I've been exploring ways to view First Touch/Last Touch marketing channel data, and how the selected date range impacts these reports. I've explored this three different ways so far with reference to one of our custom events, e131: 1) Visitor-level segmentation, 2) Visitor-level Fallout and 3) First Touch/Last touch reporting matrix.
Below, you'll see that the counts between report #2 and report #3 for the month of March 2020 are nearly identical at 536 and 541 counts, respectively. Regarding the report #1 reporting a higher 675 count...I believe this is possibly attributed to the Visitor-level segment, by its nature, having a more broad definition.
So far so good...
However, when I modify the date range to Last 90 days, the variance widens significantly for these three reports, particularly for #2 (1715 counts) and #3 (1480 counts).
Is there an explanation on why simply changing the date range creates such a large variance?