Honestly, I don't see very many conlangs these days and I see that normal languages (Such as Welsh or Irish) are much more widely used? Why is that?
Motivation is an essential factor determining language vitality. For example, the number of Hebrew speakers increased from zero to 9 million with a major spike in the past 100 years. Likewise, Sanskrit has enjoyed a significant expansion in popularity over the past century, though the existence of first-language speakers is unclear. Native American languages are also enjoying a bit of a revival, so that there are many more speakers of Lushootseed now than there were 15 years ago (again, though, none are fluent first-language speakers). A significant feature of all successful language-preservation and revitalization programs is wide-spread use including use in education, as in the case of the various Sanskrit universities of India.
Esperanto has a reasonable number of speakers including a few hundred native speakers, but this is not a geographical community. One might want to discuss with Esperanto speakers their motivation for speaking Esperanto, and compare that motivation to the motivation of speakers of Hebrew, Sanskrit, Lushootseed as well as various language with low numbers of speakers which did not have a phase of being a dead language (e.g. Welsh, the Saami languages, numerous languages of India like Dogri, Meitei, Tulu).
Many conlangs are structurally under-developed, where aspects of the language's grammar are not fixed / public (e.g. there is no reference grammar of Dothraki), so you might want to look at this list and pick appropriate languages that you think might reasonably be spoken by more people. One problem with the expansion of Barsoomian as a language is that it requires speakers to be telepathic.
Conlangs (usually artistic languages) are widely used for the sake of giving depth to a fictional world. Examples include Klingon from Star Trek, Na'vi from James Cameron's Avatar, Dothraki and High Valyrian from Games of Thrones, and the Elvish languages of Tolkien's Middle-earth.
I assume that what you're actually asking is why don't we all speak an international auxiliary language (IAL), as was the goal behind Volapük, Idiom Neutral, Esperanto, etc.
In a way, we already have one. It's called English, and is already widely used in international business, science, and technology. English is one of the six official languages of the UN, the two official languages of the International Olympic Committee, and the language of international civil aviation.
The main objection to using English as an international language (besides technical aspects like its confusing spelling system) is that it's not culturally “neutral”, unfairly privileging the Americans, British, Canadians, Australians, Irish, and New Zealanders. But hypothetically suppose that some international organization did set out to make a constructed language to use as an IAL. It would be a challenge to make one that's truly globally neutral in terms of phonology, vocabulary, and grammar; as easy for Japanese or Swahili speakers as it is for French or German speakers. (Existing proposals tend to be Eurocentric.) But since this is a hypothetical, assume that a Perfect IAL does get invented. Odds are, some countries would be more able and willing to devote resources to teaching the IAL than other countries would. And then we have a world with disparities in knowledge of the international language — which is the exact situation we find ourselves in with a global natural language. So, the project would kind of be self-defeating.
Of course, you're free to take your own initiative to learn Esperanto and have a global community of 100,000 or so L2 Esperanto speakers to communicate with. But OTOH, you could learn any one of the top 8 natural languages (Mandarin, Spanish, English, Hindi, Portuguese, Bengali, Russian, and Japanese) and have over 100 million L1 speakers to communicate with. Which do you think the average person (as opposed to a linguistics nerd) is going to choose?
Because there is little compelling reason to learn them.
Consider the very specific example of Toki Pona. It’s a relatively simple language, and most people can learn it at a basic level in a matter of days. However, even simple conversations can be difficult in it because it has such a small lexicon (fewer than 150 words, despite technically being a properly complete language), so it’s a pretty poor choice as an auxiliary language. What this ultimately means is that there are comparatively few speakers (at most about 50000 worldwide by my estimate), because there is not really a practical reason for almost anybody to learn the language other than saying they learned it.
And that last bit is the key here: Anybody learning a second language does so for a reason, and there are precious few reasons for learning a conlang.
In general, 99% of people who learn a conlang do so for one of three specific reasons:
- To use it as an auxilary language. This was the design purpose of many especially famous conlangs, such as Volapük, Esperanto, Ido, and Interlingua.
- Because they think it’s cool or fun. This accounts for most learning of conlangs that are not auxilary languages, especially languages like Ithkuil or Lojban that are built up around specific goals other than becoming an auxiliary language.
- As a means of social currency among specific groups (IOW, because a large group of people think it’s cool). This is the main reason that people learn fictional languages like Klingon, Sindarin, Dothraki, Goa'uld, Mando'a, or Na'vi, being able to speak even just a few words in these languages is a relatively fast route to being recognized as a super-fan in their respective fandoms.
For the case of learning a conlang as an auxilary language, you run into the bootstrap issue. An auxilary language is only useful if it’s widely spoken, so even conlangs built to be auxilary languages are essentially dependent on speculative learning by prospective speakers (IOW, they are only successful if enough people think they will be successful that there is a reasonably large community of speakers). But, there’s an even bigger limiting factor here: Most people never even hear about these languages, so the pool of prospective speakers is relatively small. Part of what made Volapük so successful early on was how quickly knowledge that it even existed spread (compare Universalglot, which had similar goals and predates Volapük by a decade, but never really caught on).
For the other cases though, you run into even more issues. The prospective groups who might consider learning such languages are very small to begin with. Ithkuil, for example, probably has fewer than a thousand ‘speakers’, and exactly zero people who are fluent in it (due to the extreme complexity of the language). Additionally, a lot of conlangs that are not designed to be auxiliary languages are very complex by most standards (like Ithkuil or Lojban), or very limited by most standards (like Toki Pona), which limits prospective interest even further.
Fictional languages have a slight advantage here in that they tend to have much bigger prospective audiences, but many of them are also not full languages. Klingon is a bit of a strange exception here in that it’s one of the few fictional languages that has been extended into a complete language (and there’s even an opera written in it), but most other fictional languages consist of a a few dozen words and phrases, and don’t even have a complete grammar, let alone a usable lexicon, which essentially automatically excludes them from most people’s lists of languages they speak.