HomeIQ Methodology
How the HomeIQ CMA is built, why it is defensible, and why a licensed Florida broker signs every number.
The Core Differentiator
Meydomo’s patent-pending property score
Most valuations fail at the negotiating table because they measure two things loosely: where a property actually sits, and what condition it is really in. HomeIQ measures both precisely — and applies the same measurement to your home and every comparable.
The patent-pending method scores every property across approximately 16 value-relevant dimensions: location measured boundary-to-boundary from the actual lot lines (covering water adjacency, amenity proximity, transportation exposure, easements, and adverse adjacency) and condition judged from real listing photos on a six-point scale from like-new to needs-major-repairs — the same scale Fannie Mae requires of licensed appraisers.
Because your home and every comparable are scored the same precise, patent-pending way, the resulting value is more accurate and far harder to argue with at the negotiating table. That precision is the moat. No MLS and no commercial AVM does this.
The four pillars of a signable number
Every HomeIQ CMA is a Broker Price Opinion a licensed Florida broker signs (Jose Martin, BK3177826, FL §475.25(1)(s)). A signature is a professional putting their license behind a number. That is only possible when the number is four things at once:
Consistent
The same method runs on every property. Two similar homes are never treated differently by accident or mood. The model is deterministic and re-runs statewide every night.
Accurate
Measured from the parcel's true footprint using our patent-pending boundary measurement method, and priced from its own comparable sales — not a regional average or a ZIP centroid.
Defensible
Every line traces to a measured distance, a cited decay curve, a comparable sale, and Fannie Mae §B4-1.3.
Rational
The number concludes what a careful person standing in the front yard would conclude. The method replicates how an appraiser weighs a location, but does it consistently and on the record.
Location — measured from your lot lines, not a centroid (patent pending)
Edge-to-edge, not center-to-center
How far is your home from the lake? Most valuation models measure from the center of your parcel to the center of the feature, which can report a five-acre waterfront estate as “two hundred meters from water” because its center sits well back from the shoreline it physically touches.
HomeIQ's patent-pending method measures the shortest line between your parcel's boundary and the feature's boundary — the single closest pair of points, one on each shape. If the two shapes overlap or touch, the distance is zero and you front the feature. The engine evaluates boundary segment pairs and returns the minimum on the curved surface of the Earth, in meters (geodesic measurement on the WGS-84 spheroid using PostGISST_Distance(geom::geography, feature::geography)). A waterfront lot reads as waterfront. A home one street back reads as one street back. A parcel whose boundary kisses a park reads as park-fronting even if the house sits at the far end of the lot.
Each force has its own decay curve
A disamenity does not stop at the property line — it fades. A highway's noise attenuates roughly three decibels per doubling of distance. A small park's charm vanishes when you turn the corner. HomeIQ models each effect as an exponential decay:
λ (lambda) is the “decay length” — the distance over which the force drops to about a third of its peak. Each feature type gets its own λ: a highway carries far (long λ). A small park's charm is hyper-local (short λ). A waterfront view evaporates the instant a roofline blocks it (the shortest λ of all). Each λ is set from the published research for that kind of feature — highway noise, lake frontage, golf course, transmission line — each is a literature of its own.
Your property's locational adjustment is the sum of every feature's curve, evaluated at your parcel's true edge-to-edge distance. The dollar amount is then re-anchored to your own comparable sales, so the curve sets the shape and the market sets the magnitude.
The model re-tunes itself every night — re-deriving across the state and checking the realized distribution against expected distributions. Two knobs govern it: τ (materiality, how strong a feature's pull must be before it earns a headline) and λ (the reach of each feature type). It is a control loop: measure, compare, adjust, converge. The number on your report tomorrow is calibrated to the market as it stood tonight.
Condition — assessed from actual listing photos
Condition is the single largest variable in home value. A maintained home and a gutted home on the same street can differ by over $130,000. Most valuation tools ignore condition entirely or proxy it from the year built or a price ratio — both of which are circular and produce biased adjustments.
HomeIQ rates every comparable's condition from its actual listing photos using the same engine applied to your home. The engine places each property on a six-point condition scale — from like-new to needs-major-repairs — matching the standard Fannie Mae requires of every licensed appraiser, and it returns a continuous score alongside a defect list so gradations within a tier still register in the adjustment. The subject and every comparable are rated by the identical engine — per Fannie Mae §B4-1.3-06, which requires an absolute basis for the condition adjustment. A different model on the subject would bias its position on the comp-extracted scale and invalidate the adjustment.
