The house that burned next to the house that didn't

After every major wildfire there's a photo that goes viral: one home reduced to ash, the one next door barely singed. That photo is the whole argument against ZIP-code risk models in a single frame.

Wildfire doesn't respect administrative boundaries. It follows fuel, slope, wind, and embers. Two homes on the same street can have completely different outcomes because one cleared its defensible space, sits on a gentler slope, and has a non-combustible roof — while its neighbor backs onto a brush-choked canyon that funnels wind straight uphill.

A ZIP code can contain a defensible, low-risk property and a tinderbox 200 feet apart — and score them identically.

Averaging is the enemy of underwriting

When you score risk at the ZIP or county level, you're publishing an average. And an average is exactly the wrong number for an underwriter, because it's wrong in both directions at once:

• It overcharges or declines the genuinely safe properties inside a "bad" ZIP — the ones you'd actually want on your book.
• It underprices the genuinely dangerous properties inside a "good" ZIP — the ones quietly accumulating tail risk in your portfolio.

So you simultaneously walk away from profitable business and absorb losses you didn't price for. The coarse model doesn't just lose information — it actively misallocates capital.

Parcel-level isn't a luxury. It's the right unit.

ARIS scores every property on its own site-specific characteristics rather than a regional average. Inside a single ZIP code, our model routinely finds extreme-risk parcels sitting directly next to lower-risk ones — exactly the distinction the viral photo captures, made quantitative and put on a map before the season.

That's the difference between writing off an entire region and underwriting it profitably, parcel by parcel. You don't need to abandon California. You need to see it at the resolution risk actually happens.

See the interactive parcel-vs-ZIP visual on the home page →