Methodology

1. Source

We ingest public BTOS collection periods and estimates from the Census API. BTOS is a high-frequency survey of U.S. employer businesses, with new data released every two weeks.

Official BTOS data ↗

2. Diffusion indexes

Where a direct published index is available, we retain it. Where response shares are supplied, the adapter can reproduce Census-style diffusion indexes from response weights such as Increased / No change / Decreased and the five-level performance scale.

3. Trend windows

Each entity is tracked against the prior release and, when history exists, against three and six releases back. This reduces the temptation to over-read one noisy biweekly move.

4. Statistical signal

If source index standard errors are present, the one-release difference is compared with a 95% threshold using the combined standard error. If the source adapter does not provide an index standard error, the site explicitly shows “n/a” instead of inventing significance.

5. Momentum score

The 0–100 Business Pulse Momentum score combines index levels and multi-release changes. Performance, revenue, employment, hours, demand and future-looking measures are positive contributors. Input-price pressure is inverted because higher input costs are a business headwind.

The score is intentionally an analytical ranking tool — not an official Census statistic.

6. Confidence

Confidence is based on indicator coverage and available history. “Low” generally means the entity is newly imported or lacks enough releases; “High” requires broad metric coverage and multiple historical periods.