What changed: from your claims to a predicted score

Underwriting used to lean mostly on what had already happened: your own claims, the home's age, the construction type. Insurers now lean harder on what their models predict will happen. They buy property-level data, run catastrophe models for wildfire, wind, hail, and flood, and increasingly pull aerial and drone imagery to read a roof's condition or the brush around a house. The output is a risk score that helps set your premium and decide whether to write or renew at all. A Wall Street Journal review of the market in May 2026 noted carriers using drone footage to calculate a home's risk score, a practice that is driving up costs for some owners.

What feeds the score

  • Your loss history and the property's. Prior claims on you and on the address, drawn from the C.L.U.E. database (below). Even inquiries and small claims can register.
  • Aerial and drone imagery. High-resolution photos read for roof wear, missing shingles, tarps, tree overhang, pools, and trampolines. Some carriers re-image properties between renewals.
  • Catastrophe models. Wildfire, flood, wind, and hail scores for your exact location, often from third-party model vendors, increasingly granular down to the parcel.
  • The home itself. Roof age and material, year built, square footage, systems, distance to a fire station and hydrant, and the rebuild-cost estimate.
  • Neighborhood and regional loss trends. How the wider risk pool around you has performed; a bad local run of losses repriced your slice even if your home is fine.

Why your premium can rise with no claim

Because the score is forward-looking, none of it requires you to have filed anything. A model can decide your roof is near end of life from a photo, or that your parcel's wildfire score climbed after a nearby fire season, and your premium follows. That is also why two near-identical homes on the same street can be priced differently: the inputs, not just the address, drive the number.

How to see and dispute the inputs

You cannot argue a model down, but you can fix what it is reading. The inputs are the lever:

  1. Pull your C.L.U.E. report. The Comprehensive Loss Underwriting Exchange, run by LexisNexis Risk Solutions, is the claims-history database carriers query. Under the federal Fair Credit Reporting Act you can request it free once every 12 months. Read it for claims that are not yours, duplicates, or a minor inquiry logged as a paid loss.
  2. Correct errors in writing. Dispute mistakes with LexisNexis (the FCRA requires they investigate) and ask your insurer to re-rate once the record is fixed.
  3. Send evidence of work the model has not seen. A new roof, cleared defensible space, a replaced electrical panel, or a water shut-off device can all move the score. Photos and dated receipts give the underwriter something to override stale imagery with.
  4. Ask what imagery was used. If a non-renewal cites roof condition from an aerial photo, ask to see it and supply a current roof certification if the photo is wrong or out of date.

If the score prices you out

A high enough model score can lead a carrier to non-renew or decline outright. If the standard market keeps saying no, your state's FAIR Plan is the insurer of last resort; the state directory shows what each plan covers, its dwelling cap, and how to apply. And if your renewal simply jumped rather than disappeared, the premium-jump walkthrough covers what to check and how to shop without setting off a wave of sales calls.

This explainer describes how home insurers generally use predictive risk scoring and property imagery; specific data sources, models, and dispute rights vary by carrier and by state. Confirm your rights and the current rules with the carrier, LexisNexis, or your state Department of Insurance before relying on the detail above. Page last reviewed 2026-06-01.