It is a never ending battle to find the balance between providing cool metrics and consumable metrics. Let me give you an example of one of my favorites. Happened again today. (There will be more and I'll post as often as I deal with them I promise. ;)
Let's take the concept of Sales. So, I want to run a Performance Report. Doesn't matter what really. By keyword, by product, by category, etc. For this example, we'll use keyword.
So, let's see the performance of a specific keyword. I want to look at April's results. I want leads, orders, conversion, profit, blah blah blah.
Quick run of the report and I see that the keyword has zero leads and 3 orders.
[User typically scratching head at this point.] "Hmmm, something must be wrong with this report" is the unfortunate, yet all too common, response.
Welcome to the "What does <FILL IN BLANK> really mean?" scenario. In this case, "What does ORDERS really mean?".
Well, as is typically the case, it depends right? We've got a performance report, by keyword, for April.
Does the column ORDERS mean...
1. The total number of orders that have occurred in April, due to leads of the respective keyword, regardless of whether the leads occurred in April.
2. The total number of orders that have occurred due to April leads, regardless of whether the orders were placed in April.
Hmmm again. Two different meanings that answer two different questions. Reporting services may make the most commonly requested answer (#1 I find) the default report option, but anyone expecting #2 and not knowing the option is available, is consuming bad data.
That's not a good thing...
Moral of the story? Cool metrics are only cool when a user gets what they are expecting. Perfecting that?
I'm workin' on it. ;)
Comments