Hello there,
I think most of the issues have already been addressed fairly well, so i'll just add a few comments to the bunch here.
On the subject of converter (any type) vs LDO, dont let the LDO terminology fool you as it eats plenty of power unless the input voltage is close to the output voltage, and then it can possibly eat less than many buck circuits. What this means is that the choice of LDO vs Buck would depend on the average input and output levels as well as the efficiency of the buck. A very high efficienct buck would be hard to beat for example, but a more typical one would be beaten by the LDO if the level of input voltage does not go much above the output voltage. It's something that you have to look at when you decide what buck circuit to use. Notice also that im talking about a buck circuit for comparison here, that's because you probably dont want to use a boost circuit as they usually eat more power than a buck and you dont really have to resort to boosting a single cell up to 5v anyway. So the decision to go buck or LDO is going to be mostly based on input vs output voltage over most of the run time.
There is another factor which you mentioned however and that is that you dont seem to want to go into using a large number of batteries, even if the cost is lower than one big battery. That seems to immediately rule out using AA cells as you would need too many of them to get the full 1 year run time. In this case, you'll probably want to go into a larger battery with a buck circuit of reasonable efficiency, perhaps a synchronous buck. That would take you from 12v input to 5v output with reduced current from the 12v cell, and as you may already know when you use a large battery for a small current (relative to it's normal operating current) you end up with a practical increase in ampere hour capacity. That means if you buy a 15 Ahr battery you may be able to get 18Ahr out of it simply because you're running at a very low current (like 2ma or less). That means a single lead acid battery with a decent buck circuit might just do it.
Now the questions for you include...
1. Just how important is that one year spec, what would happen if it died in 9 months. Would someone die or would you just have to replace the battery?
2. Related to that question, do you have a year or two to test this or does it have to be a one time deal where you can not afford to say double the battery size if it didnt go the full year on the first try?
3. You have any sort of budget?
4. Is there any way of charging say at the 6 month period or something like that if you use two batteries instead of one?
5. What is the working environment like, as to operating temperature? This would affect the self discharge of any battery you consider for use and thus might demand a higher capacity battery.
6. Any chance you can use li-ion ?