A comprehensive mortgage lender scoring system using Home Mortgage Disclosure Act (HMDA) data.
Overview
Our scoring system evaluates mortgage lenders across multiple dimensions to help borrowers find lenders that best match their needs. Each score ranges from 0–100, with higher scores indicating stronger performance or expertise in that area.
Score Types
Overall Lender Score
A composite score reflecting a lender's overall performance in the mortgage market. This combines:
- Pricing. How competitively the lender prices loans (rates and fees).
- Pull-through. How effectively the lender moves applications through to closing.
- Product. Demonstrated expertise in the borrower's specific loan type.
- Place. The lender's volume and reach across different markets.
These four HMDA-derived components are combined using an equal-weight average, where each contributes the same share, for a transparent raw score that fairly represents a lender's performance across dimensions.
Promptness is our fifth dimension, and it works differently. It is behavioral rather than HMDA-derived, so it exists only once a real shopper has requested a quote from that specific lender. When a lender has been asked, promptness blends into the Overall score, its weight ramping up as more requests accumulate, so responsiveness counts without a single data point swinging the ranking. Lenders who have never been asked are not scored on promptness and are never penalized for it. See the Promptness Score section below.
Product Focus Scores
These scores measure how well a lender serves borrowers seeking specific loan types:
- FHA — Government-insured loans with lower down payment requirements
- VA — Loans for veterans and active military
- USDA — Rural housing loans
- Conventional — Standard loans not backed by government programs
- Jumbo — Loans exceeding conforming loan limits
How it's calculated. Each product score considers three factors with balanced weighting (~34/33/33):
- Best market performance — The lender's strongest local market for that product, measuring their market share where they're most competitive
- Volume (min-max scaled) — Total loan volume for that product, scaled against market min/max
- Volume (percentile rank) — How the lender's volume ranks among all lenders nationally
This balanced approach ensures that both high-volume national lenders and regional lenders with strong local market presence are fairly represented. A lender with a high FHA score, for example, demonstrates strong FHA lending through a combination of local market expertise and meaningful volume.
Purpose Scores
These scores indicate expertise in different loan purposes:
- Purchase — Loans for buying a home
- Refinance — Loans to replace an existing mortgage
The methodology mirrors product scores, using the same balanced weighting of best market performance and volume metrics.
Geographic Scores (State & Metro Area)
These scores show how well a lender serves specific geographic markets.
How it's calculated. We evaluate two components:
- Market presence — The lender's loan volume in that market (log-scaled to prevent mega-lenders from dominating)
- Market focus — What percentage of the lender's total business is concentrated in that market
These components are combined and then scored using a blend of:
- Absolute strength — How the lender compares to the market leader
- Relative participation — How the lender ranks among all lenders nationally
Why two approaches? Metro areas (MSAs) use a 50/50 blend, slightly favoring absolute market leadership since local dominance matters in tight geographic markets. States use a 40/60 blend, slightly favoring participation since states are larger and regional commitment matters more than pure volume. A high geographic score indicates a lender with meaningful presence and focus in that area — not just a large lender who happens to do some business there.
Pull-through Score
Measures how efficiently a lender guides borrowers through the mortgage process to a definitive outcome.
How it's calculated. We evaluate two aspects of lender performance:
- Origination success — How effectively the lender closes loans, adjusted for borrower characteristics
- Clear outcomes — Denials for well-defined reasons also count positively, as they show the lender's ability to give borrowers a clear answer rather than leaving applications in limbo
Why this matters. A lender's raw approval rate can be misleading — a lender serving only prime borrowers will naturally have higher approval rates. Our process score adjusts for borrower characteristics and rewards lenders who help consumers get results, whether that's a closed loan or a clear explanation of why they don't qualify.
What it signals to borrowers. A high process score means borrowers who start with this lender are more likely to close with them, suggesting a smoother experience and fewer reasons to look elsewhere. A low process score means borrowers may be more likely to abandon the process or switch to a different lender before closing.
Pricing Score
Evaluates how competitively a lender prices their loans compared to the market.
How it's calculated. We analyze a blend of two pricing components:
- Interest rate — The ongoing cost of the loan
- Upfront fees — Points, origination fees, and other closing costs
This blended approach captures the full cost of a loan — a lender offering a low rate but high fees isn't necessarily a better deal than one with a slightly higher rate and minimal fees. We then compare each lender's blended pricing to expected pricing based on:
- Loan characteristics (loan amount, loan-to-value ratio, loan type)
- Borrower profile (debt-to-income ratio, income)
- Property details (location, property type)
- Market conditions at the time of origination
Lenders who consistently offer better pricing than expected given these factors score higher. This helps identify lenders offering genuine value rather than simply serving lower-risk borrowers who naturally qualify for better terms.
Promptness Score
Measures how quickly a lender answers a real quote request sent through Safe Rate. Every lender, broker, and loan officer is expected to respond within one business day.
How it's calculated. We score the speed of a lender's first response to each request, counted in business days (weekends and U.S. federal holidays excluded):
- Within 1 business day: 100
- Within 2 business days: 50
- Within 3 business days: 25
- 4+ business days, or no response: 0
Any genuine response earns full speed credit, whether a quote or a clear "we can't quote this scenario." A lender declining to quote is still engaging with the shopper; whether they could ultimately quote is tracked separately as a quote rate and does not affect promptness. Only silence (or refusing the contact channel) scores 0. A lender's promptness is the recency-weighted average of these per-request scores.
Present when there's data. Unlike the four HMDA dimensions, promptness exists only after a real shopper requests a quote from that lender. Lenders who have not been asked carry no promptness score and are not penalized, and "no data" is never treated as a zero. Once a lender has been asked, promptness blends into the Overall score, with its influence growing as more requests are observed. Consistent non-response pulls the Overall score down.
Interpreting Scores
| Score range | Interpretation |
|---|---|
| 80–100 | Excellent — among the top performers |
| 60–79 | Strong — above average performance |
| 40–59 | Average — typical market performance |
| 20–39 | Below average |
| 0–19 | Limited presence or performance |
Important notes.
- A lender may excel in some areas but not others.
- Geographic scores are most meaningful for areas where you're looking to buy.
- Product scores help identify lenders experienced with your loan type.
- No single score tells the whole story — consider multiple dimensions.
Data Source
The four core dimensions are calculated using Home Mortgage Disclosure Act (HMDA) data, which includes virtually all mortgage applications in the United States. This public dataset ensures transparency and comprehensive market coverage. Promptness is the one exception: it is measured from real shopper quote requests delivered through Safe Rate, not from HMDA.