Re: Failure or test scenario?
"Why not just turn it on and see if the system works with sats that are in such wonky orbits?"
Because "in such wonky orbits," unfortunately, the orbit is likely to be rather unstable, making constant re-calibration necessary. In the intended orbits, the satellites already up can help guide the other satellites into their proper spots, without constantly spending reaction mass that was originally intended to do course corrections for the next 10 years.
Basically, by right now spending approx. half of these two satellite's reaction fuel, they can save many times that for the (literally) up-coming satellites. The end result is most likely going to be that replacing these two birds earlier than planned is actually going to save the ESA some money. Sounds weird, but then again, this is Rocket Science ;) -- there are a couple of lessons available here (and the people at DASA are already working on them) which may make future additions to geolocation satellite systems more efficient.