How to calculate caseload from a shelter count

If registration hasn't caught up with arrivals, a shelter count is the fastest honest number you have. The method is simple on paper: multiply shelters by average household size. The hard part is getting the inputs right, and knowing where the estimate will drift.

The basic calculation

Caseload estimate = number of shelters × persons per household (PPH).

The PPH figure usually comes from your last registration round or a recent SMART survey for the site. If registration data is six months old and the site has grown since, don't assume the ratio has held. New arrivals often double up in existing structures before new ones go up, which pushes PPH higher for a stretch before it settles back down as people build or get allocated their own shelter.

Say your shelter count comes back at 4,200 structures and your last verified PPH was 4.6. That's roughly 19,320 people. Round it, flag the date the shelter count was taken, and note the PPH source and its own date. Anyone downstream pulling rations or planning latrine coverage off that figure needs to know how fresh both numbers are, not just the final total.

Where the number goes wrong

A few things throw this off more than people expect.

Unoccupied and partial structures. A shelter count from aerial imagery or a walkthrough counts roofs, not households. Abandoned plots, structures under construction, and shelters used for storage or small business all get counted as housing unless someone filters them out. On an established site this might be two or three percent of the count. On a site that's shrinking as people move to a new area, it can run much higher, and you won't know unless the count is dated and you're watching the trend.

Shelter type mix. A block of communal shelters doesn't house the same number of people as a cluster of family tents, but a raw structure count treats them the same. If your site has a mix, you need separate PPH figures by shelter type, or the number will be off in whichever direction the mix has shifted.

Household doubling. This is the one that moves fastest. In the weeks after a surge of new arrivals, you'll see household sizes climb in the existing footprint before the settlement expands outward. If your shelter count hasn't changed much but your registration lines have gotten longer, that's usually doubling, not an undercounted footprint. Cross-check against protection or food distribution line data if you have it.

Stale PPH. Running last year's household-size ratio against this month's structures is the single biggest source of error. If you haven't re-verified PPH in three or four months, treat the resulting caseload figure as a floor, not a point estimate, and say so in whatever planning note it feeds into.

Building the estimate into your planning cycle

The cleanest way to keep this from going stale is to treat shelter count and PPH as two separate inputs with two separate refresh clocks. PPH changes slowly and gets updated through registration or survey work. The shelter count is the piece that can and should move every month, especially on a growing site, because that's where the caseload number actually shifts week to week.

That's the gap a dated, repeatable shelter count closes. Refugee Camp Mapping turns a monthly high-resolution pass over the site into a shelter count and a footprint boundary, both dated, so the multiplication above is running off this month's structures instead of the sweep from however many months back. The output comes as a shelter-count layer and a footprint-change map you can drop straight into your caseload tracker.

If your ration planning is still anchored to the last full registration, a monthly shelter count is worth building into the cycle before the gap between the two numbers gets hard to explain.