Skip to content
RatePerch — real rates, real people, real insight
About

We built RatePerch because mortgage pricing is opaque on purpose.

Every other comparison site monetizes by selling your contact information to lenders. We don't. RatePerch is a public dataset of real borrower-submitted mortgage offers, published with every data source cited and every methodology decision documented. The rest of this page is how that works and why you can trust it.

An independent transparency project

What we're trying to fix

Mortgage rate shopping is broken. Lenders publish "starting at" rates designed to get you to call. Comparison sites display rate teasers and then sell your information to the highest-bidding loan officer, who calls you back with a different rate. The borrower has no way to know what a fair quote looks like for someone with their credit profile, in their state, this week. The information asymmetry costs American borrowers an estimated $15-30 billion per year in overpaid interest.

RatePerch is our attempt to fix the asymmetry from the borrower side. Every offer in our dataset was submitted by a real person who actually received it. We aggregate them transparently so the next borrower can walk into negotiation knowing exactly what other people like them are paying.

How the data works

Collection

Every offer in our public dataset is submitted by a real borrower through our submission form. Submissions are anonymous by default — names and emails are never displayed publicly. Submitters can optionally upload their Loan Estimate (the standardized document every lender must provide within 3 business days of application), which we OCR and use to stamp the submission as verified.

Validation

Every submission runs through automated fraud screening: implausible rates are rejected, APR-below-rate is rejected, statistical outliers against the same credit-and-loan-type segment are flagged for human review. Suspicious submissions land in a moderation queue that we review by hand. We'd rather lose a real submission to over-zealous filtering than publish a fake one.

Aggregation

Headline numbers (medians, percentile bands) are computed from approved, non-deleted submissions in the last 60 days. We don't adjust for points, lender credits, or special promotions because those distort the picture for borrowers comparing standard offers. We disclose sample sizes everywhere a number appears.

Credit union eligibility

We import the complete NCUA list of federally insured credit unions (4,000+) and classify each by membership type using public NCUA Call Report data. State-residency CUs appear on state pages where eligible; SEG-restricted CUs are hidden unless a user submits a community correction (which we then verify before publishing).

What we don't do

We make these commitments because the things in this list are exactly how every other mortgage comparison site quietly compromises borrower trust.

  • We don't sell your contact information to lenders.No matter how much they offer. It's a common revenue model in the rate-comparison category; we've declined it on principle.
  • We don't take affiliate fees that influence what appears.Lender rankings, credit-union ordering, and rate calculations are computed from data alone. No paid placements, ever.
  • We don't hide our methodology. Every number on this site has documented sources and a sample size. If you want to verify, you can.
  • We don't practice as mortgage lenders, brokers, or financial advisors.RatePerch is informational. We provide data; you make your own decisions. Always consult a licensed professional for your specific situation.

Where the data comes from

We use four primary sources, all public or borrower-contributed. Click through to verify.

Borrower-submitted offers
Crowd-sourced from RatePerch users via our submission form. Verified subset includes Loan Estimate documentation.
Freddie Mac Primary Mortgage Market Survey (PMMS)
The official weekly U.S. average mortgage rate, published every Thursday. We pull it via the Federal Reserve Economic Data API and display it alongside our borrower medians for context.
NCUA Credit Union directory
All 4,000+ federally insured credit unions, sourced from the NCUA Call Report dataset (quarterly public records). Eligibility classification uses the NCUA-published field-of-membership codes.
Credit union posted rates
Pulled manually from each credit union's public rate sheet on the effective date shown. Not adjusted for points, fees, or special promotions.

Who runs this

Conor Gaughan
Founder

I started RatePerch after going through the mortgage process myself and realizing it's nearly impossible to know whether the rate you're being quoted is fair. I'm not a loan officer or a mortgage industry insider — I'm a regular borrower who got tired of getting wildly different quotes from credit unions, banks, and online lenders with no way to compare them apples-to-apples. RatePerch is the tool I wished existed when I was shopping. Every methodology decision on this site is published, every data source is cited, and I cover what I'm learning week-to-week on the @RatePerch YouTube channel.

Reach out: hello@rateperch.com · YouTube

How we make money

We're in the early stages of building a verified loan-officer marketplace where NMLS-licensed mortgage professionals pay a flat per-bid fee to submit competing offers against opt-in borrower submissions. Borrowers always opt in voluntarily and can opt out at any time. The economic alignment is that loan officers compete to give borrowers a better rate — not that we sell borrowers' data.

We don't take affiliate fees from lenders. We don't take per-close referral fees. We don't accept paid placements that influence any ranking on the site. The full rules are documented in our disclaimer and on our methodology page.

Contact

General questions
hello@rateperch.com
Data corrections
data@rateperch.com
Press inquiries
press@rateperch.com
Loan officer questions
Apply to bid →

We aim to respond within 2 business days. Found an error in our data? Tell us — we read every correction email and update within 24 hours when warranted.