Housing assistance directories: verifying data for NYC clients

That task has never been straightforward in New York City, and over the past few years the structural friction behind it has become harder to ignore. From the August 2025 exposure of applicant data through the NYC Housing Connect lottery to the long-standing reality that our city's health and human services agencies operate in data silos, the information layer underneath housing assistance is fragmented in ways that directly shape whether a referral leads to a placement or to a dead end. I want to walk you through what I have learned from working inside this ecosystem, because the practitioners who hold these directories — the front-desk staff, the housing navigators, the program directors — are the ones who translate bureaucratic density into real outcomes for New Yorkers every day, and they deserve more than generic advice about "verifying data."

The Reality of Fragmented Databases: Why NYC Housing Data Remains Siloed

If you have ever tried to compile a master list of every housing assistance program operating across the five boroughs, you already know the feeling: the list is never finished, and it can never be finished, because no single authority maintains it. New York City's eight health and human services agencies each maintain their own program data, their own eligibility criteria, and their own intake workflows, and the Center for an Urban Future made this problem explicit in its November 2022 analysis of the local social services ecosystem. The report found that HHS agencies operate in data silos, and rather than recommending a wholesale restructuring of these systems, it pointed toward something more practical: the integration of application programming interfaces that would allow existing databases to speak to one another without forcing every provider into a single schema.

That recommendation matters because it respects the operational reality on the ground. A homeless outreach provider in Manhattan does not need a theoretical unified database; they need to know whether the family they are working with tonight has already been seen by a prevention program in Queens, and whether that program closed a case successfully. Without API-level integration, that kind of cross-borough awareness depends on phone calls, faxes, and the institutional memory of individual caseworkers, which is precisely the kind of knowledge that disappears when staff turnover happens. The problem compounds at scale: the more organizations you add to a referral network, the more phone numbers and fax lines and contact names someone has to keep current, and the faster that informal knowledge base degrades.

The challenge is not a lack of data. It is the absence of connective tissue between the databases we already have, and that absence has a direct cost in client outcomes.

The August 2023 audit by the New York City Comptroller made this cost concrete in the context of the Department of Homeless Services. The audit found that DHS lacks a single data system to track engagements with unsheltered individuals and relies instead on unaggregated case notes, which means the agency cannot systematically measure housing outcomes for the people it serves. For the rest of us in the nonprofit ecosystem, this is not an abstract concern about municipal data infrastructure. It shapes what we can promise the people who walk through our doors, and it shapes what we cannot.

The fragmentation also has a timing dimension that is easy to underestimate. When a new rapid rehousing program launches in Brooklyn, it may appear on one agency's website within a week and take three months to reach the databases that front-line navigators actually use. During that gap, clients are referred to programs that have changed their eligibility thresholds, moved their intake location, or closed their waitlist entirely — and the navigator making the referral has no way of knowing. The infrastructure was never designed for the speed at which program availability changes, and the people who bear the cost of that mismatch are the ones with the least margin for error.

Quantifying the Accuracy Gap: Record Duplication and Unique Listing Challenges

The numbers around social service directory accuracy are sobering enough that I want to spend some time with them, because they explain why your internal housing resource list, however carefully maintained, will always be incomplete. A comparative analysis by Benetech Service Net and Open Referral found that 74% of social service directory records were duplicated across at least two databases, while only 25% of the records in any single directory were unique to that directory. The implication is uncomfortable: most of the entries you encounter are echoes of entries elsewhere, and a meaningful share of what any single directory actually contains is information you cannot find anywhere else.

This duplication problem interacts with timeliness in ways that magnify the difficulty. A housing program that updates its eligibility criteria in January will appear with the old criteria in three different directories for months, because each of those directories has its own refresh cycle, and most of them are not real-time. The organizations most likely to be out of date are often the ones experiencing the most operational pressure: small community-based providers that simply do not have a communications staffer whose job is to chase directory editors.

The honest takeaway is that no single nonprofit housing registry can serve as the authoritative source for the entire city, and the work of building a reliable internal list is the work of reconciling multiple sources rather than trusting any one of them.

