Today I would love to hear about your personal experience back when you were learning to implement, use, or maintain Adobe Analytics. What are some of the things you had to figure out way too late by yourself that you could have used years ago?
For me personally, Counter Evars are definitely among those things. They are super cool and helpful for so many different use cases but hardly anyone ever talks about them! Especially when combined with creative expiration settings those variables blew my mind more than once.
Looking forward to your input!
Topics help categorize Community content and increase your ability to discover relevant content.
If you are implementing in 2022...
1)Prop's aren't that useful any more. 99% of what they do can now be done with eVars which are superior (in 99% of situations).
2)As @VaniBhemarasetty has mentioned - understand evars. How they work, expirary, serialisation and merchandising.Evars are so important.
3)Once you have your BRD and SDR, I would work with your product team to ensure your data layer can pass as much as possible with as little calcuation being done by Launch as possible. This will also make it easier to transisition to server side at some point.
I am in agreement with @FrederikWerner about Props. I think their value is highly overlooked these days.
One of the things I just wish I had not relied upon as much in my early days was counting "instances" for certain things, like navigation clicks. I truly wish I had enforced the implementation of a specific event with the navigation click in order to get an absolutely specific count, and then things would be accurate. Because I'm not in a position to have the control to manage that anymore, I cannot fix it, and it bothers me to this day.
Ah. Back in the day. Mine is a publishing company. They originally ignored evars (those were for commerce sites). And there were only 50 props available (content sites are heavy with variable use) so we doubled up adding two different types of values to a given prop where they would not conflict. (This is bad from a Data Architecture perspective).
Even through we now use evars regularly, have many more variables at our disposal, and have classifications, we live with that early implementation's legacy.