Case Study: A Data Hub for Affordable Housing

color post it wall · jc_peckham_peace_wall_015_0 · to-do2

Part 2: What We Know…and What We Know that We Don’t Know

One advantage of being in a social impact accelerator is that you are part of a cohort of passionate people who believe in possibility. They believe that it’s possible for people to do the right thing, even when there’s nothing in it for them; that it’s possible that government can serve and innovate and inspire; that it’s still possible to believe that something that starts small can change the world.

Another advantage is that the accelerator leadership and mentors will hold you accountable. And expect your numbers to make sense. It’s not all deep breathing and colorful whiteboards and the tantalizing fluttering of flocks of post-its — sometimes you need to make the budget balance.

As a project team, we came into this accelerator with what I’d learned in grad school was called conscious incompetence. It’s not as bad as it sounds. Making mistakes and iterating is essential to learning in this stage, and I’m ready to do both, as long as I can do it really fast and report back to the team. However, I am already sensing that some things are going to take…as long as they take.

We had already spent a good amount of time on customer personae, on journey mapping, and on user research, thanks to talented design thinking friends and colleagues at IBM, Idean, and the LBJ School. And we have top-notch geniuses from Code for America gathering data, cleaning it, organizing it, and mining it for insights as they assess solutions.

The plan, then, was to use this accelerator opportunity to address two key implementation and delivery issues: the accuracy of the data, and a business model to sustain that accuracy, as well as to add future user enhancements. Here’s what we knew for sure:

1. The system to host this could be relatively simple — we could provide just a bare-bones interface with open data, and let the market do its work to add bells and whistles. Even so, there would still be an on-going maintenance and support cost that must be covered.
2. There is a natural data “control point” for registration for voucher payments, and an opportunity to apply a surcharge there…but the potential housing properties that are not yet registered don’t have a natural incentive for landlord registration. (This is especially true in markets where demand far outweighs supply, such as Austin.)
3. Gosection8.com provides incentives for its landlord clients in terms of ease of leasing and pre-screening of applicants…but this only speaks to a certain percentage of the market.
4. The city is able to work with non-profit investment property management groups that provide exclusive Section 8 available housing, which could helps increase the accuracy of available listings… assuming they follow processing procedures and data entry standards.
5. Another option might be forging partnerships with apartment and rental brokers to include a required checkbox for Section 8 acceptance…but if they all use different forms, compliance could be sketchy.
6. We could explore property tax credits for landlords that register their housing availability. But of course, that gets into public policy…which is very different than tech and data analysis.
7. While it could save low-income applicants time and effort, its unlikely they could pay a premium for the service. And, honestly, we wouldn’t want to ask them to. They’re a key end-user…but probably not the customer.
8. Landlords in a demand-driven market like Austin would also be unlikely to pay for a tenant-matching service. Some developments already have waiting lists of 500 or more qualified families, and have no need for or interest in yet another dashboard to check.
9. According to the 2015 Census estimates, there are about 750 US cities with a population of more then 50,000, which is the kind of volume we think we’d need to have to make certain funding models work. We’d need to get more data about smaller markets and the potential market penetration rate to test whether any funding plan tied to the number of units or landlords could scale.
10. The idea of HUD adopting of this initiative at a federal level and rolling it out nationally is currently unlikely.
11. Ongoing funding will need to come out of operational expense budgets for HACA and NHCD, from transaction fees, advertisement sales, consistent renewal of grants…
12. Or some combination of those.
13. Do we do the maintenance in-house? Crowdsource it, with incentives? Both?

We knew a lot, really. But we also still did not know A LOT.

My key insight for the week came though an early-stage founders class I’m taking (yes, at the same time as this accelerator — I assure you seemed like a good idea at the time). And that insight is that figuring out your customer model, your market model, and your business model is basically a game of Whack-a-Mole. A small shift in one can send you back to the drawing board in another, and a fresh new idea can knock your founding premises out from under you. And they all need to be relatively stable and in sync before you add your for-real financial model to the equation.

It’s hard to do that kind of work with just one brain and two hands. It takes a village.

There was going to be rather more brute force and guesswork here, I realized, and rather less pure ideation, bolts of inspiration, and peaceful moments blithely scattering post-its like a flower girl at a royal wedding.

And so: we iterate.

Jennifer Houlihan is managing director for Austin CityUP, a smart city consortium in Austin, Texas.