The most reliable way to tell whether bread is done is to check its internal temperature. For plain, lean, unsweetened dough, the target is 93–99°C / 200–210°F; for soft doughs containing milk, the target is 88–90°C / 190–195°F. Insert the thermometer from the side into the thickest part of the bread so it reaches the center. Color and tapping the bottom to listen for a sound do not replace this measurement; both can be misleading.
There are two targets, selected according to the dough’s ingredients. For plain bread—dough made only with flour, water, salt, and yeast—the target is 93–99°C / 200–210°F; at this temperature, the starch is fully cooked and the interior structure sets. For soft doughs containing milk, eggs, fat, or sugar, the target is lower, in the 88–90°C / 190–195°F range. These doughs dry out at higher temperatures and lose the softness you want. The 5°C / 9°F difference may seem small, but it marks the line between dry and soft in the same dough.
Insert the thermometer into the thickest part of the bread, with its tip reaching the center. Inserting it from the side is preferred because it does not leave a lasting hole in the top crust. Do not take the reading in just one spot; check two or three places and use the lowest reading. An instant-read thermometer is more than adequate for this. For bread baked in a pan, do not let the thermometer touch the pan, because the metal is hotter than the bread and can raise the reading. Also check the thermometer from time to time in ice water to see whether it reads zero; a thermometer that is out of calibration can give false confidence.
Color is a common but unreliable measure. The crust’s color changes with the amount of sugar and milk in the dough, the oven’s upper heating element, and the pan’s position; dough with milk may darken while the inside is still undercooked. The reverse can happen, too: plain dough may remain pale even when fully baked. Likewise, tapping the bottom and listening is only a rough method; a hollow sound indicates doneness but does not show how close the bread is to the target. Time alone is insufficient for the same reason: the same recipe finishes in different numbers of minutes in different ovens.
Take the reading just before removing the bread from the oven. You can take the bread out when it is a few °C below the target because its temperature continues to rise somewhat outside the oven. Bread removed below the target will remain doughy inside, and this flaw cannot be corrected later; even if the bread goes back into the oven, only the crust will thicken. For that reason, when in doubt, baking it for a few more minutes is always safer than taking it out too early. If returning it to the oven, cover the bread with foil so the inside can continue baking without thickening the crust further.
If you do not have a thermometer, use a few indirect signs. The bread should feel noticeably lighter in your hand because it loses some of its water as it bakes. The crust should feel firm and dry. The bottom should sound hollow when tapped. It should also come out of the pan easily. None of these signs is conclusive on its own; assess all three together, though they still cannot provide the certainty of a thermometer. To measure the weight loss, you can weigh the dough before putting it in the oven; baked bread is roughly one-tenth lighter.