Built a property book that actually balances at the end of the month.

Fourteen buildings, three hundred and twelve doors, and a single line item that didn't add up.

Atlas had grown by acquisition. The original AppFolio account was set up by the previous owner in 2017. By the time we sat down with Marisol — the principal — the system held 312 units across 14 buildings, eleven different lease templates, and an "other charges" category that, between Q3 2023 and Q3 2024, was the third-largest line on the P&L.

What follows is the actual portfolio elevation we drew in week one. The wide grey building on the left is the one she bought in March 2024 — its rent roll still hadn't fully migrated when we started.

Six anomalies. One pattern.

What we found, working from the books backwards into the software. Marisol had been running her own shadow ledger in a Google Sheet for two years — that was our reconciliation key.

One schedule. Every door, every dollar, one screen.

The rent schedule is the central artefact of the property book. We built it as a single sortable, filterable view — building, unit, tenant, lease term, rent, balance, status. Below is the live view of Building 02, slice for September 2024.

Four quarters in. The leak is plugged.

Compared to the trailing twelve months on AppFolio. Numbers from Atlas's own bookkeeper, audited against the new system's source data.