Data Beginner 16 min

The reporting stack — daily, weekly and monthly do different jobs

The problem

You send a daily sales figure, a weekly summary with the same figures aggregated, and a monthly pack that aggregates the weeks. Three preparations, one message. The MD reads the monthly, skims the weekly and has muted the daily.

The workflow
1Define each report's job
2Set thresholds
3Build daily exceptions
4Build weekly patterns
5Build monthly decisions
6Remove overlap
7Deliver where read
Review usefulness

What you need

Business: FMCG distributor, ₹18 crore, 190 retail accounts, Maharashtra. Readers: MD (decides credit, pricing, hiring), Sales head (decides allocation, visits), Warehouse head (decides dispatch priority).

The tool

A scheduled query plus a short AI summary

The prompt

Subscribers. The full prompt is part of the SMEMinds AI Playbook. See plans.

What comes back

DAILY EXCEPTION REPORT — WhatsApp, under 80 words, 8am. Conditions: (1) any account in the top 40 by revenue with no order for 21 days — this is the one that would have caught your ₹31 lakh problem in week three; (2) daily order value below 60% of the trailing 14-day average; (3) any SKU below…

Verify before you use it

Human review required. Check every line below before this leaves your screen.
  • Are the thresholds right for your business? 21 days for a top-40 account is a starting point — check it against your actual order frequency data before committing.
  • Would the daily report have caught your real incident? Test it against the last three months of history. If it would not have fired, the threshold is wrong.
  • Is anyone going to be upset that a report is being stopped? Ask before stopping — sometimes a report nobody acts on is still being used for something you do not know about.
  • Check the delivery channel honestly. If the MD does not read email before 10am, an 8am email is not a daily report.