IMO, a significant confounding factor is that parser IF games tend to be puzzle-adventure games, whereas choice-based IF games typically feature no puzzles, or just a few fairly easy ones.
I cut my teeth on puzzle-adventure games and I quite enjoy them, but IMO puzzle-adventure games (both point-and-click adventures and parser-based text adventures) have a really tough market, because puzzle lovers tend to enjoy harder puzzles that are inaccessible to newbies. If you’re lucky enough to develop a customer base, and you listen to your customers, you’ll inevitably start developing games that newbies can’t effectively play.
You can do work to combat this, including “easy mode,” in-game hints, etc. but historically point-and-click adventure developers have regarded hints as an add-on profit opportunity, effectively charging newbies extra to play, which is IMO the wrong way to think about it. Charge extra for “hard mode.”
If I were to start another line of text-only IF games today, it wouldn’t be a publishing house for parser-based text-adventure puzzlefests; I’d start a line of text-based point-and-click adventures, with user interfaces that look like the sort of thing Robin Johnson has done in Detectiveland or Zarf’s Seltani, and ship layered, progressive hints as part of the product.
(I think a text-based point-and-click UI could possibly work with an optional limited parser.)
I mention all this because I’m sure I’d have no time to implement anything like this in the next year or two, but I would love it if someone were to steal my ideas and run with it!