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!
I also am wondering why can't we use calculated metrics in segment builder. We actually are leaning more and more on calculated metrics, to the degree that we are converting our main KPIs into calcualted version of themselves that can be curated as errors/issue arrise. For example, if cart views were to be comprimised by a bot, we would want a calculated official version of *Cart Views* that we can curate and remove the anomolous data. From there any segments based on cart views need to get re-engineered to include the new calculated *Cart Views*. But today this isn't possible.
The arguments I recieved from Client Care are: #1 Since the calculated metrics are based on the actual metrics, it usually is not required. #2 Also, as calculated metrics allow segments, if segments allow calculated metrics as well, it will be a circular formula. This could cause problems.
My counter arguments to CC are:
#1 What about when the calculated metric is very complicated. And, if we planned to hide the root metrics in the report suite setting menu in favor of the curated version, then it would become impossible to re-create the calc metric at that point to import into segments.
#2 Understanding there is the risk for circular logic, but not in every case, obviously. And couldn't Adobe simply display a warning for any segment where an imported calc metric is colliding/circular?
You are definitely right on point #2. That is a risk, especially if you get more than one level or recursion (e.g. seg#1 uses calc metric #1 which uses seg#2 which uses cal metric #2 which uses seg #1). Though one we might be able to mitigate that one. The other issue is that not all calculations make sense on a visitor, visit or hit level. For example, bounce rate for visits is either 1 or 0, average order value for a hit container is also meaningless. Only at the aggregate level does it become meaniningful. There are tons and tons of these types of formulas that it can be tricky and not super obvious unless you really understand what is going on under the covers.
That being said I definitely agree that you should be able to do math in segments (e.g. show me people that 5 visits with an order or people whose average order value is more than $6).
What types of segements are you trying to create with calculated metrics?
An example we would use is a segment of visitorswherea specific calculated metric occured. Following my theme in the above example, that calculated metric might contain a cart view. (So Visitors that had a Cart View). That cart view raw metric would be curated, and could have some math or might have some corrections based on bug/BOT behavior on the site. Or no changes at all, but is set-up as a place holder for future-proving. Make sense?
Yes, i have one problem with segments, where i need create segments using calculated metrics which is madeup with unique visitor. But Segments doesn't allow calculated metrics to use.
The reason i need to create segments in Adobe analytics is to match the segment value with AAM segments. Because AAM calculates count by unique visitor, where as Adobe Analytics doesn't provide options to create segments using unique visitor.
Please help me is that any way we can replicate the segments in AAM with AA
I hope Adobe will consider this idea! I'm having a similar issue-- I'd like to create a segment to see only search terms that had a CTR (calculated metric) of 0 (I manage our intranet search).
Hello-- Gigazelle recommended I add my use case here. This is the original thread where being able to add a calculated metric to a segment would have solved the issue hassle-free. There is no work around: See Unique Variables where CTR = 0
I need to filter out any search terms where the CTR (calculated metric) is over 0. By creating a segment only using clicks, and not using my calculated metric for CTR, it does not show me unique search terms with 0 clicks. As an example, 100 searches might occur for "Paper," and only get clicks 99 times. Because there was 1 time that a "paper" search did not have a click, according to the segment not using the calculated metric, the term "paper" would still show in the row.
I'll add another use case. For A/B testing, it's useful to be able to extract outliers for metrics such as Revenue. The ability to segment out automatically data points that are below the first quartile or above the third quartile would be extremely helpful.
