Join us January 15th for an AMA with Champion Achaia Walton, who will be talking about her article on Event-Based Reporting and Measuring Content Groups!
We have a number of different products and it is easy to count the number of product page views or visitors with X or more product page views. However, it would also be good to know how many visitors looked at multiple products. For this we would need a count distinct (product identifier) capability that isn't available yet.
Would such an addition be viable? And if not, what workaround could be recommended?
What I have in mind is a metric like the eVar instances, only that it is based on the distinct value eVar instances. That would work in segments and reports likewise.
Did you try implementing Event Serialization feature for each product view? Use productID as your unique_ID. This should give you the unique product counts
Event serialization doesn't count at the visit level. So, we can't answer a relatively simpe question like "how many visits did we have where people looked at 2 products?" I've worked on several different products' implementations now, and this is continues to be a question of every team focused on the ways users engage.
We can't possibly fit session ID and the product ID in to the 8 characters upon which serialization happens.
Segmenting by instances of a product view event doesn't account for the fact that many will hit back to Google and then come back in while comparison shopping or the like. Or maybe they'll click a link for more info and back.
I think this is actually available now, using the new calculated metrics features that we added in June. Try creating a metric defined either as Count Rows(), which does not take any arguments, or use the Count() function and set the argument to the relevant metric (Page Views, or Product Views, or etc.). You'll get back a number that is the same in each row, but it represents the total number of rows for the given dimension for the given date range. We use this pretty heavily ourselves to count true unique visitors (people) based on user login (in an eVar).
You can also operate on that metric by including these functions in a more complex calculated metric.
I'm going to tentatively mark this as "Implemented" but comment here if the solution I suggested doesn't work.
Hi Ben, I looked at that but am not sure it's what I'm after. To quote Danielle above: "how many visits did we have where people looked at 2 [different] products?"
We'd want a histogram (# visits with only 1 product, # visits with 2 different products, etc). Alternatively, a variable that can be the basis for a segment (a segment consisting of sessions where there were 3 or more different products viewed).
I saw that and I wasn't sure how tied to that use case the general idea was. But that makes sense. We are actually working on histograms; when we add them, they will be in Analysis Workspace only (just FYI).
