CDFI
We exist to bring trust back into the places it's gone missing. For CDFIs and community lenders specifically, that mission and yours are already the same one.
Who We Are
Two organizations, two roles. Worth knowing before anything else here.
A Washington, DC 501(c)(3). The mission side — this wasn't built to maximize a sales metric.
Our research and methodology partner, led by Dr. Bob Raleigh. Builds and validates the instrument itself.
Disinvestment doesn't announce itself. Census data can tell you a tract is poor. It doesn't tell you whether that tract trusts you enough to answer the door. That's the gap we exist to close.
Your Target Market Is Public Record. Where You're Actually Reaching Is Not.
Two tracts can clear the same CDFI Fund poverty or income threshold and still be entirely different lending environments — one where your institution already has standing, one where it doesn't. Nothing in your certification file tells you which is which today.
Your Target Market
A matter of public record, drawn at census-tract level under CDFI Fund criteria — poverty, unemployment, and area median income benchmarks you already track.
Which Tracts You're Structurally Reaching
The variable behind a shortfall you usually only see in hindsight, at your next Annual Certification Report.
Two Tools, One Instrument, Built at Your Own Unit
Everything below runs at census tract — the exact geography your CDFI Investment Area already uses. No translation layer, no proxy geography, no second system to maintain.
Where the Disconnect Is
The tested, validated core: which of your qualified tracts are most structurally disconnected from institutional trust — before your next Annual Certification Report finds them for you.
Access Isn't the Same as Income
Physical and institutional access doesn't track with poverty the way most targeting assumes. Some of your poorest tracts may be your most reachable — this shows you which. Whether that difference changes response rates for your institution is something we'd want to test with you, not assert.
A Lab, and a Partner in the Outcome — Not a Vendor Selling a Report
We Show Our Work Before We Ask for Anything
We run this as a research lab, not a locked-in vendor relationship: we predict what we expect to find, write it down before we look, and report the result honestly whether it holds or not. That's not a caveat — it's the actual discipline behind every claim in this document.
For your institution, that means the first move is ours: we'll read your own already-public Target Market, at no cost, and hand back a floor-identification result you can check against what your own lending officers or coaches already know about those tracts. If it holds up on ground you already know, we keep going together. If it doesn't, you've lost nothing and we've learned something.
Kept Honest, on Purpose
Floor Identification is validated out-of-sample. In a sealed, pre-registered test ahead of Wisconsin's August 2026 primary, we also predicted our measure would rank six neighborhoods by turnout, from most to least connected. It didn't — it identified the floor correctly and got the middle and top wrong. That's why we sell floor identification, and not ranking.
Reachability's downstream effect on response rates is not yet proven. We can show the access difference exists, independent of poverty. Whether it moves outcomes for your institution is a real question — the free pilot read is how we'd answer it together, not assert it up front.
Built to Stay a Community Tool, Not a Lending Tool
We built this to inform where an institution builds relationships and channels outreach — never who to lend to. That's not a limitation we're working around; it's a line we designed the instrument to respect, because place-level inference applied to credit decisions is redlining's exact mechanism, and we're not in that business. It's also why this pairs cleanly with your existing compliance process instead of complicating it.
You're Already Certified to Serve These Tracts.
We can show you which ones you're structurally reaching, on your own ground, at no cost for the first read. If it holds up against what your team already knows, that's worth a real conversation.
Start with a Free Read