Apologies for the hijack, but it could be closely related.
I work in a somewhat sizable commercial building with highly variable levels of use during the week. We have a bunch of fixtures very similar to the ones you linked to - some are virtually identical cans and some are a completely cylindrical can with no "taper" at the socket end. These are in the new part of the building, which has been up and running for eight months, and they are on for no more than 20 hours a week.
A quick mental count gives me a total of 60 or 64 fixtures. We use the 39PAR20 halogen lamps, and in eight months we've lost well over a dozen. 8 months x 4.2 weeks in a month x 20 hours a week = 672 hours. The originals are Philips, rated 1100 hours, and we're replacing with Satco. Haven't lost any Satcos yet, but this seems like a pretty short lifespan on the originals.
Would you agree, or do you think this is within the range of reasonableness? I'm very interested in what happens with the OP, as I'm kind of wondering if we have a heat issue as well. Maybe I should start pricing LEDs?
***edited because added an extra word word.
1100 is the average rated life, what that means is: "the length of time (in hours) required for half of the lamps in a large test group to fail. So you can have a group of 1000 lamps and 500 of them will have failed at (or before) the rated life", some will continue to operate well past the rated life.
A logical question would be what causes this wide swing?
Halogen lamps have a tungsten filament similar in structure to an incandescent type enclosed in a small glass envelope, the smaller amount of airspace and halogen gas at 7-8ATM cause the filament to reach a much higher temperature than incandescent types, thus giving a brighter, whiter light - the 39watt PAR Lamp is a redesigned, more efficient variation that has come into use in 2013 as part of a Federal Mandate. The halogen gas (iodine or bromine) and the tungsten filament when in operation for at least 20 minutes start what is called the "halogen cycle" - this is when small particles of the tungsten will come off of the filament and become airborne within the glass envelope and eventually reattach to the filament (tho not at the same place they left from) this typically helps the lamp last much longer than incandescent types which have a similar characteristic except that their filament particles do not reattach. The halogen cycle is random so the filament structure my weaken continually at one location and cause a premature failure - This usually happens when the lamp is first turned on and receives the full 120 volts as a jolt to the filament.
Generally I've heard Phillips makes a good lamp (we use Sylvania and Satco ) Satco cost less and often don't have equivalent output to the better lamps.
Your application would be best served with some good quality LEDs, they will amortize quickly (maintenance $$$ and a wattage reduction) in your building. When it comes to selecting what's right, remember that you get what you pay for, good LEDS should equal or surpass halogens in lumen output, offer some beam spread choices, have a 50K rated life, be dimmable, offer some choices on color temperature 2700-3500K, 3000k is considered "neutral", and a Color Rendering Index number of 85 or above. I like Sylvania, and Soraa LEDs as they offer all of these things, an $8.00 special from Walmart will not - it's best to think of the LED market as a lot of Snake Oil, buyer beware! Depending on the housing (Can) and it's physical dimensions it's often advisable to use a "short neck" halogen or LED so it regresses and becomes less visible - works great with a Specular or Alzak trim piece to get rid of the glare normally associated with recess cans.
