A link popped up on Reddit showing the relationship in each of the major sports leagues between wins in one season to the previous season. Originally, there was no data posted there for the NHL, so I made the same sort of scatter plot for 2005-2013.
Teams just haven't shown many wild swings, for better or worse, from the lockout in 2003-2004. Suter leaving Nashville certainly hurt, leading to their worst finish in 10 years. Philadelphia fired Ken Hitchcock, traded away Peter Forsberg, and got sub-.900 goaltending from Niittymaki. In the same year, Pittsburgh made the era's largest single-season leap when Evgeni Malkin joined the team.
The average standard deviation for each team was 11.80. If you back out the first two years of the salary cap era, and the lock-out shortened season last year, that number falls to 8.55. This means that in a given year, the average team would see their win total fluctuate by about 4 or 5 wins per year. Between 2007 and 2012, the most consistent teams, for good or bad, were Calgary, Pittsburgh, Winnipeg/Atlanta, and Toronto. Tampa Bay and Colorado were the least consistent, fluctuating between good and terrible.
The correlation between any given team's point total and their total the following season is just 0.4602. The link isn't particularly strong. This may be a result of the salary cap causing talented teams to disband their rosters, like Chicago's fire sale after winning their first Cup, or of swings in shoot-out luck, like New Jersey dipping from 12-4 in their Cup Final season to 2-7 last year.
Would the correlation be stronger if there weren't a hard salary cap and shoot-outs to worry about?
Looking at the ten seasons before the lockout supports that idea. The tighter cluster of dots shows a stronger correlation in point totals from one year to the next - stronger by about 35%. Even before analyzing the data, we could have guessed this would be true, recalling the powerhouse franchises; Detroit, Colorado, Dallas, and Philadelphia all dominated the regular season annually. The expansion teams didn't make many great leaps forward, and middling teams stayed middling.
In these ten years, Ottawa drafted Alexandre Daigle and made an immediate jump to respectability; Alexei Yashin proved the Islanders could outscore their defensive problems; and San Jose drafted Vesa Toskala and were rewarded with six years of year-over-year improvements.
I admit that I'm not as good with interpreting data as some of the other contributors here, so I was a bit confused to find that the variability between seasons pre-lockout was higher than post-lockout, with the average standard deviation of year-over-year point total differentials (of non-expansion teams) was 12.09, or about 6 wins. I would have assumed that if the correlation from year to year was stronger, that there would be less variability. I open that up for interpretation by those more qualified than I.