Microsoft Ads Connector // campaign_performance_daily_report:
- The problem: The Account Name and Campaign Name are not available columns.
- Current workaround (partial): To get the Account Name and the Campaign Name into the campaign_performance_daily_report, one must join the account id columns to the account_history and campaign_history columns. However, the campaign_history schema does NOT include deleted/removed campaign data. I have been unable to find another way to include this data from Fivetran, and have resorted to merging with an offline file sourced directly from the platform where deleted items are available.
Microsoft Ads Connector // campaign_performance_daily_report
- The problem: The report has conversions, revenue, all conversions, and all revenue as available columns, but the goal name is not available. If the advertiser has more than goal defined in the Microsoft Ads UI, then the data from these columns will be an aggregation of all goals.
- The workaround: use the goals_and_funnels_daily_report which has goal name, all conversions, and all revenue (but does not have conversions or revenue). Join this data with the cost, clicks, and impressions from the campaign_performance_daily_report. However, you have to be careful because the campaign_performance_daily_report segments the data by campaign ID, by day, by device type, by device os, by network by ad distribution. The goals_and_funnels_daily_report segments by campaign id, by day, by device type by network, by ad distribution (does not have device os), so you must first summarize the data from the campaign_performance_daily_report to remove device os before the join.
This brings me to the last ask. Without a custom connector to select the columns we would like, we end up with a much larger data set than is necessary because the default schemas include every single segment available vs the ones you want. (i.e. I don't need to know/have any use for device os). It would be extremely helpful to have a clean Custom Microsoft Ads Connector with just the fields necessary because each segmentation can easily 2x, 3x, etc the # of rows of data.