How to build an automated weekly business report that pulls from Stripe, GA4, YouTube, and email

Every small business owner I know has the same Monday morning routine. Open Stripe to check revenue. Open Google Analytics to see traffic. Open YouTube Studio for views. Open the email tool for open rates. Then paste it all into a spreadsheet and try to figure out what it means.

That takes about 45 minutes and produces a snapshot nobody enjoys looking at, partly because the act of assembling it is exhausting and partly because the numbers sitting in isolation do not tell you much.

I built a different version of this for OliviaGRE, a GRE tutoring business that runs almost entirely on autopilot. OliviaGRE gets students through a combination of YouTube content, a free study plan lead magnet, and an automated email sequence. To manage it properly, you need visibility across all those channels at once. Logging into four separate dashboards every week was not sustainable, so I built a weekly report that assembles everything automatically.

What the report shows

The weekly snapshot covers six areas, each comparing the current 7-day window to the previous 7 days so you can see direction quickly.

It covers YouTube performance (views, external clicks to the website), website traffic from Google Analytics, form submissions and new leads broken down by source, revenue and new payments from Stripe, email list growth and open/click rates from MailerLite, and chatbot engagement. The last section is a list of recommended actions, written automatically based on what the numbers show.

That last part is where it gets genuinely useful. Raw numbers are not a report. A report tells you what the numbers mean and what to do about them. A script that looks at the data and writes specific action items — "the free trial conversion rate dropped 12% this week, check whether the Tuesday email link was working" — is more useful than a dashboard that just displays the numbers and leaves the interpretation to you.

How it gets built

The script runs weekly and pulls from five places at once. YouTube for views and external link clicks. Google Analytics for sessions and user counts. Stripe for charges and revenue. MailerLite for email stats. Google Sheets for the CRM data — leads, bookings, call outcomes.

It assembles all of that into a single snapshot, then runs analysis on it and generates a readable summary. The output is a styled HTML dashboard and a plain-text version, both saved automatically. No one has to run it manually or copy anything anywhere.

The payments and booking data also sync to Google Sheets on an hourly basis as a separate process. That keeps the CRM current throughout the week, so by the time the weekly report runs, it is pulling from clean, up-to-date data rather than trying to interpret raw payment events in real time.

What goes wrong

The most common problem is login credentials expiring. Google requires periodic re-authentication for Analytics and YouTube access. When those expire, the report shows zero traffic while everything else looks normal. In May 2026 this went unnoticed for nearly two weeks, because Stripe and MailerLite were still running fine and the business felt healthy.

The fix was a built-in sanity check. If Stripe processed payments and MailerLite added 40 subscribers but Google Analytics shows zero sessions, that is a broken connection, not zero traffic. The report now flags conflicts like this explicitly rather than silently reporting zeros.

The other issue is when platforms update their API response format without warning. The script starts returning empty data instead of an error, which is harder to catch. Thorough logging at every data-parsing step makes these much easier to spot quickly.

What it replaced

Before this, assembling the weekly snapshot took two people about 45 minutes and still produced a spreadsheet that was slightly wrong. After, it runs automatically in about 2 minutes and surfaces things that would never have been caught manually.

The clearest example was discovering that OliviaGRE's long nurture email sequence had a 43.6% open rate but only a 2.2% click rate. The open rate looked fine in isolation. The click rate looked low but not alarming. It was only when the report calculated the ratio between them that the problem became obvious — people were opening the emails but nothing in them was compelling enough to act on. That is an actionable insight. A spreadsheet with two separate numbers would not have surfaced it.

How long it takes to build

A basic version that pulls from Stripe and generates a plain-text summary takes a day or two. Adding email and website analytics adds a few more days. YouTube has more moving parts and slightly finicky quota limits, so budget extra time there.

The part that took longest was the recommendation engine — the bit that looks at the numbers and writes useful action items rather than just displaying them. That involved writing a prompt that could receive a data snapshot and return specific, prioritised suggestions. Getting it right so the recommendations were genuinely useful rather than generic took several weeks of iteration.

Frequently asked questions