Every report reads from the flat file, not from the ad platforms. If two reports pull from two places, they will eventually disagree and nobody will know which is right.
Setting it up (once)
1
1. Read the framework
Reporting framework concepts covers the north star, counter-metrics, and how the layers relate. Data quality and rollup concepts covers what breaks when weekly numbers don’t add up to the monthly ones.
2
2. Define every metric once
Fill in the marketing reporting data dictionary: one definition, one owner, one grain per metric. Then agree the headline set in KPI definitions.If a starter pack fits, begin there: ecommerce or sales-led.
3
3. Build the flat file
Set up the flat file. Its grain and its column names come from the data dictionary, and its campaign column joins on your naming conventions and UTM values.
4
4. Set the cadence and the dashboards
Agree who reports what, to whom, how often in reporting cadence, and how it’s presented in dashboard standards.
Producing a report
Each period rolls up into the next. A monthly report is not written from scratch; it’s assembled from weeks that already reconciled.
1
1. Load and check the rows
Add the period’s rows to the flat file and check them against the rules in data quality and rollup concepts. Catch the gaps before writing, not after someone questions a number.
2
2. Check the change log
Anything that changed mid-period (a budget shift, a tracking fix, a naming correction) explains a movement that otherwise looks like performance.
3
3. Write it
Use the template for the period. Report writing concepts covers how to say what happened without overclaiming a cause.
4
4. File it
Save to the archive for that period. The archive is what makes a quarterly report possible without re-deriving three months of numbers.
Related resources
- Marketing reporting data dictionary One definition per metric.
- Flat file The single source every report reads.
- Reporting framework concepts The reasoning behind the structure.
- Report writing concepts How to write the narrative.
- Set up tracking What has to be right upstream for any of this to work.
- Reporting best practices How to produce reports people trust.