The continuous score (not just the discrete tier) is the value axis, so a cosmetic improvement that does not move a full tier still registers in the comp adjustment. Condition is then incorporated as a predictor in the same controlled multivariate OLS that derives the GLA, age, and waterfront coefficients — isolating its marginal price net of every other physical difference.
Comparable selection — Fannie Mae neighborhood-first
Fannie Mae §B4-1.3-03 and -08 require that comparable sales be drawn from the subject's own neighborhood or competitive market segment first. The segment is bounded by real dividers — streets, waterways, freeways, desirability lines — not a radius and not a comp-count target.
HomeIQ selects comparables through a resolved segment key in this order:
- Subdivision or plat — parsed from the legal description. Per Fannie Mae, comps from the same subdivision or project are preferred.
- Census tract — the populated bridge that beats ZIP or radius when subdivision data is unavailable.
- Radius ladder — 1.5 → 3 → 5 → 8 → 12 km as the last resort, never lifting to unlimited and never reaching into a non-adjacent price tier.
Location is controlled by selection — not adjustment. As long as every comp sits in the subject's own segment, the location line is zero. A competing-area comp is admitted only with an explicit location delta derived from the submarket fixed-effects model.
The data infrastructure
HomeIQ is powered by a unified property database that does not exist anywhere else outside the licensing catalogs of the data oligopoly — and it was built from free public sources, on a durable natural key, with a proprietary valuation signal layered on top.
11.05M
Florida parcels in one table — unified from 67 county schemas. One identity model, not 67 incompatible databases.
95.2%
Of parcels carry actual boundary geometry — a MULTIPOLYGON, not a centroid. The real shape of the real lot.
10.4M
Parcels scored with Meydomo's patent-pending property score — measuring locational externalities across ~16 feature types, boundary-to-boundary.
167
Attributes per parcel — geometry, value prior, sales history, permits, taxes, school zones, and more.
The database is built from free public sources only — no ATTOM, CoreLogic, Regrid, or Estated licensing. Those run into the millions of dollars per year and restrict redistribution. Our marginal cost per parcel is compute, not license. This is both a cost advantage and a freedom-to-operate advantage.
Data by SparKLAWThe spine includes statewide sales and deeds, building permits at both county and city level, code-enforcement cases, county clerk legal actions, 365K reference geometries (coastlines, water bodies, freeways, rail, golf courses, parks, transmission lines), and 20+ geo-context layers — Census ACS, BLS employment, IRS migration data, HUD Fair Market Rents, FHFA House Price Index, and school records with boundary geometry.
The architecture is national by design. Florida is the reference implementation — 11M parcels, ~6.5% of the national target. The same machinery (data klaws, curators, bulk-load doctrine, big-box migrations) is designed to onboard every other state through the same repeatable process.
Broker sign-off
Every HomeIQ CMA is a Broker Price Opinion prepared pursuant to Florida §475.25(1)(s) and signed by Jose Martin, Qualifying Broker, License BK3177826. Stage-1 does not include an interior inspection. Comparable condition is assumed at average, move-in ready condition unless listing photos or county records indicate otherwise.
The broker signs only when the four pillars are met — consistency, accuracy, defensibility, rationality — and only when the methodology traces to Fannie Mae §B4-1.3 at every line. Every adjustment is market-derived; no magic constants, no rule-of-thumb adjustments, no floor substituted for the comp reconciliation.
The same methodology feeds all three HomeIQ reports — Price Check, Gameplan, and Outlook — because consistency across the report family is the product. If they disagreed, the broker could not sign any of them.
Sources
- Fannie Mae Selling Guide §B4-1.3-03 (Neighborhood)
- Fannie Mae Selling Guide §B4-1.3-06 (Property Condition)
- Fannie Mae Selling Guide §B4-1.3-08 (Comparable Sales Selection)
- Fannie Mae Selling Guide §B4-1.3-09 (Adjustments to Comparable Sales)
- Fannie Mae Selling Guide §B4-1.3-11 (Sales Comparison Approach Reconciliation)
- UAD Condition & Quality Rating Definitions v1.0 (Apr 29 2025)
- Florida Statute §475.25(1)(s) — Broker Price Opinion authority
- Anselin, L. (1988) Spatial Econometrics: Methods and Models
- Brunsdon, Fotheringham & Charlton (1996) — Geographically Weighted Regression
- Goodman & Thibodeau — submarket water-view premium range (5%–54%) vs. pooled +18%
See the methodology in action.
Your HomeIQ CMA is one click away. $4.95. About 60 seconds. The work an appraiser does in two weeks, pre-computed and ready for you.
Get Your HomeIQ Report →