We're talking about company wide project management where engineers are on the bottom. Engineer's workflow is usually quite simple and limited to assigned issues. They'll never be happy with Jira UI as it's indeed, um, crappy but something like a CLI app using standard REST API with zsh completion may be enough to make most of them happy. Engineers mostly don't use Structure, BigPicture, filters, reports, they don't make complex workflows etc. Oh, maybe the kanban board.
And that's exactly the problem with Jira. It's unsexy, the REST API lacks important functionality (like adding a new option to select list custom field, but one can write custom endpoints in Groovy using ScriptRunner and Java API), it's relatively slow, the code base is old, some issues are never going to be fixed etc. But it has the app ecosystem and the community which is really what makes it, sort of, best in class as apps cover most users' cases. The question usually is 'what app can do this?'
I think providing a viable alternative is quite a challenge as that means building an active community and ecosystem. Plus, imagine a company invested quite much into building their processes on Jira with some set of apps. They have scripted custom fields, workflow post-functions, custom events, event listeners, timesheets, planned vs actual reports, time to market reports, SLAs, Structure, custom REST endpoints, behaviors and other ScriptRunner stuff. An alternative would have to somehow provide all that either out-of-box or via extensions, plus a feasible migration path. What benefits would justify such migration?
It seems that Atlassian know that and therefore reshape their business accordingly by raising the prices to cut away the small fish and move such customers to the cloud. They may loose many of them b/c there are better tools for simpler needs but it's really difficult to beat Atlassian on the on-premise cover all solutions market. They've been there for quite a long time.