dang look at the 2500k... 5 years old and still only a few frames off the leaders. Who knew you'd basically be set for life with such an economical cpu back in the day? I mean I get that the majority saw it as a steal back then, but even they have to be surprised that 5 years later it hasn't suffered at all especially given that 5 years before that we had Athlon 64 X2's and Cedar Mill based Pentium 4's (conroe didn't release until july of that year) they weren't exactly keeping up in 2011 yet in 2016 the 2500k does just fine.
Yeah it all comes down to the incredible performance jump Intel made with Sandy Bridge. One arch to rule them all, to be honest. I think in hindsight, 20-30 years from now, we will still be reminiscing that architecture as one of the greater leaps in CPU land. So far Intel has been feeding off that ever since, because lets face it, under the hood all they've done from that point is move around different CPU parts between chipset and die and keep everything up to date (PCI 3.0, SATA, M2, etc.).
With that in mind, it makes sense for the Zen release to aim for 'Sandy Bridge IPC levels' because really that's where the holy grail's at. Once they've reached that point, they can eclipse Intel's meagre performance gain over the years with one or two architectural refinements.
In hindsight, it will probably also explain why Intel CPU's have progressed so little over the past five years. I am almost convinced that with the current CPU tasks, there is little to gain in terms of IPC or efficiency - something that is contrary to the thought of 'no competition = no perf gain' many people like to think.
ONLY when a new kind of CPU task comes to light that really tanks performance, will there be a new incentive to make CPU's a lot faster. Let's face it, the only real need for more performance is whatever has a good business case to it. If the vast majority doesn't need more perf, the market won't get it. Thát is why CPU's haven't progressed too much, they are capable at all the tasks we can throw at them.