Yes there is: it's the number of currently running processes described by that language. In practice this means counting every tab in every browser on every device on Earth. Even the largest DC operators in the world, Amazon, Google, Microsoft, NSA, etc who run sophisticated bespoke software distribution systems that support all kinds of runtimes, cannot compare in number to everything else. (Although it wouldn't surprise me if one or more of them attempts to maintain a stateful process for every man woman and child on Earth. That's still not enough.)
I take it as you see "success" in terms of programming languages as something objective then, not subjective?
Languages are tools that solve a problem, and it is possible for them to both be used by relatively few people and still successfully solve a problem.
For example, though Elixir is not as popular as Go, Elixir might win for me because I think it has a better concurrency model. Elixir is not globally used more than Go, but nonetheless it is my winner, and the people who pour their time into the product have still done an admirable job.
Every programmer needs to confront that the fact that their opinion about what really matters, isn't the same thing for them as it is for others. And even more so, what really matters changes based on what context you're using the language in, and what you're building.
There is not a single metric that decides if a programming language "wins" or not. What does that even mean?
For me, Clojure won a long time ago, because what does matter is simplicity. But I also realize that this is a personal opinion, not something that I spread as "fact" for others to finally realize.