It's common for folks to say "don't worry about the commercial success of the sim" and enjoy the driving/experience. But unfortunately, my experience at iRacing suggest these two things go hand in hand, and its why I frequently bring them up.
In my case, I'd hope for a good experience in what I'll call "open world" online racing. This is different to leagues, which I'll call "closed world" - i.e. limited numbers of participants with specific goals/scenarios they are trying to emulate. I've helped run successful rF1 leagues, and know just how difficult that was, finding a group of people willing to compromise on all the competition aspects and join together for either one-off races or championships. Thinks like, race length, how/when to qualify & how many laps, use of rain/weather conditions etc. etc. There are so many options, getting consensus on these is almost impossible with a small group (say 50 or less). Its much more likely with a big group that you can find subsets who will compromise to give a result that "works" for most, or many, enough to run such events. At iRacing, even with 100,000 members, trying to run some series like IndyCar has been an exercise in frustration. Typically they've had only about 30-50 people committed enough to do a full season in the road/combined series (less than 10% of those who compete in one race per season at least). You really have to get the numbers up to get enough for some niches. Then if you want to have a skill-based matchmaking system on top of it, and get enough "splits" to create good groupings of like-skilled drivers, then you need even more. This is why you need many thousands of people online.
I'm sure there are many games where there is not much online activity, but they are still a sales success. I'd bet this applies to the Codemasters F1 games for example, and Gran Turismo. But to run successful online, its impossible to not have significant sales.
When we ran our rF1 league, we piggy-backed off another organizations developments of tools to run championships, servers etc. which primarily promoted sportscar racing (and was run by/with many of the best sportscar mod makers of the time). Our IndyCar and historic F1 championships were much lower participation rate. There was significant work to promote the league/find drivers, decide on the race schedule, and get the cars/tracks we wanted to run in a position to be used (testing etc,). For example tracks, we would often have to fix pitstalls and such. On the car mods, we'd go from 'cosmetic' things like changing the engine sounds or mirror resolutions, to more extensive tire and engine performance changes. For something like an IndyCar, the car performance for things like turbo-lag were not something directly modelled, but done via engine torque curve changes. This meant the torque curves weren't "real", but provided a similar net result to that seen in the real world, at least directionally. I'm sure some of the best rF2 leagues still do a lot of this; and even S397s own GT competitions, they have bespoke BOP packages not applied to the 'standard' DLC content the rest of us use. It's this stuff that has to stop; because you are only fragmenting the userbase by doing so, and creating 'cliques' and 'insiders'. Put everyone on a level playing field, and the (online) sport will grow. That's been the lesson of iRacing.
