You may be right, Jan, if you've had extensive or substantial experience in 3D chess. I'd had little in 50 years and when I went to the chess club to play, the perspective was so different that I had no idea of what I was doing. Sure, I wasn't as good as the others, well, far from it but I simply played and looked like a dolt. I was totally out of my depth. I was like a ten year-old who was comfortable, swimming", in waist-deep water and then jumping into 10 metre-deep, diving pool. *Gurgle*. (I drowned). (Oh, diving from the 10 metre platform is exhilarating but less so if you don't enter perfectly straight! )
What Are Your Goals?


In support of that, comrade, I can recount that one of the new blokes at the chess club didn't reappear because he was flummoxed by the lack of an analysis board.

Can confirm, online chess is a completely different experience than OTB, even if a 60 : 30 game or Daily.

When I started with online chess some years back (then it was mainly tactics training for me), I needed to "reconstruct" more complicated positions on a "physical" chess board. I couldn't see anything when the pieces were so flat! Later, after some time, I found chess.com, and when I tried daily games and vote chess games, I needed to help myself with a real chessboard as well to see the position. It certainly is different to play 2D or 3D, but I can't agree one of them is harder than the other in general. Both need practice, off course.
Now, it only makes sense to compare apples and apples. Online blitz can be compared to OTB blitz, online rapid to OTB rapid and online daily to correspondence. OTB tournaments to online tournaments with same (or simillar) time controls.
Different people recommend different things. When I was going to the grammar school, there was a trainer for smoe time who recommended online chess a lot (he was a FM).
I don’t understand all the statements about online chess being a different game than OTB.
It is not. You still move pieces. You either see good moves or you don’t. If one’s results online are a lot better than OTB, then they’re doing something wrong. Perhaps it’s caused by the fact that people can’t use opening DBs during OTB games, in which case the solution is to build an opening repertoire. 
If you observe differences, it’s most likely due to two of the statistician’s enemies: small samplesize and confirmation bias.