View Measurement File Stats

  1. From the Connect navigation menu, click Files to navigate to the Files page.

    C-My_Files_menu_selection.png
  2. Select the desired Measurement audience from the Audience dropdown.

The "recognition rate" column will appear for each file. By clicking on the additional file information box you will be able to see additional statistics.

Why can’t I see the recognition rate column for all of my audiences?

This feature is only available for Measurement audiences at this time.

What is a RampID?

A RampID is an anonymous representation of an aggregation of touchpoints, representing a person.

How do you determine what is invalid data?

Invalid data are data that we deem unmatchable. For PII files, we check the validity of the identifiers by looking for blanks, bad format, or known invalid email domains, for instance. For files containing devices, we check that the device is not a ‘0’, blank, or a permutation of N/A, nil and null.

Can you point me to which line of data is invalid?

We do not retain this information in the system and we shuffle rows for each file to comply with privacy requirements. It is therefore not possible to point out specific rows that were considered invalid.

How can I prevent my data from being invalid?

For first-party data, make sure that all of the fields are filled, that the identifiers look correct (for instance, you can use an email regex to ensure basic email correctness), and that there are no obvious invalid data points, such as noname@nodomain.com. For third-party data, you can work with your DSP and/or agency to ensure that they do not include blank or ‘0’ devices in their exposure files.

Is anything changing in my file output?

No, this feature is entirely separated from any matching or formatting processes. It is solely a statistics feature and we have not changed the way we process and match files.

What is the difference between recognition rate and match rate?

Almost none. The name "recognition rate" better captures our matching process and our ability to recognize touchpoints by excluding invalid data.