to us, first of all, we're only able to serve people who are residents of the city of Cincinnati.

Oh, I see.

And we're only serving families. Meanwhile, you know, more than two thirds of our homeless population is single individuals. So, you know, this project not only doesn't have the capacity to serve everyone who we are currently targeting families in the city, we're also completely disregarding single individuals and people who do not live in the city of Cincinnati simply because we don't have the resources to serve them. So the potential for the project is enormous because, again, if given enough resource, could expand to single individuals, which is the majority of our homeless population. We could expand to all of Hamilton County or even all of Greater Cincinnati.

And, you know, we've even had conversations about expanding what we're doing to other communities around the country and offering them what we've learned through doing this because we are one of only less than a handful of communities in the world that are trying to use predictive data analytics to reduce evictions and homelessness.

If a family if you can help a family stay in their home, it costs x amount of dollars per month per year. If they lose their home, it costs x amount of dollars per month per year. Fill in those gaps for me.

Well, when a fan again, the numbers I have are primarily about families because that's who we're serving. So the average family that walks into a homeless shelter, you know, that becomes literally homeless. Again, because we administer all this funding, we're able to calculate what on average do we spend on that family. So $22,500 is the average cost of helping a family that becomes homeless.

For you?

For an episode of homeless. For an episode. An episode

of could be a

very Some are longer, some are

shorter. But

the average cost is $22,500.

Okay.

In the housing stability collaborative where we identify that same family while they're still in an apartment and we help them keep that apartment rather than losing it, $3,600 is what it Wow. So with six to one, essentially, we can serve six families in this predictive analytics project for every one family that comes into a homeless shelter.

Did you have a sense of that before you started this housing stabilization project?

We were already running a program called shelter diversion that targets families in the middle. Families that have already lost their own housing and are sleeping doubled up with friends and family, but are not yet in need of a shelter bed kind of thing. That project costs about half as much as helping a family when they're so it's about $13,000 to help when a family's already doubled up with friends and family versus the 22,500 when they walk into a homeless shelter. So we already knew that by going upstream, it got cheaper. Mhmm.

We didn't realize how much cheaper it would be. We figured that it would cost less to keep a family in their current apartment than it costs to provide them with shelter diversion services, we did not necessarily anticipate just how much lower the cost would be. It is significantly lower. That's not the coolest thing, though. We are 97% effective at keeping families in their apartment for $3,600.

Whereas, if you help a family after they're in a homeless shelter, almost a third of them will become homeless again.

Oh my.

So when you spend the $22,500 to help a family, almost a third of them will become homeless again. And you're insane. $100, we're 97% successful. So

That's

interesting. The outcomes are so much better in addition to it being so much more cost effective.

And and you think you can take that to other communities and show them and teach them how to do this?

You know, the the legwork it has taken to pull this off is considerable. Yeah. So what we think we could potentially do is that we could offer what we have learned and what we have built to other communities at a way lower cost than what it would cost them to go out and build it the way we

did. True.

So it would be a cost savings to those communities, but by giving it to other communities at a reduced cost, we could create some revenue to help us serve more people in here in Greater Cincinnati. So, like, that's sort of the model that we're looking at is they get it at a lower cost than they could build it themselves, and we potentially generate some revenue that allows us to serve more people here. And that type of revenue might be what would allow us to expand to serving single individuals, people outside the city of Cincinnati, all that sort of thing. But, you know, that's all sort of in the planning stages at this point.

Well, when I think about these families and I think of the children, if you can keep them in their homes and keep them in the school and and avoid some of that disruption to those young lives, what a difference that makes long term. That's huge.

Well, Joan, you point out to me that I'm missing the most I've I've I've failed to mention the most important part is that, you know, there's so much trauma that people experience when they become homeless. Yeah. And, oh, by the way, none of that happens because you know what what frequently happens with kids kids is really the proper place to focus is you know between when a family loses their own apartment and when that family might later end up in a homeless shelter, it's not uncommon for kids to change schools somewhere between four to six times because, you know, every time they change from staying with this family member to that family member to this friend of the family, each one of those moves may necessitate changing schools because their address is changing so frequently. And that just creates a trail of disaster for those kids educationally in terms of their health, mental health, all that sort of thing. So again, for $3,600, we keep all of that from happening.

The kid stays in the same apartment, stays in their school of origin, never has to experience, you know, so many negative things. So, I mean, that really is the most important thing. And until you brought it up, I forgot to mention it.

You can't forget it. That's So I I tutor at a little inner city school and to keep those family stables. I mean, what what a difference it makes on these children and how they learn and what they bring to the classroom. I mean, it's Well, I mean, I difference.

Yeah. I can only imagine, like, one of my kids changed schools at one point in the sixth grade and went from one really good school to another really good school. But just so happens that there was some American history that the one school hadn't covered yet and the other school had already covered it. So my child missed it and that was a very intentional very well thought out school change and still there was educational content missed as a result. So now imagine you're changing schools four times in one school year.

Like, how much stuff do you miss? Yes. And that sort of thing. So yeah. And reducing all of that for children is critically important because it it puts those kids behind in ways that they may never be able to catch up.

