I’m committed to posting to this blog every day for Lent. Almost didn’t have a post for today, and then I stumbled at work.
I’m putting data into Looker (Formerly DataStudio). I had a bunch of rows collated to one. These rows contained the regular campaign, some experimental campaigns, branded ad groups, unbranded ad groups. I was looking for ROAS (return on ad spend) or conversion value / cost. I was getting the proper answer when there was only one row of data, however, Looker was just adding the results together if there were multiple rows. Why is this important?
Well let’s look at the data:
- Unbranded –
- Conversion value – $100
- Cost – $50
- ROAS – 2
- Branded
- Conversion value – $100
- Cost – $10
- ROAS – 10
- Total
- Conversion value – $200
- Cost – $60
- ROAS – 12…Wait a second.
If we could multiply our cost by 12, we would have returned $720 in value. Not a measly $200. The real ROAS calculation would be $200 / $60, or 3.33.
To properly have Looker deaggregate your data, simply encapsulate your calculated fields in SUM( ). You can find a better explanation on the Looker Help Forum.
