This data is also extremly valuable for journalists. I can envision someone becoming a better journalist by optimizing for lower refunds. A good journalist will use this as a weapon. Quite often you can't really be sure about the quality of your piece. Now you have an interesting data point.
If Blendle continues to be successful this could get very interesting though. As described in the article linked here, it's a great metric. And their publishers will also start to use their metric. So -while they now describe the Blendle-revenue as an extra- this will actually change the way we deal with the 'print' data which is now accessible online, creating a new dynamic. This will surely have an impact on the print version of these magazines as well, I can imagine, making Blendle a content driver(shaper?) and not 'just' a source of 'extra revenue' as they tend to see it now.
If it all takes off impact will, imho, be different than we can probably predict now, even though we can make some estimates of possible directions. Some people will be happier with the results after -say- five years than others.
Time will tell.