That's true. That's true. It's and this is a huge, huge project, and I'm glad that the success rate is there. Ninety seven percent is just an an unbelievable number.

I I do think it's important to keep in mind that with that you know, we've only been at it for a little over a year and a half, so the sample size is still not huge. We're actually later this year we're beginning a randomized controlled trial with Notre Dame. They're going to do a research study on the project that'll probably take at least two years to complete. So, and they're already doing a research study on our shelter diversion program that I mentioned as well, because I think it's very important that we document how the effectiveness of these types of interventions because these types of interventions do not exist in most communities. And if we want other communities to try doing something similar down the road, we we have to have research to back up that that it works.

So we're working with what's called the Lab for Economic Opportunity or LEO, which is a part of the economics department at Notre Dame. And they're doing studies on both of our homelessness prevention programs so that we have that documentation. The biggest frustration in my professional life is we get over $30,000,000 a year from the Department of Housing and Urban Development every year. Exactly zero of those dollars can be spent on prevention. Oh my.

So the main thing I would like to do is complete these research studies and then take them to Washington DC and beat HUD over the head with them and say, you know, you should be funding prevention because it's cheaper. It saves people the trauma. It has better outcomes. All these sorts of things. Part of the reason for engaging in the research with Notre Dame is specifically so that we can try to advocate for policy change at the federal level because if we could you know, our total budget for homelessness prevention in Hamilton County for for our homelessness prevention projects is less than $3,000,000 a year, but HUD gives us over $30,000,000 a year that we can only use to help people after they're homeless.

So if we had the flexibility just to shift 3,000,000, let's say, of that 30,000,000 over to prevention, we could double our prevention capacity, serve six families for everyone on the homeless side and just have a much better run system in that regard. But the current reality is that we can't use the HUD funding for prevention.

So if our if our listening audience wants to help support your efforts, what what is there what can we do? What

Well, you know, the reality is because there's so much funding for housing post homelessness from the federal government, all of our fundraising and efforts to secure additional resources are all on the prevention side. So, I mean, it's I always feel like it's sort of a lazy answer to say donate, but all the money that we raise goes to prevention because that's where there's the least resource. So that's one way people could help. The other thing that we have is and Flywheel was instrumental in helping us think in these terms. You know, there's data out there that has predictive value to what we're trying to do, and we're still on a regular basis learning about data sources we were unaware of previously.

So who knows? Some of your listeners might be aware of something that's out there that would have a predictive value for us, you know, because we're currently using data from, you know, people who go into Saint Vincent Paul and request services or who are requesting food from the Free Store Food Bank and obviously completing the tenant guard survey. We're using Clerk of Courts data on eviction histories and data from a variety of other homelessness prevention programs that exist around town to help us identify who is out there that is at risk. But there may be other ways that we haven't thought of to identify those people and other data sources that we have not yet tapped into that would be helpful.

And how do we find you? What's your website? What's all your Internet stuff?

Well, the the organization's name is Strategies to End Homelessness, and our website is just straight out our name, strategiestoendhomelessness.org. That's the easiest way to

find it. Alright. Well, this is quite the journey from, teenagers in Peru. So is it what you expected?

My wife laughs at me because she she'll regularly just sort of say to me, you know, we employ data analysts. We have an entire data team that is a part of our organization. So there's four people in that department. We also have five people who just run the homeless management information system. So, I mean, we got, you know, nine or 10 people who do nothing but data.

Yeah. And my wife always just sort of kids to me. She's like, that doesn't sound like social work. You know? Because my my you know, I have a master's degree in social work, and she always is just sorta like, that doesn't sound like social work.

But but it is. It's trying to take a innovative new approach to sort of an entrenched social problem. And I think that people, you know, I imagine most of your listeners are here in Cincinnati kind of thing. I think what the average Cincinnatian doesn't know is that the most innovative things going on around homelessness, I think, anywhere are happening right here in Cincinnati. So we don't have to look to I have received calls from San San Francisco and Los Angeles and other communities wanting to learn about what we're doing.

But people tend to perceive that well, I'm sure the most innovative things around reducing homelessness are happening in those places. No. They're calling us. Yeah. So and I think they're you know, Cincinnati should be proud of that because if we didn't have Eighty four fifty one who obviously works closely with Kroger, if we didn't have the city of Cincinnati putting up the Impact Award funding I mentioned before, if we didn't have Flywheel helping us identify ways to improve the project, like, wouldn't be happening.

Yeah. But it's really only happening here.

That is so wonderful to hear. We do punch above our weight in a lot of different areas, and I think social impact is one of them. It's intentional. And it's groups like Flywheel that keep it in the forefront and keep the intention.

I agree completely, but I don't think the average Cincinnati necessarily understands that.

Yeah. Yeah. It's it is kind of a big deal. So thank you for reminding all of us, and thank you for all that you do. It has made a huge difference in this town, and we could not be more pleased and proud to be a part of your success.

I have great staff. They do the heavy lifting.

You do a lot of good. Thank you so much.

Thank you.

On The Fly! is produced by Joan Kaup and Joey Scarillo with support from Gianna DiMona. Music by Ben Hammer. It is recorded at 1819 Innovation Hub courtesy of the University of Cincinnati.