Should I use Custom Dimensions or Metrics for these variables?[ Edited ]
December 2015 - last edited December 2015
I have to spin up some quick GA tracking for a very simple site that allows users to search and buy books, and I don't have time to learn/implement Enhanced E-Commerce. So, for the time being, I want to augment some of the standard pageview data with events, custom dimensions and metrics.
Some of the use cases are:
- User views search results page: I want to know for every pageview, how many results were displayed
- User views book page: I want to log the Price of the book that the user viewed
- User completes book purchase: On the "thank you" page, I want to log the Price, Shipping Cost and Total Price that the user purchased
I would then use this augmented pageview data, combined with event data, to answer questions such as: (a) how many search queries returned zero results?, (b) of those pageviews that returned > 0 results, what was the CTR into the book page?, (c) what was the purchase rate of books with price < $10 and books with price >= $10? (i.e. detail view to "thank you" page).
I'm having a hard-time if these numerical variables (result count, price, etc) should be sent to GA as a hit-level dimension or as a metric.
My understanding is that these should be dimensions, since they are characteristics of the item that the user is viewing or transacting with. While metrics would be reserved for variables that increment, such as "Number of Books Purchased", "Money Spent", "Number of Books Shared", etc.
However, if I try to create a report and I wanted to do something like filter by Price > 10, then I can only do that with metrics and not dimensions.
What is the right way to think about this? In the context of being able to answer those questions above.
Re: Should I use Custom Dimensions or Metrics for these variables?
as per your use cases I would set it as:
1. Custom dimension (as I do not know what to do with the search result count as a metric
2. Custom dimension as you are giving context to a product (you may opt to track value buckets instead of real value such as low, high, mid)
3. Custom metrics - just make sure the user cannot have multiple hits for the same purchase which would inflate the numbers on each hit