From AVM to Actual™: PropTechUSA.ai Research Blueprint (2025 Datasets)
Local Home Buyers USA — powered by the research of PropTechUSA.ai
Our platform converts noisy valuations into certainty-priced, as-is offers using three models: the Hyperlocal Sentiment Score (HSS), the Cost of Certainty Index (CoCI), and SpreadGuard (mortgage-spread risk). Below we publish our 2025 internal datasets (rolling 90-day and YTD views), methods, governance, and reproducible case artifacts that unify product, acquisitions, and compliance around a single standard: close with certainty.
Our 2025 Datasets (Rolling 90-Day & YTD)
These are our operational datasets. Values here reflect current internal pulls.
Coverage & Volume
| Metric | Rolling 90-Day | 2025 YTD | Notes |
|---|---|---|---|
| Markets observed (ZIP+4 cells) | 18,420 | 46,900 | HSS coverage grid |
| Seller inquiries analyzed | 7,312 | 22,804 | Deduped by address + device |
| As-is offers issued (CoCI) | 2,186 | 6,540 | Signed LOIs + formal contracts |
| Closed transactions | 418 | 1,206 | Funded, title disbursed |
| Median close time (days) | 14 | 14 | 7/14/21 day paths |
Model Inputs (Aggregated)
| Signal | Mean | p25 | p75 | How we use it |
|---|---|---|---|---|
| HSS (0–100) | 53.2 | 41.0 | 65.7 | Adjusts carry & fall-through probability |
| Repair Scope (USD) | 18,900 | 7,800 | 29,600 | Bid-backed; feeds CoCI |
| Mortgage Spread (bps) | 82 | 58 | 112 | Stress tier in SpreadGuard |
| Appraisal Δ vs Contract (%) | −1.8% | −4.5% | +0.9% | Fall-through amplifier |
Outcome Deltas vs. Financed MLS (Matched Sets)
| Outcome | Our CoCI Path | Financed MLS | Delta |
|---|---|---|---|
| Fall-Through Rate | 3.9% | 17.6% | −78% |
| Days to Close (median) | 14 | 36 | −22 days |
| Post-Close Occupancy Offered | 72% | 11% | +61pp |
Matched by ZIP+4, bed/bath, condition grade; excludes edge cases with legal encumbrances.
Models & Methods
Hyperlocal Sentiment Score (HSS)
Purpose: quantify micro-market demand pressure to modulate time-to-close and fall-through risk. Scale: 0–100 (50 neutral; >65 hot; <35 soft).
- Inputs: listing velocity, DOM deltas, price-cut frequency, absorption vs. new listings, seasonality, micro-migration indicators.
- Use: feeds carry-time penalties and probability of disruption inside CoCI and SpreadGuard.
Cost of Certainty Index (CoCI)
CoCI_Net = AVM_Baseline − RepairScope − Carry(HSS, Days) − RiskPremium(Spread, HSS, Contingencies)
- RepairScope: vendor bids + permit/variance + GC overhead.
- Carry: taxes + insurance + utilities + vacancy × expected days (HSS-weighted).
- RiskPremium: fall-through probability × re-marketing time × spread-adjusted discount.
SpreadGuard (Mortgage-Spread Risk)
Purpose: convert credit-market stress into a practical fall-through probability and discount factor.
- Inputs: primary rate vs. benchmark spread, lock volatility, denial/withdrawal rates, appraisal variance.
- Output: stress tier (low/moderate/high), fall-through probability, discount factor → piped to CoCI.
Public KPIs & Outcome Benchmarks
Rolling 90-day across certainty-priced, as-is transactions.
Matched sets, same micro-markets and condition grades.
Bridge possession aligned to move dates (where permitted).
Reproducible Case Artifacts (Proof-of-Concept)
Each closed deal creates a JSON case file and a human-readable report. Anonymized versions appear in our Cost of Certainty in Practice series.
{
"case_id": "coci----",
"avm": 325000,
"hss": 46,
"repair_scope": 22000,
"spread_bps": 82,
"risk_premium": 18000,
"coci_offer": 285000,
"path": "14-day",
"milestones": {"offer":"YYYY-MM-DD","walkthrough":"YYYY-MM-DD","title_clear":"YYYY-MM-DD","close":"YYYY-MM-DD"}
}
The renderer converts case JSON into a public article with diagrams and breakdowns.
Read a live example: Jacksonville (3bd/2ba) — AVM→Actual with CoCI.
Model Governance & Compliance
Versioning
- HSS_v2025.1 • CoCI_v2025.2 • SpreadGuard_v2025.1 (semantic; change-logs maintained).
- Auto-rollback if 14-day MAE ↑ beyond threshold or fall-through deltas widen in backtests.
Inputs Excluded
- No protected classes or proxies; no school-name, name-ethnicity, or sensitive demographic signals.
Explainability
- Every offer stores
breakdown { avm, repairs, carry, risk_premium }with links to evidence and bids.
Security
- Verified title partners; wire-fraud controls; documented escrow disbursement chain. See Inside the Title Company.
Related Research & Tools
FAQ
Are these numbers real?
Yes—metrics shown are from our internal rolling 90-day and YTD datasets.
Do you publish raw data?
We publish aggregate KPIs for privacy and compliance. Partners can access raw streams via our API under agreement.
How do you compare to financed MLS paths?
We use matched sets by ZIP+4, bed/bath, and condition grade, excluding edge cases with legal encumbrances.
Turn Research into Your Result
We’ll run HSS, CoCI, and SpreadGuard on your property and issue a clear, as-is offer with a 7/14/21-day plan.
Get Your Veteran-Focused, As-Is Offer© 2025 Local Home Buyers USA • PropTechUSA.ai Research • Updated Nov 11, 2025
Real-World Seller Insights
Fresh how-tos and market tips from Local Home Buyers USA.
