I’m gonna fold this because it’s likely a personal “Quint from Jaws is drunk and ranting about what’s stuck in his craw” type of post that might be way too much stream of consciousness blather if you’re not interested.
(Hanon is not actually drunk...)
The other factor which I’ve already sort of mentioned - Fallen London remains popular because it’s a game you can play for months and years due to the metered pace of play and the vast amount of content, and they way Failbetter craftily structured “grind” mechanisms as gameplay and made them interesting. Often to advance in a story arc, a player would have to repeatedly attempt an action which had a chance of success or failure based on “qualities” which are basically variables representing physical resources, relationships, mental and physical state.
Due to the finite number of turns, this “chance to succeed” in some cases might have the addictive real-world allure of a slot machine and consequences: if you don’t succeed the choice repeatedly and your candle goes out, you’ve got to wait and return to try again. Sometimes choices would offer a better chance of success or a guaranteed chance of success if a player spent extra turns or burned precious rare qualities on a specific choice, so risk-reward.
Variables are commonly used in most IF, but a QBN (quality-based narrative) often relies on thousands of them. In an RPG you might manage a pageful stats like hit points, armor class, STR/INT/WIS etc and be aware of all of them throughout the game, in a QBN like FL they are not finite. Exploring a graveyard and meeting spirits there might produce a random new quality you’ve never heard of called “Haunting Essences” and start increasing it the more you grind. What the heck is that? Is it good? Is it bad? It’ll likely do something. One of Failbetter’s stated philosophies was to make what would be considered a failure state and a game over in other narratives just as interesting as success. As I spend time in this graveyard to gain Haunting Essences (which turn out to be a social currency in the underworld) if I occasionally fail a stat check choice, I might instead start accumulating “Terrifying Shocks” and if that gets too high, I won’t be allowed into the Masquerade Ball because I’m a raving lunatic after hanging out too long in the cemetery. But now I’m a haunted terrified lunatic seeing ghosts, which give me cred with the Society of Psychics since I believe in ghosts and a new storyline will open up.
You might think of a QBN as the IF equivalent of a “loot” game, only with interesting and unique quality variables instead of treasure. These thousands of qualities eventually shape the player’s character into a unique individual - I can’t go to parties because I won’t stop talking about ghosts, but now I’m a ghost-hunter. And I might also garden and bake bread, so my character’s individual standing in this enormous world is unique.
To get players to commit to that kind of narrative and come back several times a day requires a large amount of content. Fallen London had that and developed a fanbase. The idea was that players who liked the QBN style might want to play a similar but new games. What you’re seeing left on the site were the few larger moderately successful worlds - I think Zero Summer had a team committed to producing content. But the Directory of Worlds link which has nothing behind it now originally had maybe 30-40 small experimental playable worlds being developed and tested as the authors slowly figured out how to make the system work, and the scope of creating the content. I think we were just getting a handle on it when they shut down. I actually entered a Storynexus game into IFComp back in 2013
The other factor with creating a game system of any kind is you really need a “weenie” to draw people to it - to accumulate players and authors who go “Hey, I want to make something exactly like this! How do I do it?” It’s why when people announce they are creating a revolutionary new IF system the first question is “Where is the example demo game so we can see what this great system is capable of in action?” For Storynexus this was Fallen London. For Gruescript this was Detectiveland. For Inform 7 I believe this was Reliques of Tolti-Aph (and basically every Infocom game that Inform provides tools to replicate in style). If someone is going to make a Storynexus-alike, it needs to be compelling enough to attract players and authors. If Fallen London - possibly the longest-running persistent choice-based story world in existence - tried it and failed, it’s a hard road for any developer.