Short answer: in many small marketing teams, someone spends Monday morning copying numbers out of Google Ads, Meta, Google Analytics and the lead sheet into one slide. That can now arrive on its own. Connect the sources to one report, schedule it to land in inboxes every week, and let AI write the five-line summary on top. The tooling is the easy bit. The harder work is accepting that the platforms will never match, and deciding which number your business trusts.
Why the Monday report eats a morning
Ask whoever makes your weekly marketing report where the time goes. It's rarely the analysis. It's logging into four dashboards, picking the right date range in each, pasting figures into a sheet, and then explaining to the owner why Meta says 42 leads, Google Analytics says 31 and the sales team says they got 19 calls.
That last part is where reports lose trust, so start there.
The numbers will never match, and that's fine
Each platform counts by its own rules.
Meta. Meta's Marketing API reference says its default reporting window is 7-day click and 1-day view. So a person who saw your ad yesterday, didn't click, and then filled your form today can count as a Meta result.
Google Analytics. Google's attribution settings page explains that with the data-driven model, credit for one key event is shared between the ads, clicks and other touchpoints that led to it, so a channel can be credited with part of a lead. The default lookback window is 90 days for most key events (30 days for first visits and first opens). The same page notes that changing the reporting attribution model applies to historical data too, so last month's numbers can shift after someone changes a setting.
Google Ads. By default, most settings for conversions built from Google Analytics events match between Google Ads and Analytics, per the same page. One difference by default: Google Ads gives credit only to Google's paid channels, while Analytics also credits other paid and organic channels. And Google Ads doesn't see your Meta campaigns at all.
None of these is wrong. They answer different questions. The fix is a rule, written at the top of the report: leads and sales come from our CRM or lead sheet; platform numbers show which channel is moving. Once everyone reads it that way, the Monday argument stops.
What one weekly report should hold
Keep it to one page. A useful version for an owner or marketing manager:
- Leads this week, from your lead sheet, against last week and the four-week average.
- Spend by channel and cost per lead, using your own lead count, not each platform's.
- Website visits and enquiries, from Google Analytics.
- Social: posts published and the one post that did best.
- Three lines from AI on what changed and what to check.
- One decision for the week: pause, raise, test or leave alone.
If a number doesn't lead to a decision, drop it.
The tooling you probably already have
Google's Data Studio builds reports from connected data sources and can email them on a schedule. Its scheduling help page says the email carries the report as a PDF, with a preview of the first page and a link to the full report. Under the Repeat option you can choose Custom for patterns like every Tuesday and Thursday, or every two weeks on Friday. Limits worth knowing:
- up to 50 recipient addresses, and the person who made the schedule must be on the list;
- the email shows up to five pages as a preview, while the attached PDF holds every page you picked;
- the schedule follows your computer's operating system time zone, so check it's set to IST;
- multiple schedules per report, and delivery to Google Chat or Slack, are Data Studio Pro features.
Before you promise anyone a date, check how each source gets in: a built-in connector, a partner connector or a weekly export.
Where AI fits
AI earns its place in two spots: the summary and the follow-up question.
The summary. A model can read this week's table and last week's and write three plain lines, for example: "Leads up 18% on last week, almost all from the Kannada video ad. Cost per lead on Search rose for the third week. Website enquiries flat." The rule we use: AI writes the words, the numbers come straight from the source. Never let a model retype a figure. Have it refer to the cell, and have a person read the summary before it goes out for the first month.
The follow-up. Google announced Ads Advisor and Analytics Advisor on 12 November 2025, rolling out to English-language accounts from December 2025. In Google's examples, you can ask Ads Advisor "How is my campaign doing?", and Analytics Advisor answers questions like "Why did my active users spike on September 25th?" by running a key driver analysis. Google says Ads Advisor can apply its suggested changes to your account, with your review and approval. Both look inside their own product, so they answer the "why" for that channel, while your weekly report still holds the whole picture.
If you'd like to see how we set up marketing reporting and automation, our marketing automation page shows the approach.
A two-week start
- Days 1 and 2: Write down who reads the report and the one decision each person makes from it.
- Days 3 to 5: Pick the five or six numbers above. Write the rule on which source counts leads.
- Week 2, first half: Connect the sources in Data Studio. Check each number against the platform for the last four weeks. Mismatches you can explain are fine; ones you can't explain mean a tracking problem.
- Week 2, second half: Add the AI summary, read it against the table yourself, then schedule the email for Monday 9 a.m. IST.
After a month, ask the readers one question: did this report change a decision? If not, change the report.
Arthat AI
Working on something similar at your company?
Arthat AI builds these automations for Indian businesses every week. If you want to talk through whether this fits your situation, we'll give you a straight answer — even if the answer is "don't build it."
