https://imgur.com/a/ONaMAtz

So then it's
Passing 0.29*0.04= 0.012
Rushing 0.05*0.10 = 0.005 (RB rushing yards only)
Receiving .34*1.00 = 0.34 (this is receiving wr/te, ignores RB receiving so Utah returns effectively everyone - for now at least, their #3 in receiving yards is in the portal)
OL snaps .33*~.90 = 0.297 (Don't have the snap count data, but Umana was out most of the year, and all the backups are also returning)

So then the biggest things I see are that the categories where Utah has a bunch returning (WR/TE and OL) are the heaviest weighted and that the percent returning at WR/TE and at OL is even higher than you thought.

I still don't see how it gets to 81% though. QB alone should bring them below that. My back of the envelope math has them closer to 65%, which is good, but not 81%. My guess is that he missed Bentley transferring, which would amount to a difference of 23.7%, and closely matches the difference between my numbers and his (possible he also missed that Lisk is done or I could have underestimated the OL snap count returning)

He also notes that defensive returning production is more important and that's obviously where Utah has a TON returning.