He doesn't address the obvious question: Why? Why for the end user? Why for Jolla?
For users, a far better user interface (still prefer the user interface of the N9 over android one), a new app ecosystem (work most android apps, plus jolla native/meego/mer ones, maybe will be a compatibility layer to run ubuntu touch/tizen ones, and as it is based on qt/qml, probably will be easy to port BB10 ones too. And of course, ports of linux apps in general (the N900 had a lot of them)
For Jolla is, of course, improving installed base, enabling people to try, or even improve the OS, and generate a critical mass of developers and fans that will make the rest of potential users to see with good eyes the phones with jolla built in.
Remember, part of android success was that people was able to install cyanogenmod in their old windows phones. If this is more efficient than android giving new life to old android phones could give enough visibility to the new OS.
Not really, no. Only a tiny fraction of Android users even know what cyanogen mod is, let alone install it on their phone.
That's not at all what made Android successful.
Jolla's apparent killer feature is that it runs Android apps, like, erm, Android. It's not really any more open or cheaper than AOSP.
The best chance they have to gain traction is to attract users. You can buy marketing until your budget is gone, but that only goes so far; word of mouth is almost certainly the best chance they have to expand their user base, improving their brand awareness, drawing new buyers.