I use depreciation in calculating car costs, but that's because I set a certain time frame where I am relatively sure a given situation won't change (I usually use 36 months), for instance, commuting distance, family size. The assumption is at the beginning of the period I acquire the assets and at the end, I dispose of them. It's very much like a corporation but it makes sense to me.
I missed out on a Cruze Eco that was priced about 4k under market for no apparent reason a couple weeks ago. Plugging the belief that it would be at least worth what I paid for it, even with miles added on, in 36 months, depreciation was effectively zero. If it had been priced at market, depreciation would have been about 4k. Of course as you said that is not on top of purchase price but depreciation is a factor when you aren't keeping an asset forever which applies in my case. Anyway, based on the zero depreciation, lower maintenance costs, and better mileage, it made more sense to buy that newer "fancy" Cruze than to keep my paid for, fully depreciated Focus going. The wife still doesn't get my maths on this.
But yes depreciation + payments assuming total payoff is wrong.