Raw data export

How it works

An MMP stores a row for every event it receives. Each row holds a time, the event name, device and campaign details, and the attribution result. A dashboard report adds those rows up by channel or by day. Raw data export hands you the rows themselves.

Vendors deliver them in a few ways:

• scheduled files written to cloud storage that you own;
• an API you call to download rows for a date range;
• a real-time message to your server for each event, often called a webhook or advertiser postback;
• a file download from the dashboard.

From there your team loads the rows into a warehouse and joins them with other data, such as subscription records or support tickets.

When it matters

You need raw rows when a question cannot be answered inside the MMP's reports. Typical cases are joining attribution with revenue from your own backend, building a model on user-level history, or keeping a full copy of your data in case you change vendor. If nobody on the team writes SQL, you will not use it yet.

It is one of the capabilities often left out of an entry plan. Airbridge, for example, lists raw data export under its Growth Plan and not under Core.^[1] Check the plan table before assuming you can take event-level data out.

How it relates to attribution

Each exported row carries the attribution decision the MMP made: which touchpoint got the credit. That lets you check the MMP's totals against your own counts and apply your own rules on top. Needing this is a common reason to move up a plan, because exported rows also protect you against losing history in a later switch.