What this looks like in practice for a mid-size community-based organization is a reconciliation routine that most teams do not have time for but cannot afford to skip. You pull the latest listings from two or three directories — your own internal list, a borough-specific referral database, and one of the citywide platforms — and you compare them line by line. Where records overlap, you note which source updated most recently. Where a record appears in only one source, you flag it for a phone call to the program directly. This is tedious, unglamorous work, but it is the single most effective thing a housing navigator can do to reduce the rate of referrals that go nowhere. The time spent on a monthly reconciliation is almost always less than the time spent chasing a referral that turned out to be based on outdated eligibility rules.

The practitioners who hold these directories are not short on effort or care. They are short on infrastructure that does the reconciling for them, and that gap costs real people real housing opportunities.

Lessons from Recent Security Failures: Protecting Client Data in Housing Registries

The security of housing directory data became a public concern in a way it had not been before when, in August and September 2025, a data breach in the NYC Housing Connect lottery system exposed applicants' personal information in public search engine results. City Council responded by convening hearings on data security, and the conversation has since extended well beyond the lottery itself to the broader question of how housing assistance data is stored, who can access it, and how long it is retained.

For nonprofit providers, this is a moment to audit the way we handle our own directories, because housing assistance data — referral histories, eligibility documentation, intake notes — can sit in exactly the same kinds of loosely governed systems that raise alarm when they appear in a municipal platform. The most common exposure I have seen is the directory shared by email or stored on a shared drive without row-level access controls, where a single overshared link can effectively publish a client's history of housing assistance requests to anyone in the organization who happens to open the file. The Housing Connect incident is a reminder that these risks are not hypothetical, and that the organizations holding the most sensitive referral data are often the ones with the least formal information-security infrastructure.

The standard we should hold ourselves to is not "could a sophisticated attacker get in?" but "could a curious colleague with the wrong link see something they should not?"

Three operational practices have served the organizations I work with well in this regard. First, segregating the directory of available programs from any record that connects specific clients to specific referrals; second, applying access controls at the file or database level rather than relying on folder-level permissions; and third, scheduling a quarterly review of every external integration — every API key, every shared spreadsheet, every mailing list export — to confirm it is still necessary and still appropriately scoped.

A fourth practice, which I have come to consider equally important, is building a culture where staff understand why the segregation matters. Technical controls are only as effective as the people operating them, and in my experience the most common security gap in nonprofit directories is not a software flaw but a well-meaning caseworker who exports a client list to their personal laptop to finish intake notes at home. The conversation about directory security has to include a conversation about workflow: if the only way your staff can do their work is by moving data outside your system, then your system needs to change, not your staff.

Operational Strategies for Closed-Loop Referrals and Agency Coordination

The phrase "closed-loop referral" gets used loosely in our sector, but what it actually means is straightforward: when you refer a client to a housing program, you eventually learn whether that referral resulted in service, and you record that outcome. In 2023, NYC Health + Hospitals transitioned its social care referral platform from NowPow to Findhelp, and within the first six months achieved an 86% closed-loop referral rate for housing and food insecurity needs. That figure is worth sitting with, because it represents a real operational benchmark for what is achievable when a referral platform is paired with disciplined follow-up.

NowPow, meanwhile, continued to maintain its NYC community resource directory through a bi-directional data-sharing partnership with the Greater New York Hospital Association's HITE site, serving nearly 250 community-based organizations. The lesson from this arrangement is that platform choice matters less than the discipline of the workflow around it. A closed-loop referral system is not a software purchase; it is a set of organizational habits that include confirming the receiving agency received the referral, contacting the client within an agreed window to ask whether they were served, and recording the result in a way that future staff can interpret.

1. Confirm receipt within a defined window. A referral that is not acknowledged within 48 to 72 hours should trigger a follow-up, not a wait-and-see approach. The window itself matters less than the consistency: pick a number and build it into your workflow.

2. Contact the client directly. Assume that even the best-intentioned receiving agency will lose a small share of referrals, and design your workflow to catch those losses rather than absorb them. A brief call or text to the client after the expected intake date is the single most reliable way to surface a breakdown.

3. Record the outcome in structured fields. Free-text notes are easy to write and difficult to aggregate; a few structured fields (status, date of contact, result) make your directory auditable and your program evaluation possible.

