Start with a sit your dog already knows
Stay is easier to teach on top of a solid sit. If your dog does not sit reliably yet, train that first, then come back. A stay built on a shaky sit will always be shaky.
Pick a release word and use it every time: "okay" or "free" are common. The release word matters as much as the stay itself, because it tells the dog when the job is done. Without it, dogs decide for themselves, usually at the worst moment.
The 3 D's, one at a time
Duration: Ask for a sit, say "stay," count one second, then release and reward. Build to 5 seconds, 10, 30, a minute, over sessions. If the dog breaks, you raised the bar too fast. Smile, reset, and make it easier.
Distance: Once your dog holds 30 seconds with you close, take one step back, then return and reward. Build to across the room. Always return to the dog to reward; do not call them out of the stay, or you punish the very behavior you want.
Distraction: Last and hardest. Add mild distractions one at a time: you clap, a toy rolls by, someone walks past. Each new distraction drops the difficulty back to easy duration and distance, then rebuild.
The golden rule: only ever raise one D at a time. New distance means short duration and no distractions. New distraction means close distance and short duration.
A two-week practice plan
Week one: duration only, building to 30 seconds in quiet rooms. Week two: add distance in the same rooms, then take the stay to new rooms (each new location restarts at easy). Week three and beyond: layer in distractions one at a time. Five minutes a day is plenty. Two short sessions beat one long one.
Common stay mistakes
Repeating "stay stay stay" teaches the dog that the first three do not count. Say it once. Reaching for the dog when they break teaches them that breaking gets attention. Just reset cheerfully. And never punish a broken stay: the dog was not being defiant, the criteria were too hard. Lower them and try again.