Warehouses

The market indices (default rate, average collection period, occupancy/capacity, average ticket, margin) by SECTOR (line of business/CNAE) and by REGION (state), from an ANONYMIZED aggregate of the network — one average per company, never anyone's row. 🔴 The index is only published with at least 30 companies in the sample (k-anonymity); the table does not even store the company identifier. The n is always in view and each index carries its methodology (formula, unit, period). Below 30, the number is NOT shown. For the SECTOR dimension, this index is the publishable face of the SAME aggregate that serves as the prior for credibility (A1) — the sector experience that every new tenant inherits. Read-only.

Each row shows the key (the sector or the state), the indicator, the calculation period and — when published — the value with the "n = X companies" badge right below: the n is part of the number, not a footnote. The methodology column gives the formula and unit of that specific indicator; read it before comparing with your internal number, because a difference in formula explains a lot of "divergence" that is not divergence.

For the SECTOR dimension, this index is the publishable face of the SAME aggregate that serves as the prior for the credibility-weighted estimate of the Actuarial Core — the sector experience that every new tenant inherits while its own history is short. In other words: the number you see here is what goes, with the weight stated in text, into the "sector" half of the risk estimates. One source, two readings — never two different numbers for the same thing.

How your sector (line of business/CNAE) and your region (state) behave on market indicators — default rate, average collection period, occupancy, average ticket, margin — and whether the sample is large enough to trust. Two tables, one per dimension; each row shows the value, the period, the number of companies (n) always in view and the methodology (formula, unit, k-anonymity). Read-only screen.

Because the number is only published when at least 30 distinct companies have contributed to that breakdown. Below that, showing the average would almost be showing the data of whoever contributed — with 3 or 4 companies, the "sector average" is practically the neighbor's number. We'd rather say we don't have the number than publish a fragile one: a lying number destroys more trust than its absence.

No. Your company enters as ONE POINT: first we calculate your average, and only that average goes into the sector or region calculation — never a customer, a receivable or an invoice of yours. The tables behind these panels don't even keep the company's identification, so going back from the number to whoever contributed is impossible by construction, not by promise.

Once a day, by an automatic routine. It recalculates the current period and replaces the previous value — running it again on the same day doesn't duplicate anything. When there isn't enough activity, the routine simply doesn't publish, and the panel keeps saying that the sample is insufficient.