We noticed a general issue with how Adobe processes its data. If 2 Hits happen in the same second - e.g. a tracked click on a Teaser link and the Pageview of the next page, Adobe Analytics seems to order them in the wrong way - or in a random way. This leads to problems with Fallout Reports and Sequential Segments because the assumption "User first clicked link x (Link Click was tracked), then viewed Page y" will not work!
Example:
We want to create a fallout report where a user clicked on a tracked Teaser Link on the Homepage that leads to the Page X.
Since this teaser link leads directly to the X page X, everyone should be in the Fallout Step 2 (yes, apart from some very rare cases where the next page loaded very slowly and the user left the page too fast). But we should not see a drop-out rate of 39.3%!

When we look at a subset of the users who clicked the teaser, we see that some of them appear to first have viewed the destination page and THEN clicked on the teaser that leads to that destination page.
This of course is "the other way round" and thus near impossible.

However, when segmenting for "
Visits with Element ID ab0015-case-2a-..." almost EVERYONE who clicked also saw the Page X in her Visit (122 to 120 Visits)!
So they did get to that page somehow.
When we looked at these cases in detail (eg the raw data), we saw that the timestamps of the click and the ensuing Page View were identical (same second) which confirms the hypothesis.
See this extract from the raw clickstream data which confirms this hypothesis. Both users had the click Hit and the ensuing Pageview Hit in the same second.
See hits at 10:11:51 (User 1) and at 16:46:03 (User 2):
Adobe then does not order by the time the Hits were received, instead it seems to simply put Event Hits after Pageview Hits. This is a major flaw in the processing logic and needs to be fixed very soon! We had to alert all our users to stay away from Fallout and Flow Reports due to this!