How the renewal tracker works.

Inputs

A premium contribution contains four fields: a supported U.S. state, a renewal month in the current tracker year, last year's annual premium, and this year's new annual premium. These four fields are the only data elements the tracker records for a premium submission.

An issue contribution contains a supported state, renewal month, and issue type — either nonrenewal, coverage restriction, or deductible increase. Issue contributions do not require a premium figure. A homeowner who received a nonrenewal notice and no renewal quote can still submit a signal.

We deliberately do not request ZIP code, street address, insurer name, policy type, policy number, claim history, or uploaded policy documents. Excluding these fields keeps the submission useful as a community signal while significantly reducing the amount of information that could identify or narrow a specific household.

The annual premium fields accept the total annual premium shown on the renewal notice — the insurance cost only, not the total mortgage payment. If a homeowner pays insurance through escrow, we suggest using the annual insurance premium from the declarations page or renewal notice, not the total monthly mortgage payment multiplied by 12, because the mortgage payment includes property taxes, reserve adjustments, and other items unrelated to the insurance change.

Calculation

The dollar change is the new annual premium minus the prior annual premium. The percentage change is the dollar change divided by the prior annual premium, rounded to one decimal place for the individual receipt and to the nearest whole number for the published state median.

The individual receipt is shown immediately after submission and is private to the submitter. It shows the homeowner's own dollar change, percentage change, and renewal month so they have a record of what they entered. The receipt does not include state-level comparisons until the state's privacy threshold is met.

Calculations are performed server-side at submission time and stored with the report. The percentage is recalculated if a report is later reviewed or corrected through the administrative interface, but the raw input fields (old and new premium) are what drive all downstream calculations.

Publication threshold

A homeowner's personal receipt appears immediately after submission. A state-level median premium change, report count, or issue count is displayed publicly only once at least 10 qualifying reports of that type are available for that state in the current tracker period.

The threshold applies separately to premium reports and issue reports. A state with 10 qualifying premium reports but only 4 nonrenewal reports will show a premium median but not a nonrenewal count. Each type of data must independently meet the threshold before it is published.

The threshold serves two purposes. First, it reduces the chance that a published figure is heavily skewed by a very small, unrepresentative sample. A median from 3 reports is far less reliable than a median from 30 or 300 reports. Second, it reduces the risk that a published figure could be narrowed to a small group of identifiable households in a thinly populated state or market.

States that have not yet reached the threshold display a "Building" indicator rather than a number. This is intentional — it tells visitors that data exists but is not yet sufficient to publish responsibly, rather than showing a potentially misleading figure based on one or two reports.

Quality rules

Several automated checks run at submission to reduce noise and prevent manipulation.

Duplicate submissions from the same browser are blocked using a local browser flag. If a homeowner submits a report and then reloads the page and tries to submit again, the second attempt is rejected. This is a soft block using browser storage rather than a hard account-based block, but it prevents accidental double submissions and makes casual repeat submissions more friction-heavy.

Submissions with identical state, renewal month, and premium values as an existing record are flagged for moderator review rather than automatically accepted. Two identical reports can legitimately exist — two neighbors with identical situations — but a pattern of identical submissions is a warning sign of synthetic data.

A bot-detection layer is active in the submission flow. In the production environment, Cloudflare Turnstile provides an invisible challenge layer that rejects automated submissions without requiring homeowners to complete a visible CAPTCHA. Bot-trap fields in the submission form are also used to reject non-human submissions that fill hidden fields.

Outlier handling

Reports with premium changes below minus eighty percent or above plus three hundred percent are treated as outliers and held out of public aggregate calculations. These thresholds were chosen to capture the range of plausible market-driven changes while excluding values that are most likely errors or policy-type changes rather than pure rate changes.

A change of minus eighty percent or more would mean a homeowner's annual premium dropped from, for example, $3,000 to $600 or less. While significant discounts happen, this magnitude almost always reflects a major coverage reduction, a policy type switch, or a data entry error — not a market-wide rate drop. A change of plus three hundred percent would mean a premium more than quadrupled, which is at the outer edge of documented market events even in the most stressed states.

