I have a question regarding normalization. Here is the scenario,
- I have 2 website template offerings (i.e. skus), say one is called "Metro" and other is called "Relic"
- I have 10,000 "Metro" sites and 5000 "Relic" sites
- In google studio, I create a chart with "Time" dimension, breakdown dimension as "Template" and metric as sessions. This gives me two lines per SKU over time
The issue is that it shows "Metro" performing better than "Relic" because "Metro" is double in size. I am trying to figure out the best way to come up with the formula that shows performance of "Metro" vs "Relic" without size effect. I tried creating two separate calculated fields using CASE and check the "Template" dimension and dividing its size, so "Metro" would be `sum(all sessions of type "Metro") / total metro sites`. The issue is that this isn't a true indicator of performance because it will show Metro far lower than Relic, even if Metro could be performing better than Relic.
Do you guys have any idea on how to go about approaching this?
for you, what is your KPI (key performance indicator) when it comes to "Metro" vs. "Relic"?
Can you also explain a little bit more how the setup works? What does it mean - "you have 10,000 'Metro' sites and 5000 'Relic' sites"?
Hope that helps.