It occurs to me that many contractors like NyteTime are trying to evaluate LED's in the field to see if they hold up.
I think it would be very helpful to develop some simple field-testing methods for LED's so a contractor could quickly check the fixtures during a yearly maintenance.
To test luminance: An accurate lightmeter (+/- 1%) would be essential. At the minimum it would test both footcandles and color - those meters are pricey. Simple footcandle meters are fairly inexpensive.
Next, you'd need to create a black box of some sort that could be positioned with the fixture in one end and an opening for the meter at the other end. There would need to be some way to exactly duplicate the position of the meter every time. You may also want to have a few holes for different meter positions since light distribution may change in time.
To test color: Either a color meter could be used or you could buy an extra fixture at the time of the project. The extra fixture would be kept as a never-used control. During testing (at night) you could illuminate a white board with both old and new fixtures side-by-side. Color differences would be readily seen this way.
There are some confounding factors to consider such as the presence of condensation, the ambient temperature and others.
I think this testing is very important since we are looking for relatively minor changes each year. Most mfg. now use an L70 standard for life - that means the LED life is predicted as the time when 70% luminance is reached. An LED operated for 6 hrs. per day reaches a 50,000 life in about 23 years. If the LED is living up to it's promise then we'd expect a lumen depreciation of about 1.5% per year. This is far below our ability to perceive difference in brightness. Even a side by side comparison could have a 5% difference and it would be hard to detect. An LED operating 6 hrs/day that loses 5% per year will have an actual life closer to 13,000 hrs.
Let's not let mfgs. get away with inflated claims. Let's accurately test the fixtures in the field.