Esperanto didn't catch on because a lot of people feel like it's not the right tool for the job - there are forks and whole different projects that try to mitigate the issues.
The job was to be an easily learnable lingua franca, a common second language. Esperanto (also the forks) is a better tool for that job than english. All else equal, it's much quicker to learn.
Returning to the metric analogy, I think it's interesting to see the whys.
Ultimately, the quality of a language system is not the only or main thing leading to its adoption. There's a massive cultural element. The metric system was ideologically compatible with the French revolution and enlightenment/scientific/modern cultural ideals generally.
It was also useful individually, often by influential individuals (eg scientists). Even if everyone uses another system, it's still worth learning metric so you can do math, engineering or land surveyance more easily. You can always convert units for communication purposes.. I assume it's still like this in the states. If you need to calculate volume, you use metric and then convert to square feet or gallons.
This gave metric systems a bootstrap that esperanto didn't have. This has a parallel to network-effect products. You can't displace Twitter or facebook just by making a better alternative. That only matters once you have millions of users. You must have a bootstrap... a reason for individuals or small groups to use it regardless of overall usage.
The literal "killer app" for Esperanto :D https://www.amazon.com/Aggressor-Maneuver-Esperanto-Language...
To that question, I found this YouTube polyglot explaining why he doesn't think esperanto is a language to learn raises some very good points: (https://youtu.be/n4AklOtfhMo - in french) The one I found compelling is that, in the absence of an economic motivator, you really need to fall in love with a language to remain motivated to learn it well. Where esperanto doesn't really have much of a cultural history or literary tradition, it just doesn't offer most people a lot to love.
Nowadays there's couchsurfing.com, which simply has a larger subculture, and as such the pasporta servo is not that useful any more. But it was a clever way to address that challenge!
I would say that esperanto does offer some inherent reasons to learn it. It's very easy to learn, relative to natural languages. That's what it was invented for. It's fun to make fast progress, and I think this is the biggest motivator for people to learn it.
It's also useful to learn a second language, because learning a language makes it easier to learn languages. There are some school programs teaching esperanto for this reason. It's part of their path to learning english.
Besides that, there's linguistic interest & idealistic interest.
You're right though, that it's all about the inherent reasons. There are no books to read or people to talk to. That's one hell of a downside.
I think klingon and dothraki are good reference points. People learn those for no practical reason.
It's definitely different if you really like the ideal of a language that doesn't belong to any one group of people. But I'm guessing that, nowadays, that's a compelling motivating factor for fewer people than, "I really like Star Trek," is. Klingon's rich cultural history may be a constructed patchwork, but at least it exists.
https://dvd.ikso.net/pagxo/en/libro.html
Note that these are merely the books where the copyright has already expired, so they can be freely put online. There is much more of the fresher content.
Esperanto would probably have to be 10X better than existing languages in some sense to be a viable replacement.
Esperanto is very easy to learn, compared to a natural language. English is the right one to compare with, since that became the international lingua franca esperanto wanted to be.
There are even some studies suggesting that the quickest way to teach english is to teach esperanto first, and the language learning skills acquired shorten the path to fluency enough to be worth it.
I think 10X is true in certain situations, but displacing incumbents can be more about path dependence than degrees of improvement.