What is the difference between RAW data and report data?
The three important concepts are raw data, report data, and archiving.
Raw data
Raw data arrives constantly and is available immediately. One of the few places where raw data appears in your Matomo interface is: Behavior >> Visits Log.
In both Matomo Cloud and Matomo On Premises, raw data can be immediately exported any time from the API interface. This interface is accessible at Administration (cog icon) >> Platform(export) >> API >> Live.
If you are running Matomo On Premises, you can directly inspect the tables of your database: the tables prefixed by **matomo_log_ **are raw data.
Archiving
The action of turning raw data into the other kind of data (report data) is called archiving. It occurs whenever the back end of Matomo executes the Symfony PHP function console core:archive.
Report Data
The vast majority of what you see in Matomo, and especially every tabular table, is report data. It does not exist until archiving runs. If you ever open a report and see all zeroes, it either means there is no raw data for that time period, or more likely, raw data has been collected for that time period and archiving did not successfully happen.
Report data is sometimes called enriched data or aggregated data. Again, almost everything you see in Matomo is Report Data.
Troubleshooting
If you see a gap in your report data, and your report data seems inaccurate, especially too low, the most basic thing to check will be the following.
The main way to assess the health of your raw data came in for a suspicious time period and then do the same for a healthy time period. You can then compare the two.
If you running Matomo Cloud, you will need to do this by checking your API >> Live.
But in Matomo On Premises, since you have direct access to the database, it is often simpler to run a query. For example:
A good start would be a query similar to this
SELECT COUNT(*) AS hits,idsite
FROM matomo_log_link_visit_action
WHERE server_time > ‘2025-01-01’
AND server_time < ‘2025-01-31’
GROUP BY idsite;
You should try further queries as well, as suggested in our SQL recipes.
Compare your current time period to other time periods when the data was as expected. Hopefully it will confirm that your raw data is reasonable. If however you find a problem, your troubleshooting will turn to the Developer Tools in your browser, which should help show what kind of data is sent to your Matomo server when a visitor opens your app or web pages. Focus on both the browser’s Console tab and its Network tab.
If the raw data seems reasonable, the next failure point to check is the archiving process.
Is the archiving process completing on time? Is the scheduled archiving still running when the next archiving is triggered? Are there warnings when the archiving executes? Are there PHP errors?
Some places to look for these answers:
* Matomo’s back end will log messages to either the default location tmp/logs/matomo.log, or an override location you set in the config.ini.php file at the line logger_file_path =
* PHP errors will show up in the location set by your PHP config file. Note that PHP FPM module handles your Matomo interface but archiving uses a separate PHP configuration for the PHP CLI module. To find this configuration file, open your Command Line Interface (CLI) and type
php -i | grep error_log
Deleting data
Raw data is risky to store long term: penalties in privacy laws are more severe for excessively storing raw data than for storing report data. A further reason to delete old raw data is that it usually occupies disk space which is expensive. Once the report data is generated, if the report data does not need to be changed in any way, you can consider deleting the older raw data. As a rule of thumb, you can consider the default setting we use in our Matomo Cloud: delete old raw data after 24 months.
You can also configure Matomo to delete old report data. Look in your interface at Administration (cog icon) >> Privacy >> Anonymize Data >> delete old Report data (archived data).. As a rule of thumb, you can consider the default setting we use in our Matomo Cloud: delete old report data: never.
What issues arise if the RAW data is not available?
Please see the section in this FAQ called: What is the impact of deleting historical raw data?