A thermometer is often described as a precision tool, which understates what it does. It replaces a specific judgement that most home cooks have too few repetitions to build.
Doneness is a temperature, not a time
Cooking times in recipes are estimates derived from a particular thickness, starting temperature, pan and heat source, none of which match another kitchen exactly.
What actually determines doneness is the temperature reached inside the food, since proteins denature and starches gelatinise at reasonably consistent points.
Measuring the thing that matters removes the chain of assumptions that a time in minutes quietly depends on, which is why timed instructions fail unpredictably.
Professionals rely on repetition instead
A cook who prepares the same cut hundreds of times learns to read it by touch, by how the surface behaves, and by the sound in the pan.
That judgement is genuine and fast, but it is built on volume, and it does not transfer well between cuts, sizes or equipment.
A home cook roasting a large joint a few times a year has no realistic path to that calibration, which is exactly the situation a measurement suits.
Carryover cooking is invisible without one
Food removed from heat continues to cook as heat moves from its hot exterior towards its cooler centre, and the internal temperature keeps climbing for several minutes.
The size of that rise depends on the mass and the surface temperature, so a large roast overshoots considerably more than a small piece of fish.
Without a reading there is no way to account for it, which is why food that looked correct at the moment of removal is frequently overdone by the time it is carved.
Placement decides whether the number means anything
A probe resting against bone reads the bone temperature, which runs hotter, and a probe pushed too far through reads the pan on the other side.
The useful reading is the coldest point, usually the geometric centre of the thickest part, which requires moving the probe rather than trusting the first number.
Instant-read instruments settle within seconds; leave-in probes report continuously but need positioning before cooking starts, when the food is easiest to misjudge.
Where it changes nothing
Stir-frying, sautéing and most vegetable cooking are governed by surface behaviour rather than internal temperature, and a probe adds nothing.
Bread and pastry benefit occasionally, mainly for enriched loaves where colour is a poor guide to whether the crumb has set.
The tool earns its place on large roasts, whole birds, custards and confectionery, where the margin is narrow and the cost of being wrong is an entire dish.