Weekday morning business metrics anomaly digest in Slack

Every weekday at 8am, run a BigQuery SQL query against our warehouse that returns two things for each of my key business metrics: the last seven days of daily values and a trailing 28-day baseline. Use one CTE per metric so the query stays readable. The four metrics are daily revenue, new signups, daily active users, and conversion rate. Use the Run Query (Synchronous) BigQuery operation and return the combined result set so the whole thing comes back in one call.

Once the rows are back, for each metric compute the mean and standard deviation of the 28-day baseline window, then compare the most recent day (or the seven-day trailing average, whichever is more meaningful for that metric) against that baseline. Flag any metric where the current value is more than two standard deviations away from the baseline mean.

Write a short narrative digest for the flagged metrics only. Lead with the biggest mover in absolute terms, then work down the list. Use plain language and always include the actual numbers, for example: "Signups fell to 150, well below the expected 400 to 500 range." Skip any metric whose delta is within the normal range so the digest stays scannable. If nothing is anomalous today, post a single line saying all metrics are within their expected range.

Post the digest to the #metrics channel in Slack using the Send a Message Slack operation. Format it as Slack mrkdwn: a bold header line with today's date, then one short paragraph per flagged metric. Keep the whole message under about ten lines so it's readable in a glance from the channel sidebar.