Outlier reports remain stored in the database and continue to be listed in the submitter's personal receipt. They are not deleted. If a moderator determines that an outlier is genuine — a homeowner in a highly distressed market who actually received an extreme notice — it can be reclassified and included in the aggregate through the administrative interface. The default behavior is exclusion until reviewed, not deletion.

Outliers are also separated so they can be studied as their own signal. A cluster of extreme-high-change reports in one state over a short period might be meaningful even if they are excluded from the median calculation. Future versions of the site may display outlier patterns separately from the primary median.

Report types and how they interact

RateReceipt separates three types of market signals: premium change reports, nonrenewal reports, and coverage restriction or deductible change reports. These signals interact in ways that are easy to misread if they are mixed together.

In a deteriorating market, some homeowners are nonrenewed while others face steep premium increases. If nonrenewals were averaged into the premium median alongside renewals, the median might look more moderate than it is — because the homeowners facing the worst outcomes were removed from the renewal pool. By tracking nonrenewals separately, the site shows both the premium experience of those who were renewed and the availability experience of those who were not.

Coverage restriction and deductible increase reports capture a third dimension: the market is not just about whether coverage continues and at what premium, but also whether the terms are changing in ways that shift risk from the insurer to the homeowner. A deductible that doubles from $1,000 to $2,000 is a real cost change even if the annual premium stays flat.

Why the median is used instead of the average

The published state figure is the median percentage change across qualifying reports, not the arithmetic average. The median is used because it is less sensitive to extreme values than the average.

If 9 homeowners in a state reported a 15% increase and 1 homeowner reported a 400% increase (an outlier that passed the threshold but was an extreme value), the average would be around 54% — far from the typical experience — while the median would be 15%, which correctly reflects what most homeowners in the sample experienced.

In a small sample, even one extreme value can significantly distort an average. The median is more robust in this situation. As sample sizes grow, the difference between median and average typically narrows, and future versions of the site may display both alongside a confidence range.

Removal process

Any submitted report can be deleted at any time using the removal code issued at submission. The removal code is a randomly generated string shown on the individual receipt screen after a successful submission. Homeowners are encouraged to save their removal code.

To remove a report, the homeowner enters their removal code at the removal endpoint. No other information is required — no email, no account, no identity verification. The code is the credential. When the removal is confirmed, the report is deleted from the database. If the report contributed to a published state aggregate, the aggregate is recalculated at the next update.

Removal codes are hashed before storage. The site does not store the original removal code in plain text — it stores a hash that can verify a code but not reveal it. This means that if the database were compromised, the removal codes themselves would not be directly exposed.

Limitations

This is an opt-in, self-reported sample, not a scientific survey, a statistically representative panel, or a verified insurance rate filing database. The participants who use RateReceipt are not a random sample of all U.S. homeowners. They are homeowners who received a renewal notice significant enough to prompt them to seek information, found the site, and chose to submit a report. That selection effect almost certainly biases the sample toward larger-than-typical changes.

Coverage differences materially affect premiums, and the tracker does not capture those differences. A homeowner who doubled their dwelling limit and saw their premium increase 20% is in a very different situation from a homeowner whose coverage stayed flat and whose premium increased 20%, but both contribute the same signal to the tracker. Interpreting the median as a pure rate-increase signal requires treating coverage as stable across the sample, which is an assumption that does not always hold.

State boundaries are coarse. The Florida median encompasses coastal Southeast Florida, the interior, the Gulf Coast, and the Panhandle — regions with dramatically different risk profiles, insurers, and market conditions. A single state figure tells you that something happened in a state, not where or to whom.

The tracker should be read as a community signal: a starting point for asking informed questions, not a substitute for professional advice, rate-filing data, or independently audited market research. Homeowners with concerns about a specific renewal notice should consult their insurer, a licensed agent, or their state insurance department.