You pay for a select potential efficiency when you buy Cree (CXA, CXB) or Bridgelux (Vero) COBS.
What I mean by potential efficiency is the possible efficacy's given by the product and because Cree sports the highest potential, they also get to charge the highest dollar for their products. Bridgelux doesn't focus on efficiency as being it's selling point and rather found it lucrative to establish their own niche, which is a packaged deal of affordability, simplicity and with potential efficiency slowly rising.
The catch that people sometimes forget when running COBS low is that they are ultimately reducing the total watts being ejected from the COB, thus requiring more COBS to be positioned in a given area. So on one side, by lowering current, you get less heat and a better mixture of PAR W (due to higher efficiency, lm/W), leading to cheaper electricity bills, but that money you save goes right back into buying more COBS in order to cross the required PAR/W finish line.
In short, there are two extreme for a given COB: running it as low as possible but getting a great PAR W % but emitting little of anything at all or running it as high as possible, getting more heat than anything but still producing a bunch of PAR W.
The best thing a person can do is find a peaceful middle ground for the time being and in a couple of years, splurge on newer versions and not get heavily invested in current gen technology, because it hasn't hit it's peak.