It shows who did the most work in an unambiguous way. If you have never collaborated on a paper, you'll probably think this is not important, but in reality it is.
Interestingly, some fields where alphabetical order is used have started adding a note that explicitly shows who did how much work.
I think the problem is that it's really not that unambiguous. There are general rules to the order but it's not explicit, and being second author doesn't explain in any kind of detail what was done.
I'm quite a fan of explicitly tagging authors for what they did. That way you avoid deciding whether the person who did the stats did "more work" than the person who did the code, etc.
Explicit tagging is the best option, yes, but contribution-based is significantly better than just alphabetical.
Typically, the first author did most of the intellectual work. Rarely if ever would the first author be a person who did a lot of practical work but little to no intellectual work.
I agree it's largely better than nothing, though there was that paper recently that got pulled because the authors couldn't agree on the ordering of the names.
Interestingly, some fields where alphabetical order is used have started adding a note that explicitly shows who did how much work.