Depends. A single watt saved could suffice, if it is from a device that is on 24/7 and will be kept a long time. Say you find out that your intercom is using unnecessary idle power. And there's an easy fix for that. If that intercom were to be in use for 20 years (independ of who lives there), then that's 175kW of unnecessary power. It will most definitely not need 175kW to produce a Kill-A-Watt. As such, your purchase would have paid off even if you'd never use it again.
Reasonable point. Around here the power savings would have to be a little more to make it pay for itself (monetarily), but still as you say they don't have to be huge if you allow a long enough time.
It would be interesting to find out how much a "typical" or average user of a Kill-A-Watt meter saves in the medium to long term. (i.e., after the initial fever of testing things or turning things off.)
But I don't know mate, I think you're reaching a paranoid stage of greeness.
"Just because I'm paranoid doesn't mean they aren't out to get me"
Kill-A-Watts are mass-produced, as such, technically speaking, the more of them are produced, the smaller the CO2-footprint of a single individual unit, because the vast majority of energy is used by the factory.
The energy used by the factory is part of the embodied energy (along with that used in making the parts and packaging, shipping the parts and the final device, warehousing, marketing it, recycling, etc.). One may be able to improve efficiency, but you can't make it go to zero unless you stop making them.
Also, the impact of you personally not buying one is so irrelevant, that it wouldn't matter anyway. The manufacturer is not gonna "notice" that and will produce one less.
Of course that is illogical - by that reasoning, one shouldn't vote, and likewise by that reasoning beaches contain no sand. (if I take away one grain of sand, it is still a beach. Therefore I can repeat that operation, and still have a beach. There are only finitely many grains of sand on a beach. Therefore after removing all the sand I still have a beach.)
(Actually, I would just as soon anybody who is swayed by that reasoning doesn't vote - since that means that the people who understand the flaws in that argument will have more influince.
Nobody made a greater mistake than he who did nothing
because he could only do a little.
Consider also Superrationality - people may think like me and act like me, whether I communicate with them or not.http://en.wikipedia.org/wiki/Superrationality
And of course gathering information is an early step towards influencing.
Now, if you'd convince thousands of people to not buy Kill-A-Watts because of their potential CO2-footprint, it would have a bad impact because if just a handful of these people would save a few dozen watts after checking their consumption, the accumulated impact over years of saving will be much greater than that little bit of CO2-footprint a few thousand Kill-A-Watts have.
Not buying the meters does not preclude using them. If we persuade thousands of people to borrow one of the meters instead of buying it, then not only have they saved energy (assuming that use of a Kill-A-Watt results in energy savings for the average user - which seems a reasonable assumption, but worth checking), but we have also cut down on energy to create those meters, reduced clutter of gadgets in their homes, reduced the need to recycle the meters eventually, and the users will recoup their investment even faster because they don't have to recoup the full cost of the meter.
And, most importantly: people have to make a living. If too many people stop buying a certain product, the manufacturer will stop producing it. But he and the people he once employed still have to earn their livelihood, so they have to find other jobs. To be really sure, you would need to calculate the probability that the new jobs they're getting have less environmental impact than their jobs at the Kill-A-Watt factory.
Speaking of paranoia.