Data Standards
Premium changes and coverage issues are kept separate.
Last updated June 24, 2026. These standards explain what RateReceipt collects, what it avoids, and why public summaries use privacy thresholds.
Coverage issue reports
Coverage issue reports include three separate categories: nonrenewal notices (where the insurer does not plan to continue the policy at the next renewal), coverage restriction notices (where the insurer is reducing what the policy covers), and deductible increase reports (where the deductible amount or structure changed at renewal).
These reports do not require a premium figure. A homeowner who received only a nonrenewal notice without a new premium can still contribute a useful data point by submitting a coverage issue report. The report records the state, the renewal month, and the type of issue — nothing else.
Coverage issue reports are displayed as separate counts for each state, not as percentages of change. They are not mixed into the premium median calculation. A state with a high number of nonrenewal reports may show a normal-looking premium median, and that separation is intentional — both signals are informative on their own terms.
Why separation matters
If nonrenewals were folded into a single premium change number, the data would be harder to interpret. A state where many homeowners are being nonrenewed and only the higher-risk remaining policies are renewing could show a premium median that looks stable, even though the market is under significant stress. Keeping the signals separate makes both more readable.
Similarly, a homeowner who received a deductible increase without a premium increase has experienced a real cost change — they now face a larger out-of-pocket amount after a covered loss — even though their annual premium did not move. The deductible increase counter captures that signal without distorting the premium median.
The three categories together — premium changes, nonrenewals, and deductible or coverage changes — give a more complete picture of a state's homeowners insurance market than any one metric alone.
Privacy threshold
State-level premium medians and coverage issue counts are withheld from public display until at least 10 qualifying reports of that type are available for a state in the current tracker period. This threshold exists for two reasons.
First, a very small sample can be misleading. A state median based on two or three reports might reflect unusual situations rather than a genuine trend. Withholding the number until a minimum threshold is met reduces the risk of readers overinterpreting a small, unrepresentative sample.
Second, a very small sample is more likely to reflect a small, identifiable group of households. Requiring 10 reports before publishing reduces the probability that the published median could be narrowed to a specific neighborhood or community even through indirect means.
The threshold is applied per report type and per state. A state may have enough premium reports to publish a median but not enough coverage issue reports to publish a nonrenewal count, or vice versa. Each type is gated independently.
Outlier handling
Reports with premium changes below minus eighty percent or above plus three hundred percent are treated as potential outliers and held out of public aggregate calculations pending review. These thresholds were chosen to exclude changes that are most likely the result of data entry error, a change in coverage scope rather than a pure rate change, or a switch from one type of policy to a very different one.
A homeowner who reduced coverage significantly and then compared a very small premium to a previous full-coverage premium might show a change of minus ninety percent that does not reflect a market rate change. Similarly, a report of a plus five hundred percent increase might reflect a policy that was previously misquoted or covers very different terms. Excluding these from the median calculation reduces noise without discarding them entirely.
Outlier reports are stored but not published in the public aggregate. Moderators can review outlier submissions through the administrative interface to determine whether they should be included (if they reflect a genuine extreme change) or excluded (if they appear to be errors). A moderator decision to include or exclude a specific report requires the admin token and is logged.
What RateReceipt does not collect
The tracker is designed around minimum data collection. The following categories of information are not requested, not accepted in submissions, and not stored:
- Full name or household identity
- Email address or phone number
- Street address or ZIP code
- Insurer name or policy number
- Claim history or claim details
- Mortgage lender or loan information
- Uploaded policy documents, declarations pages, or renewal notices
- Income, credit, or financial information
Each submission is associated with a randomly generated removal code — a string of characters that the submitter can save and use later to delete their report. No account, login, or persistent identifier is created. The removal code is the only connection between a submitted report and any future action on that report.
Aggregate product events (for example, whether the submit button was used, or whether the state dropdown was changed) are counted anonymously without attaching a user identifier. These counts are used to understand how the site functions, not to profile individual users.
Report removal
Any submitted report can be removed at any time using the removal code issued at submission. The removal process requires only the removal code — no identity verification, account, or explanation is needed.
When a report is removed, it is deleted from the database. If the report was included in a published aggregate, the aggregate figure will be recalculated at the next update cycle. Removed reports are not retained in any backup that feeds public-facing calculations.
If a homeowner submitted a report and lost their removal code, they can contact privacy@ratereceipt.com for assistance. Because there is no personal identifier stored with reports, the assistance process involves matching against the data fields submitted (state, renewal month, premium amounts) rather than an account lookup.
Data limitations
RateReceipt's tracker data has several important limitations that readers should understand before drawing conclusions from the published summaries.
The data is self-reported and unverified. Reports reflect what homeowners entered, not figures audited against actual policy documents. A homeowner who mistakenly entered an annual premium as a monthly figure, or who confused their total escrow payment with their insurance premium, will produce a report that does not accurately reflect their insurance change. The outlier thresholds reduce the impact of extreme errors, but moderate inaccuracies are not detectable.
The sample is not random or representative. Homeowners who use RateReceipt are self-selected — they typically received a renewal notice significant enough to prompt them to look for information about it. States with more reports may have more engaged homeowners, more dramatic renewals, or more effective referral channels, none of which are correlated with the true underlying market in a predictable way.
State-level medians conceal within-state variation. Florida's median encompasses everything from coastal Broward County to inland North Florida, which have very different risk profiles and insurer availability. A state figure is a blunt instrument that should prompt further investigation rather than serve as a precise benchmark.
For these reasons, the published summaries are described as "community signals" rather than statistics. They are intended to give homeowners a rough sense of whether their renewal experience appears typical or unusual, and to prompt informed questions to their insurer or agent — not to replace professional advice or independently audited market data.