Part of the answer will depend on what you're planning to do when you convert to mobile. Is it just going to be a single mobile stylesheet, or multiple stylesheets with responsive design, or a completely new codebase? Because obviously each of those creates a different amount of work – so if all you're going to do is put a simple skin on the existing page structure then it's worth doing for a lot less than 8% of your visitors, whereas if you're creating an entirely new template and structure then it may be worth holding fire.
You also need to look at (a) if there is any suppressed/latent demand that you're not seeing because the site doesn't work properly on mobiles, and (b) whether a mobile site is actually necessary.
First - suppressed demand. It's been a common refrain of web designers throughout the ages that "I don't need to cater for <x> because nobody with <x> ever comes to my site", whether <x> is a particular browser, or visual impairment, or whatever else. And no, if your site doesn't work properly for Netscape then you won't see many Netscape hits in your logs because they won't get beyond the first page, and they won't come back ... but that doesn't mean they didn't try. How sure are you that people don't actually want to use your site on their mobiles, but are being held back because they can't?