The S7 will never achieve ROI at .15 cents/kwh - unless bitcoin price rises a lot more than it has been doing, or diff suddently stops increasing.
Even at a 3% average diff increase and 6.7 cent / KWH power cost, the S7 costs more than it will ever return - and recent diff increases have been a LOT more than 3%.
I'm not sure at this point if an S7 will be profitable AT ALL past the halfing, unless your electric is in the UNDER 5 cents/KWH range.
if you're talking about 3% diff increase per month, it's more than doable, look at the antminers s5, its profit remained at 0.01 for a very long time, at least since january which is 9 months already, this is more than enough to roi with everything out there
only recently , dropped to 0.009, so the whole diff increase is a bit overestimated...
3 % diff increase per increment, or about double that per month - the last couple months have been closer to 5% per increment.
At 3% the S5 has ZERO profitability after the halfing, and it's noticeably UNDER .01BTC/day even now on GROSS income unless you overclock it a LOT.
The December-July or so timeframe saw an average diff increase of appx. 1.6% - MUCH flatter than the last couple months, and VERY ABNORMAL for Bitcoin in general to date.
If you bought an S5 late last year or very early this year, it's probably hit RoI already and will make you some decent money - but the market has CHANGED since then, if you don't already own the hardware, get it VERY cheap, or have VERY CHEAP electric your probability of achieving RoI on currently available mining gear is almost zero.
Don't forget to factor in cost of power supplies too, that can make a noticeable difference - though at least most power supplies can be "recycled" to other usages after your mining gear goes unprofitable or you sell it.