4. Review the closed-loop rate monthly. An 86% benchmark is achievable, but only if the gap between that number and your current performance is visible to the team. Share the rate in your monthly staff meeting and make it a standing agenda item.

For agencies coordinating across boroughs, the practical implication is that the closed-loop discipline should extend to the relationships your organization has with peer providers, not just to the platform you use. A monthly check-in with two or three partner agencies, focused on shared cases, will surface referral breakdowns faster than any audit. The conversations that matter most are the ones where someone says, "We got three referrals from you last month and only one showed up — what happened?" That question, asked consistently, is what turns a referral list into a functioning network.

Navigating Official Channels: Accessing DHS Case Files and Verification Protocols

When a client comes to you with a question about their housing history, the pathway to official verification is specific and worth knowing in detail. Clients who are working with the DHS PATH intake center and need to verify their housing history must submit documentation to [email protected] or by fax to 917-637-7756, and they can request their full case files via [email protected]. These are not the kind of details that survive a casual glance at a city website, and the fact that they are not widely known among smaller community organizations is part of the reason data silos persist.

For aging-related housing resources, the verification cadence is more structured. The New York State Office for the Aging requires that Area Agencies on Aging and Independent Living Centers submit a monthly attestation confirming that they have reviewed and updated their NY Connects Resource Directory listings. This creates a documented review cycle that is more rigorous than what most community resource databases maintain, and it means that when you pull a listing from the NY Connects directory, there is a structural expectation — backed by a formal attestation process — that someone at the relevant agency has looked at that entry recently. It is not a guarantee that every detail is current to the day, but it is a layer of accountability that most directories simply do not have.

Knowing which directories carry formal review requirements, and which do not, is itself a navigational skill, and it is one I encourage every program director to build into their team's intake training.

The agencies that hold these directories well — DHS, the NYS Office for the Aging, NYC Health + Hospitals — tend to publish their procedural details in places that are easy to miss. Building a one-page internal reference document, with the right email addresses, fax numbers, and update cadences, gives your front-line staff a way to verify housing resources without having to reconstruct the pathway each time. That document should include a note about which directories have attestation cycles and which do not, because the presence or absence of a formal review requirement is one of the most reliable signals of how much trust you should place in a given listing.

There is a broader point here about the relationship between verification and advocacy. Every time a nonprofit navigator has to send a fax to verify a client's housing history, or wait days for an email response that may or may not arrive, the friction is a reminder that the systems holding our clients' data were not designed for the speed at which housing crises unfold. Part of our role as a sector is to push — through City Council testimony, through coalition letters, through the quiet persistence of individual program directors calling their agency counterparts — for the kind of integrated, real-time access that would make these verifications routine rather than heroic.

A Strategic Roadmap for Directory Stewardship

The work of verifying housing assistance data in New York City is not glamorous, and it will never be finished. The city's nonprofit ecosystem is too large, too distributed, and too dependent on the operational choices of hundreds of individual organizations for any single directory to be both comprehensive and current. What we can do, as individual providers and as a sector, is build the habits that make our internal references reliable enough to act on, even when the larger infrastructure remains imperfect.

My suggestion, drawn from years of working alongside housing navigators across the boroughs, is to treat your directory as a living system rather than a reference document. That means reconciling at least two external sources before adding a new program, scheduling a monthly review of your most-used entries, separating program data from client data at the storage level, and building a short list of official channels — DHS, NY Connects, HITE, your local AAA — that you can reach directly when verification matters. It also means accepting that the 22.1% return-to-shelter rate for unsubsidized single adults within one year is not just a statistic; it is a reminder that the gap between a referral and a stable housing outcome is wide, and the accuracy of the directory at the moment of referral is one of the few variables we can actually control.

If you are reading this and feeling the weight of how much there is to maintain, I want to leave you with something I have come to believe firmly: the practitioners doing this work are not failing the system. The system is failing them, and the most useful thing any of us can do is build our corner of it with as much rigor as we can muster, share what we learn, and keep asking city and state agencies for the API integrations and attestation standards that would make this work lighter. The people on the other side of these directories deserve nothing less.