Projects using Google BigQueryProjects using Google BigQuery
Projects
People
Cover image for How do you automate a
How do you automate a daily sales report from BigQuery to Slack with Make.com, when the data is 250,000 rows and the reader is not an analyst? Six decisions make or break it. A direct-to-consumer brand's daily sales report took one to two days to compile, which meant Monday's numbers arrived on Wednesday. A 7 AM Make.com scenario now pages the data out of BigQuery, computes the KPIs and posts a plain-English summary to Slack and email before anyone opens a laptop. The six decisions: 1. Page, do not pull. A day can be up to 250,000 rows. The scenario pages through BigQuery in chunks and aggregates as it goes, so no single step holds the whole table. 2. Compute the KPIs in the scenario, not in the prompt. Revenue, orders, average order value, top products and day-on-day change are arithmetic. The model only writes the sentences. 3. Give the model numbers, never the rows. A short structured summary goes to OpenAI with a fixed template; the output is two paragraphs a founder reads in thirty seconds. 4. Check before you post. If a KPI is null, if the row count is far below the trailing average, if a query failed, the scenario alerts the owner instead of posting a confident wrong report. 5. Post where people already look. Slack channel at 7 AM plus an email copy for the people not in Slack. No dashboard to remember to open. 6. Keep a paper trail. Every run logs the query window, row count, KPIs and the text it posted, so a number can be traced back a month later. The diagram shows the run: schedule, paged query, KPI step, summary, checks, Slack and email. The full case study is on my profile. The same scenario shape works with Shopify, Stripe, a CRM or a Sheet as the source. Prem Patel, Nex Automations. Make.com Level 5 certified, official Make partner, Zapier Certified Expert.
1
240