I'd be looking through all my balances on my reports and seeing what was where when, it's more than just the one AU from what you describe. As well as all the other suspects that have been suggested, 12/11 and 12/14 reports are darned close together so usually can get close to determining a discrete change.
Yeah there's a lag on inserting data with the bureaus unfortunately, and EX is probably the best of the 3 on that front.
Unless it says 30 or worse on the tradeline, it's not late from a FICO perspective. If it has Term 1 month it should discount it from revolving utilization but not from number of revolvers with balance, that's it.
@newhis wrote:2 interesting things happened the last few weeks:
- another AU reported after reaching 800 and back to 784
- the AU is Chase, paid to $0 mid December and CK still reports the balance Jan 3rd
For some reason DWs score drops when one of her AU cards report a balance (this time $35).
Chase didn't report the $0 balance for AU, sorry I don't have a way to check if they updated my report.
I'm paying all AU cards before statement close. Let's see if she gets her score back up.
I guess as history increases (AoOA and AAoA) then the score will not go up/down so many points.
I'm trying to only 3 cards report a balance (no AU) and only 2 by March.
Someone's AU finally counts? 
IF you can check, what's the last report date on the Chase card for the primary cardholder? I have noticed if I have charges running through even if I zero it out it may skip the report... I need to basically let the tradeline idle to ensure it but it could just be a Chase bug too I suppose, only happened twice but seemed consistent.
It would be interesting to see though if the primary accountholder got the midcycle report and the AU did not, that might actually be helpful in some situations to know for others.
As the file gets thicker on the clean side, not sure on the up/down so many points; I have some guesses, a working theory of mine (and this is not supported by fact) is that the wider the scorecard is in absolute numbers, and I think some of the clean scorecards are larger than the dirty ones I've been in, that the corresponding changes for any little thing are magnified. That fits what we typically see, where something like maxxing out an individual tradeline on my file isn't particularly awkward score wise but we've had higher complaints on others, and also my installment utilization is another part, 22 points vs. 35ish.
To be fair I may stuck up against the edge of a bucket, maybe, for having a bunch of negatives likely sorting me into the 3rd or 4th dirty bucket if I had to guess, my file is pretty solid compared to the majority of dirty scorecard players I would suspect. Too many suppositions and anecdotal data but it's a theory, unfortunately can't really test that.