If you haven't got a torque wrench, then it's all down to judgement calls - and there's no universally agreed method to calculate judgement. Judgement comes from experience, and how would I know that the experience of one person would apply to my situation?
Bike technology is generally quite forgiving, which can lead people wildly astray before causing any harm or becoming noticeable.

Case in point: over here, the favourite inner-city shop who does wheel building for MTBers who mainly ride in a nearby(reachable by bike) recreation area don't use tensiometers. Don't even have one. Won't take the job if asked to check/touch up a wheel to a certain range.
This works fine for the overwhelming majority of riders, as the rather technical tracks they ride usually means they'll smash their rims by impact before spokes start dying of fatigue anyhow.
Now, see what happens when a commuter rider gets his wheels worked on in that shop - failure upon failure upon failure. Shop is clueless - they think they build good wheels - rider/customer is annoyed. Shop concludes that the rider is doing "something special", and that they can't help him.

There's also the point of failure investigation. If I have something that fails after being tightened by feel, then installation error becomes an immediately possible reason. If something fails after being assembled with a torque wrench, there's one uncertainty removed or at least substantially reduced.
If I want to be really particular, I can go back and check the calibration of the torque wrench with basically a cheater bar and a known weight, or a luggage/fishing scale.

I'll use torque wrenches om bottom brackets, crank sets, post mount brake calipers and stems. Don't use them on levers, chain ring bolts and other minor assemblies.