Matomo Automated Reporting: Three Ways to Build a Weekly Briefing
Compare Matomo scheduled reports, on-demand AI summaries and external workflows. Choose what to automate and understand each option’s limits.
Matomo automated reporting can mean three different jobs: sending existing reports on a schedule, writing an explanation of the figures, or running a custom workflow across several services. Choose the job first. You may already have everything you need inside Matomo.
Option 1: Schedule the reports Matomo already provides
For a recurring attachment with standard metrics, start with Matomo’s built-in scheduled reports. Its report delivery guide describes recurring delivery and export formats. This is a good fit when recipients know how to interpret the charts and tables.
Open the report scheduling area in your Matomo administration, create a report, choose its contents and recipients, and configure the frequency and format. Use the current configuration instructions for your version. Send a test before relying on the schedule.
On a self-hosted installation, include archiving and scheduled-task health in the operational checks. An attractive email is not useful if its data is stale.
Option 2: Generate an explanation with an AI assistant
If the recurring work is interpreting results, use an assistant connected to your reports. With Savigny, connect your Matomo in setup, then use chat in the dashboard or a compatible MCP client.
Start with a reusable reporting instruction:
For the selected website, compare the last complete calendar week with
the preceding complete week. State the exact dates and website timezone.
Read visits, traffic sources and top pages for both periods.
Write a 150-word briefing with the biggest supported changes, the
absolute figures, and two suggested checks. Separate facts from
hypotheses. Flag missing data. Do not infer revenue or conversions.
This produces an on-demand draft. Savigny’s current MCP tools do not schedule or send reports. Any scheduled execution or email delivery needs a separate system. The assistant’s ability to run a recurring task also depends on its own features and permissions.
Option 3: Build an external workflow
Use a workflow platform when you need a fixed schedule, data validation, AI summarisation and delivery in one process. A typical sequence is: schedule, retrieve the previous period, check the response, generate a summary, review or deliver, then record success.
The Matomo and n8n guide explains this using Matomo’s Reporting API. That API route is separate from Savigny’s hosted MCP connection.
Decide what “finished” means
Before automating, write one report manually. Define the audience, website, period, timezone, required metrics and action expected from the reader. Use the report template to capture those choices.
Then decide what should happen when a metric is unavailable. Zero visits, an empty response and an authentication failure are different states. A workflow should stop or flag the report when required data is missing, rather than turning an error into a reassuring summary.
A recurring report needs an owner
Assign someone to check delivery failures and periodic changes to tokens, tracking and goals. Include a source date and period in every report. Compare like with like, and never label a partial month as a completed-month decline.
Use native scheduling for regular report delivery, an AI assistant for interactive interpretation, and an external workflow when the whole process needs to run unattended. Start with the smallest process that solves the reader’s problem.