I think it would be worth looking at this using "degree days". (Degree day stats from Wunderground, I eyeballed his KW since no firm numbers are provided).
* August 2011 vs. August 2012: 809 cooling degree days vs. 594 (1.36x). 2100 KW vs. 1500 (1.4x)
* January 2012 vs. January 2013: 544 heating degree days vs. 595 (0.91). 3000 KW vs. 2000 (1.5x)
So accounting for the weather with "degree days", with A/C it looks like there was no savings, but with heat it made a massive difference.
* August 2011 vs. August 2012: 809 cooling degree days vs. 594 (1.36x). 2100 KW vs. 1500 (1.4x)
* January 2012 vs. January 2013: 544 heating degree days vs. 595 (0.91). 3000 KW vs. 2000 (1.5x)
So accounting for the weather with "degree days", with A/C it looks like there was no savings, but with heat it made a massive difference.