Preliminary Damage Assessment Software: Getting Consistent Numbers to the State Before the Window Closes

Thirty six hours after the storm, fourteen teams, one deadline

The tornado came through on Tuesday evening. By Thursday morning fourteen two person teams are working assigned grids with clipboards and phones. Each team has a form, a photograph habit and an opinion about what counts as major versus minor damage. Team four is generous because the lead has seen worse. Team nine is conservative because the lead is a building inspector who knows what a repair actually costs. Nobody is wrong, exactly, and that is the problem.

Friday evening the county emergency manager is in a conference room turning fourteen sets of paper into a spreadsheet. Photographs are on phones, named by camera timestamp, and no longer clearly connected to the addresses they belong to. Half the destroyed properties have no dollar estimate because the team was moving fast. Sunday it goes to the state. Wednesday it comes back, because the destroyed and major counts do not reconcile against the photographs, three addresses appear twice, and the valuation method varies by team. The window to demonstrate the threshold has just lost five days it did not have.

This is a money document, not a status report

The assessment is the evidence base for a declaration request. Getting it right decides whether individual assistance and public assistance reach the jurisdiction, and the amounts involved dwarf the cost of any software you will ever buy to produce it. It also decides something more subtle: how credible your agency is the next time. State reviewers remember which counties send numbers that hold up and which counties send numbers that need rework, and that memory shapes how much benefit of the doubt you get on a marginal event.

The uncomfortable truth is that most counties are not failing at data collection. They are failing at consistency and at aggregation. Fourteen teams collecting reasonably good local observations, then combining into something the state cannot use, is the standard pattern. Fixing the collection app without fixing the consistency layer just produces inconsistent data faster.

Where Crisis Track, Survey123 and Veoci stop

Crisis Track is purpose built for this and does the federal side math well, including rollups against the thresholds and reporting in shapes reviewers recognise. If your assessment practice fits its model, it is a strong and honest buy. The limit is that its model is its model: if your state uses its own damage categories or valuation approach, or if you want assessment data joined into your own parcel, permitting and recovery systems rather than living in a subscription platform, you are working against the grain.

Esri Survey123 is excellent at the thing it is: a form that works offline, syncs when connectivity returns and drops points on a map you already own. Counties reach for it because it is fast to stand up and it is already licensed. What it does not do is the part that gets you rejected. The category logic, the valuation method, the deduplication, the threshold rollup and the submission package are all left to you, and in practice that means Excel two days after the event with a tired analyst doing it by hand. The collection improved. The failure point did not move.

Veoci is flexible and will let you configure a workflow around all of this inside their environment, which suits agencies that want one platform for many emergency management functions. The trade is that you are building in someone else's abstraction, and the parts that need to reach into your assessor database or your recovery accounting are the parts that get awkward.

Consistency between teams is the entire game

The single highest value feature in a custom build is not a nicer form. It is forcing every team to reach a category the same way. That means the app asks about observable conditions rather than asking for a conclusion: roof structure condition, wall integrity, water depth inside the structure, foundation displacement, whether utilities are cut, whether the structure is accessible. The category is then derived from those answers by rules you control, uniformly, across all fourteen teams.

Do that and three things change. Your destroyed and major counts stop depending on which lead was in which grid. A reviewer's challenge is answerable, because you can show the observations behind the category rather than defending a judgement. And you can recalibrate mid event: if the state signals that your major threshold is reading high, you adjust the rule and reprocess, instead of sending fourteen teams back out.

Offline is a requirement, not a preference

Assessment happens where towers are down. The application has to hold the assigned grid, the parcel layer, the address points and the imagery basemap on the device before deployment, capture everything locally including photographs at full resolution, and sync opportunistically when a team passes through coverage or returns to the staging area. Photographs must carry the structure identity in metadata at capture time, not be matched up later by timestamp, because timestamp matching after the fact is exactly where the photographic evidence quietly disconnects from the record.

Sync also has to be conflict safe. Two teams will assess the same corner property, because grid boundaries are imperfect and everybody is tired. The system should catch it at sync and route it to a reviewer rather than silently keeping the last write.

Join to parcel and assessor data before you deploy anyone

• Pre load every structure in the impact area from the parcel and address point layers, so teams confirm and assess rather than typing an address into a phone in the rain.
• Pull assessed value, year built, construction type and occupancy from the assessor, so valuation starts from a defensible baseline instead of a team member's guess.
• Flag the ones that matter differently: mobile and manufactured homes, multi family properties where one structure means many households, and businesses, which count under a different part of the picture.
• Capture insurance status and flood insurance status, because uninsured loss is what drives individual assistance need and it changes the argument entirely.
• Track access and re inspection, since a property nobody could reach on day two has to come back into the queue rather than fall out of it.

The parcel join is the highest leverage half day of work in the entire project. Counties that skip it spend the rest of the event reconciling addresses that were typed twelve different ways.

What this costs and how long it takes

A first release covering offline collection with observation driven categories, the parcel and assessor pre load, photograph capture bound to structures, deduplication and the threshold rollup runs $45,000 to $95,000 and ships in 8 to 12 weeks. A full platform adding the state and federal submission package, valuation modelling, insurance capture, team assignment and progress management, public reporting, and the handoff into cost recovery runs $110,000 to $250,000 across 5 to 9 months.

What moves the number: whether you are building for one county or standardising across every county in a state, because a state build means configurable categories per jurisdiction and a reconciliation layer above them; the condition of your parcel and assessor data, which is the usual schedule risk; and whether the assessment has to hand off cleanly into disaster cost recovery, which is worth doing and is a genuine second integration rather than an export button.

When you should not build

Do not build if you activate rarely, have a small assessment cadre and no standing analyst. The value of a custom system compounds with repetition, and a county that does this once every four years will find the system unfamiliar exactly when it matters. Crisis Track or a good Survey123 project, plus a half day of calibration with your team leads before each deployment, will outperform an unfamiliar custom tool.

Build when you are a state agency trying to get consistent numbers out of many counties, when you are a large jurisdiction that assesses more than once a year, when your submissions have been returned for category or valuation inconsistency, or when you need the assessment data to flow directly into recovery, permitting and cost tracking rather than being retyped.

How to choose a developer

Ask them how they would derive a damage category. If the answer is a dropdown with four options, they have built a form and your consistency problem is untouched. The right answer captures observable conditions and derives the category through rules you can change and reprocess against.

Ask what they have shipped that works with no connectivity for a full day, including full resolution photograph capture and conflict safe sync, and ask specifically what happens when two teams assess the same structure. Ask how they will handle the ten percent of impacted structures that do not match a parcel record, because there always is a set like that and dropping it is how a mobile home park disappears from a count.

Ask whether they have worked with assessor data before, since assessor exports are idiosyncratic and every county's is different. Then settle ownership before kickoff, including the data, which belongs in infrastructure you control. At Digital Heroes the jurisdiction owns the code and the data from the first commit. Start scoping by pulling your last assessment, taking one grid, and checking whether you can still connect each photograph to the structure it documents. That answer usually decides the project on its own.