Youtube Analytics

The most common Youtube Report, or Endpoint, is the Basic Stats from Video Reports (see screenshot below)

This report type works in a specific way. Every fetch will in fact collect the data of every video published since the account's creation, but will only include the data within the specified Time Range when populating the data extract file. 

For instance, when fetching September data, all videos, including the ones published in 2012, will be recovered, but only the performance data (shares, views, likes, dislikes, subscribers gained etc.) from September will be displayed. 

The main issue when pulling data from Youtube Analytics comes from the missing Time Stamp or Date Column. The extract is implicitly reporting the videos on the Time Range specified, without any daily attribution.

Therefore, we recommend the following best practices to fetch, import and overwrite Youtube data:

FETCH LIFETIME EVERY DAY

When fetching Lifetime, overwriting only on Datastream is sufficient as every new extract will contain accumulated data from account creation to Date.

FETCH CURRENT/LAST MONTH + METADATA EXTRACT DATE 

If you want to visualize data from the ongoing Month, selecting "Current Month" and fetching every day is necessary. If, however, you can wait for the beginning of the next month to visualize data, fetching "Last Month" on the first day of next month is sufficient.

Since there is no field "Date" pulled from the API, it is also important to set correct overwrite options as Adverity cannot simply overwrite "Video_published_at", without the risk of overwriting all data with each new import.

1. To get a timestamp, we need to use transformation scripts to extract the date from the meta_data as follows: 
   
2. Once the fields "start_date" and "end_date" have been created, we need to map them as Key Columns in Schema Mapping. Select the 2 following Key columns: "Start_date", "Video_id", (Or "Start_date" and "Video_published_at").
3. Finally, in Overwrite options, we need to select "Datastream" + "Key Columns".

In this way, every new import will overwrite previous import if the data are from the same month (based on Start_date which will always hold the value of the 1st of the month).

This would result in Monthly accumulated data. Extract from the 25/09 will overwrite extract from the 24/09 containing the latest data of the current Month. On the 01/10, I will now start importing the data for October without impacting the data from September anymore as my Start Date now changed from 01/09 to 01/10).

EXAMPLE IMPORT

1. Fetch "Current Month".
2. Extract "Start_date" from the Metadata.
3. Set "Video_id" and "Start_date" as Key columns.