What are the "other variable costs"?

Publishing costs, server capacity... 

This is what will always happen when the primary cost is based on streams or listening hours becuase the total listening will go up with each added user. The ratio of income to liabilities will never change, and there is never the benefit to costs that comes from scale.

The reverse of this is why streaming services will not pay out more than the hundreths of a penny that a stream is currently worth when they get more users. Each new user brings the same amount of cost with them, and again there is no benefit from the larger scale.

I dont think Pandora is a good short term investment but I do think there is a long term opportunity here, who knows if Pandora will be around when its time truly comes.  Commerical radio monetization is substantial and the more Pandora (or like services) integrates the better chance they have at grabbing the incredibly high CPM radio ad spots.  

The problem with a service like pandora (and where IHeartRadio may beat them) is their lack of localization.  A huge amount of radio advertisers are locally based which is the opportunity at radio.  It's the last medium really geared towards influencing individual markets (I suppose you can include local newspapers as well) which is why they're able to sell at such high cpms.

Pandora needs to focus locally and they'll have an opportunity.  If they can sell small businesses on their ability to influence locally, they'll succeed.  If not, it will be increasingly difficult to crack the CPM needed to be profitable.

I love Pandora, but if even they can't get to profitability...

Hard to believe.