Preliminary damage assessment vs verified field assessment: what goes into the appeal

Every appeal document eventually asks the same question: how many structures were damaged, and how do you know? Most teams are working with two different answers measured at two different times, and funders who read a lot of these documents can tell which one you're giving them.

Two different jobs, not two tiers of the same job

A preliminary damage assessment is a count. It tells you how many buildings in the affected footprint show visible damage, roof loss, partial collapse, total collapse, pulled from very-high-resolution imagery against a pre-event baseline. It doesn't tell you who lived there, what they lost inside, or whether the water system still runs. That's not a gap in the method. It's not the method's job.

A verified field assessment is a different exercise entirely. Teams walk the ground, talk to households, confirm occupancy, and record loss in terms an appeal's needs-analysis section uses: displaced families, destroyed livelihoods, damaged water points. It answers questions imagery can't, because imagery doesn't know who was home.

The mistake is treating the second as a more accurate version of the first, when it's really a slower process answering a different question. A field team covering a district takes days to weeks depending on access, security, and how many enumerators you can deploy. An imagery-based count over the same footprint runs in a day or two once a clean post-event image is in hand. If your appeal deadline sits inside that gap, and it usually does, you need the count, not a faster version of the field assessment.

Remote count vs ground truth: what each one tells a donor

Donors reading an appeal are checking for a number that's internally consistent with the scale of the event and traceable to a method. A structure-by-structure count from satellite imagery gives them that: here is the affected footprint, here is the count against a known baseline, here is the delivery date relative to the event. It's defensible because the method is visible and repeatable, not because it claims to be ground truth.

What it can't give a donor is confirmation of occupancy or severity of personal loss, because roof damage visible from above doesn't map cleanly onto household-level need. A collapsed roof on an empty storage shed and a collapsed roof on an occupied home count the same from an optical sensor. That distinction is exactly what a field team is for, and it's why serious appeals eventually want both: the fast count to open the window and anchor the ask, the field data to refine who gets what once teams can reach the area.

This is also where imagery-based damage count accuracy gets misread. The count is accurate to what the imagery can see: damage visible at the delivered resolution, checked against the affected footprint. Holding it to a different standard, a structure-for-structure match against a field tally, misses what the method measures. A count built this way holds up in an appeal annex because its method is stated plainly.

When field verification is required

Certain lines in an appeal require field verification, and the rest don't. If the ask includes per-household cash transfers, livelihoods compensation, or anything tied to a named beneficiary list, a funder will expect field-verified figures before money moves, full stop. A remote count isn't built to satisfy that requirement. Where a remote count earns its place is in the opening sections: scale of the event, affected footprint, order-of-magnitude structural damage, the numbers that justify triggering the appeal in the first place and that a pooled-fund allocation table expects on day one, before any team has set foot on the ground.

The practical sequence most emergency response teams land on: lead the appeal with a fast, standardised structure count so the ask has a number behind it from day one, then layer field verification in as access allows, updating the beneficiary-level figures without re-running the whole appeal from scratch.

If you're building the opening numbers for an appeal and the clock is the constraint, a standardised structure-by-structure count delivered within days of the event is the piece built for exactly that window. Get in touch when you've got an affected footprint and an event date.