I'd agree it's not good taken in isolation as the ONLY criteria, especially as "usage" is really pretty subjective from person to person (my "light" can be somebody else's "heavy" - plus what I've got installed/disabled will probably differ greatly from another device).

But taken in context, like with me and the comparison that I mentioned before versus similar prior usage patterns on a previous device, it MIGHT actually work rather well as ONE OF the indicators of battery life.

In most cases anyway.

Especially when its so easy to just look up SOT and compare - though again with the sheer subjective-ness of individual use cases, it's really only good as a simple point of reference.

So, AS LONG AS you're aware of the shortcomings of basing overall battery life estimations on SOT, I'd say its fine for simple comparative purposes...

... BUT, if you're taking it as the "be all end all" measure of battery life... eyyyyy, I'd say you're correct in saying that it's shouldn't be treated as such - considering just how unreliable it is from a consistency between setups/use cases point of view.

But TBH, while we're on this subject, it's actually pretty hard to cross measure battery life performance in smartphones "in-actual-use" quantifiably considering how varied the use cases are for everyone.

Like, even if you resort to benchmarks like Futuremark or something, unless you're running a clean bone stock device, what's installed (and thus is running in the BG) on your device isn't gonna be the same as another either.