How's this:
Of 199 games Yusupov played starting 1.d4 d5 as White, he played
2.c4 156 times.
Of the remaining 43, he played 3.c4, transposing to QGD/Slav lines 25 times.
Of the remaining 18, he played 4.c4, transposing to QGD lines, 15 times.
So, of 199 1.d4 d5 games, he didn't play a QGD system in three games. That's just over 1.5% of the time.
Hmm okay, so I think that I am beginning to see Yusupov's approach now. If the opponents replied 1.d4 with 1. ...d5 he went for the Queen's Gambit, but if they replied with 1. ...Nf6 he used more off-beat lines like the Colle, Colle-Zuckertot, the Torre etc. Makes sense in practical terms.
This lines don't have perfect classifications or move ordering, however most of these simple openings seem to have potential ... I am not a fan of advancing the c-pawn as white in the QGD since I think the pawn structure can be undermined by black with moves like e5, c6 etc.