Filling gaps in a sparse schedule of values before renewal

Suburban homes along the edge of a neighborhood, overlooking an open field under a partly cloudy sky.

Every submission has them. A schedule of values comes in from the broker with half the secondary characteristics blank, year built missing on a third of locations, and construction class entered as "Unknown" more times than you'd like to admit in a renewal memo. What you fill those gaps with is the real question, and that choice has to hold up when someone asks you to defend the number six months from now.

Start by triaging, not patching

The instinct is to run a scrubbing routine and move on. Resist that until you've sorted the blanks by what they cost you if you're wrong.

A location with a $400,000 TIV in a low-hazard zone can take an industry default without much argument. A location with $40 million in TIV sitting in a wind-exposed coastal county, or a cluster of warehouses in a flood plain where the schedule just says "commercial, unknown," is a different problem entirely. That's where a default value stops being a convenience and starts being a modeling decision someone will eventually have to explain.

The first pass sorts blanks by TIV concentration and peril zone, separating rounding errors from the ones that could move your modeled loss, before any cleaning routine touches them.

Default values vs imputed exposure

Default values come from lookup tables: occupancy-based averages for construction class, regional norms for year built, standard COPE assumptions baked into the vendor model. They're fast, and for the long tail of small, low-value locations they're honestly fine. Nobody is restating a treaty over a few thousand dollars of TIV sitting on an unremarkable default.

Imputed exposure is different. It means replacing a blank with something derived from an actual observation of the location, not a category average. That could be cross-referencing a prior schedule, pulling assessor records where they exist, or using a current read on what's physically on the ground to confirm footprint, rough height, and whether the parcel is even still a single structure or has since been subdivided into three.

The difference matters most in two situations: where defaults systematically understate density, in growth markets where the schedule was built two or three renewals ago and the submission just carried old secondary characteristics forward, and where the default occupancy assumption doesn't match what's actually built. A "light industrial" tag on a parcel that's now a distribution center with four times the footprint is a materially different risk.

A workable process for renewal

For the locations that matter, the approach that holds up at renewal usually looks like this:

Pull the high-TIV, high-hazard blanks into their own list, separate from everything else. Check whether a prior-year schedule or any secondary source confirms construction type, year built, or number of stories before reaching for a default. Where nothing confirms it, impute from the best available current signal rather than an occupancy table, and note which locations got which treatment. Keep that note. It's the difference between a defensible exposure build and a guess that happened to look clean.

Stale data causes more trouble at renewal than missing data does: a schedule that was accurate three renewals ago, rolled forward since while the built environment underneath it changed anyway. A market builds out, a parcel gets redeveloped, a single-story warehouse becomes a multi-building complex, and none of that shows up until someone goes looking.

That's the gap Exposure Mapping is built for: a current, territory-wide view of structure footprints and counts from recent high-resolution imagery, so the locations you'd otherwise be defaulting get a current read instead. It complements a good SOV, covering exactly the locations where the SOV can no longer describe what's there now.

If your sparse schedule has more of those stale locations than clean defaults can cover, it's worth seeing what a current built-inventory layer looks like for your next renewal.

Get started

← Back to the blog