Paul Tomkins and Graeme Riley did a huge research project into transfers as part of their book, Pay As You Play. They found that 50% of transfers are a success, and that figure applies as much to high priced signings of elite players, as it does to lower priced transfers. If you go on Transfermarkt and look at a few clubs' signings over say the last 8 to 10 years, you'll see that this 50% figure seems realistic.
Van de Beek was signed using the same process as Maguire, Wan-Bissaka, James, Fernandes, Telles, Cavani, Pellistri and Amad. The recruitment team have a veto, as does Solskjaer. Both parties must have been sufficiently happy to agree to the signing, otherwise they would have used their veto.
Donny was signed as a back-up player. Since the first choice in his position, Bruno Fernandes, was fit through-out the season, Donny hasn't featured that often. Liverpool signed a left-back last summer, Tsimikas, who has barely played for them. Ake cost City £40m, he hasn't played much either. Real Madrid have bought numerous back-up players over recent years, that haven't featured that often.
Teams have a 25 player squad. We buy players to be part of the squad. No-one knows which players will get injured during the season. No-one knows which players will have a Martial season. Suppose we bought Sancho for the price that Dortmund asked for, but missed out on Cavani, Amad, Pellistri and van de Beek, in order to part fund the deal. Sancho creates tons of chances that Martial, our main striker, without serious competition, then misses because he is having a terrible season.