Daily signup digest in Slack

Every weekday at 9:00 AM Eastern Time, run a daily signup digest for me and post it to Slack. The goal is a short, qualitative read on yesterday's new signups, not a raw numbers dump. The value of this workflow is the written interpretation on top of the numbers.

Trigger: cron, every weekday (Monday through Friday) at 09:00 America/New_York.

Step 1. Run a PostgreSQL Custom Query against my users table. I will plug in the exact table name, the signup timestamp column, the plan tier column, the acquisition source column, and the country column when I configure the workflow, so write the SQL with clearly labeled placeholders for those fields. The query should return: (a) the total count of new signups in the last 24 hours, (b) yesterday's signup count broken out by plan tier, by acquisition source, and by country, and (c) daily signup counts for each of the previous 7 days so a rolling 7-day average can be computed and compared against yesterday's number.

Step 2. Write a short narrative summary in plain English based on the query results. The structure should be: a single big headline number at the top showing total signups in the last 24 hours and how that compares to the 7-day average (e.g. up 18% vs average, down 30%, in line). Below the headline, use bullet points to call out what's up, what's down, and anything anomalous worth attention. Examples of things to flag: a spike from a specific country, a quiet day on a paid plan tier, a new acquisition source appearing for the first time, an unusually strong or weak channel, a streak ending. Do not just dump every count. Interpret the data and only call out what changed or what's notable.

Step 3. Send the message to a Slack channel I will specify, using the Slack Send a Message action. Format the message with Slack mrkdwn: bold the headline number, use bullet points for the breakdown, and keep the whole thing skimmable in under 20 seconds. No tables, no code blocks, no giant data dumps.

Edge cases: if yesterday had zero signups, still post a short message that says so, plus the 7-day context, so the channel knows the workflow ran. If the database query fails, post a one-line error message to the same channel so I notice it the next morning instead of silently missing a day.