Celtic have the best defensive record in the SPFL both in terms of goals conceded (15 in 26) and in xGA (0.65 per 90m).
Despite this Celtic are on track for conceding a goal a game overall (41 in 44) mainly due to early season European profligacy.
The Mid Half Blip
More specifically, the recent 3-2 win at Aberdeen saw 2 goals conceded in the period between the 50th and 70th minutes. That seems to be when Celtic’s all action pressing style slows down, prior to substitutions being made.
At least that is what my eyes were telling and I put this graphic up on the latest Huddle Breakdown episode (please subscribe!) to support that:

The red line shows when Celtic concede this season. The Black line is this seasons SPFL (over 300 goals) and the Green line Celtic’s average over 2014-21. Those latter two lines trend very close.
The overall trend is that teams are more likely to concede the further into the game you get, logical as teams tire. Also, there are spikes at near half and full time as “time added” counts as the 45th and 90th minutes.
But the Red spike between 50 and 70 minutes is very pronounced for Celtic this season.
xG
You are learning that actual goals conceded and scored is, whilst not random, a relatively poor indicator of performance.
Expected Goals offer more performance-based trends. So, for this hypothesis to be true (Celtic are vulnerable between the 50th and 70th minutes) then the xG trends would back that up.
So, here they are:

Once again, the trends are consistent – rising gradient towards the 90th minute, but with some deviation.
Most eye-catching was the level of chances given up in 18/19 either side of half time. Celtic only conceded 38 goals in 63 matches that season so not sure what was going on!
The Red line is this season and there is indeed a (more modest) spike from 50 minutes for xG conceded. It is certainly higher than that period in any other season.
Reasons to Worry?
Like Postecoglou I can always find things to want to improve on.
The reason for this spike, I believe, is that the manager is sometimes slow to bring substitutes on. Now, I will hold my hand up and say to do an analysis of when subs come on versus game state would be a mini dissertation that I don’t have time for. So, I am going with eyes and gut on this one!
We know the bench has been thin for much of the season due to injury. However, this situation has eased. As the team plays such an aggressive and physically demanding style, it is natural players tire into the second half. So does the opposition!
Where Postecoglou can improve is sensing these momentum shifts and getting ahead of them. A good example was the Aberdeen game.
When subs do come on, Celtic are remarkably efficient at closing games down. They have conceded just ONE goal in the last 20 MINUTES of SPFL matches and that was the first game of the season at Hearts (John Souttar’s header). Incredible. Subs do work.
Recent matches against Aberdeen, Hearts, St Johnstone and even Alloa have began to drift in the second half before action is taken. With 5 subs allowed, there is really no excuse. Celtic, with the largest wage bill, have the strongest bench (that’s how money and football works – honest!).
Dare I say, also, that earlier subs may be preventative as regards the huge injury list Celtic have carried this season?
So, small adjustments, but those margins could be crucial at seasons end.