"The problem with these sorts of statements is that nearly everything is insignificant compared to something else."
Yes. Yes it is. Welcome to reality, and engineering. The problem with your statement is that if you're burning a lot of time and mental energy on the insignificant things, then you are wasting your time.
If you, personally, have already taken the time to secure your housing's heating situation, already have an energy-efficient car and minimize driving anyhow, then maybe for you worrying about your kid's diapers is the lowest hanging fruit. I mean that completely straight, not sarcastic at all. However, if you're sitting here encouraging others to worry about their diapers while their homes sit there blowing 75% of their heat out a leaky door, then you're just wasting time.
It's the 80/20 principle, only when it comes to things like energy budgets it's more like the 95/5 principle; one or two things make up 95% of your energy budget and the rest is fiddling around the edges. Optimize the 95% first. Worry about the 5% later, if at all.
(Fiddling around making optimizations smaller than the noise threshold is also a waste of time. "Noise threshold" may sound like a weird concept to apply here, but it does; making a small food optimization that, even if applied by a million people, is still less than the daily variation in the amount of food thrown out by restaurants daily would be a complete waste of time. This is an example, I have no specific optimization in mind. The temptation is to say "Well, these things add up", but in fact when you're dealing with these sorts of distributions, sometimes they don't! Even ten .001% optimizations don't add up to much, and for the effort of getting those you probably missed out on a 1% somewhere else. These distributions can be very counter-intuitive, we don't do well with such large order-of-magnitude ranges.)