Personally, I gauge by difficulties 15, 20, 25.
Because most checks add an ability and a proficiency, there is roughly about +5 to every check even at the lowest levels.
Thus 10 is normally trivially easy.
The difficulties of 15 and 20 are the respective challenging and hard of combat math. These two are my goto numbers.
I think DC 10 works as a better go-to number because:
1) The +5 only appears when a character who is GOOD at the skill is making the check. This isn't always true for skills. Sometimes a player is forced to make a check their PC is not optimized for (actually this happens a lot at my tables).
This is a major difference between skill math and combat math. Unless someone has deliberately gimped their character, a 1st-level PC should be making every attack with a +5, or maybe a +4 if you're not precisely optimized ("wrong" race, for example). There is no way that a 1st-level PC should be making every single skill check with a +5; it's much too constraining to ask players to only ever do things that jive with their very best skills.
2) DC 10 can still have tension and be interesting if the consequences for failure are interesting. If the consequence is a boring "didn't work, try something else," then sure, use the higher DC. But if there's a real downside for failure, even the PC with the good bonus is going to seek out advantage, even for DC 10.
Of course context matters:
I tend to use DC 15 as my go-to number in combat, often because the consequences of success and failure are often analogous to an attack, so it makes sense not to overpower skills by giving them a low DC.
The other time I use DC 15 as the go-to number is when the whole party gets to attempt something. When a bunch of PCs are rolling, the odds of success shoot way up. DC 15 makes it more likely that the PC who actually has the good modifier is the one who succeeds.