How to Teach a Dog to Stay (The Three Ds)

Stay is the command most people think their dog knows and most dogs do not, because it is trained the way it is imagined: dog sits, owner walks backwards across the room while saying "staaaay", dog gets up, owner repeats the word louder.

The fix is unglamorous. Stay is three separate skills, and they have to be trained one at a time.

The three Ds

Every stay problem traces back to raising more than one of these at once:

• Duration — how long your dog holds it
• Distance — how far away you are
• Distraction — what else is happening

Train one, keep the other two easy. When you raise one, drop the other two back. Walking three metres away and expecting the same thirty seconds your dog managed at arm's length is asking for a new skill and calling it disobedience.

Start with a release word

This is the piece most owners skip, and it quietly causes most of the failures.

If there is no signal for "the exercise is over", your dog has to work it out. And whatever they guess becomes the real rule — usually "when the owner looks away" or "after about four seconds".

Pick a word: "free", "okay", "break". Then use it every single time you end a stay, without exception. The stay does not end when you return; it ends when you say the word.

Duration first

Your dog is in a sit, you are directly in front, nothing else is happening.

1. Ask for the sit. Our guide to teaching a solid sit covers getting that fluent first — stay is built on top of it
2. Wait one second. Mark ("yes"), treat, release with your word
3. Wait two seconds. Repeat
4. Build to five, then ten, then twenty

Two rules that matter more than the counting:

• Reward while your dog is still in position, not after they get up. You are paying the holding, not the finishing
• Vary the duration. Three seconds, then eight, then two, then twelve. A steadily increasing count teaches your dog to anticipate the end

If your dog breaks, say nothing, reset, and make the next repetition easier. A broken stay is information about your criteria, not defiance.

Distance second

Only once your dog holds twenty or thirty seconds at arm's length.

Reset the duration to one or two seconds, then:

1. Take one step back. Step in, mark, treat, release
2. Repeat until boring, then two steps
3. Build gradually to three or four metres

Return to your dog to reward, rather than calling them to you. Otherwise "stay" starts to mean "wait until you get called", and the stay dissolves into a recall — which is a different cue with a different job, covered in teaching a reliable recall.

Watch your own body language: leaning forward, staring, or moving fast all pull a dog out of position. Stand up, move calmly, keep your shoulders relaxed.

Distraction last

Now, with duration and distance both dropped back to easy, add mild difficulty and build up:

Move up only when the current level works nine times out of ten. If it fails twice in a row, you are a level too high — go back rather than repeating.

Sit-stay or mat?

A sit-stay is physically demanding, so keep it to short, defined jobs: waiting at a kerb, holding position while you open the car boot, staying put while a visitor comes through the door.

For anything long — dinner, a coffee, a visitor staying an hour — use a mat or place cue instead. The dog must stay on the bed but may lie down, roll over and settle. It is far more sustainable, and it is the more useful skill for real life. It is also the practical answer to jumping on visitors: a dog on a mat is not at the door.

Practice that fits real life

Stay is a good candidate for tiny sessions rather than dedicated training time. Three repetitions while the kettle boils; a ten-second stay before the lead goes on; a wait at the kerb on every walk.

This distributed pattern is how the training schedule in Paw Plan is arranged — short daily sessions of five to ten minutes drawn from a course of around eighty lessons of four to seven minutes each, with rest days built in rather than squeezed out. The reason is not convenience. Spaced repetition consolidates better than a single long block, which is why a stay practised for two minutes on six days beats twelve minutes on one.

On the days when nothing goes to plan, do two minutes anyway. Showing up beats catching up.

A first session, start to finish

If you want something concrete to do today, this is five minutes:

1. Reps 1–5. Sit, wait one second, mark, treat, release. Stand still throughout
2. Reps 6–10. Vary between two and five seconds. Still directly in front
3. Reps 11–15. Add a weight shift or a half-step to the side, back to two seconds
4. Finish on a success, release, and stop while your dog still wants more

That last point is the one people ignore. Ending a session on a failure — "one more, and this time hold it properly" — is how you teach that the exercise gets harder until it becomes impossible. Stop early, always.

The three places it is actually worth having

Stay is easy to over-train as an abstract skill and never use. The versions that earn their keep:

• At the kerb. Sit and hold while traffic passes, released when you step off. Trained on every walk, free of charge
• At the door. Hold position while the front door opens, so a delivery does not become an escape. This one is genuinely a safety cue
• At the car boot. Wait until released rather than leaping out into a car park

Each of these is a specific context and needs its own practice — a dog with a perfect living-room stay may have no kerb stay at all, because the two look nothing alike from a dog's point of view.

Common faults, and what they mean

• Creeping forward — you are rewarding slightly late, or leaning in. Deliver the treat to their mouth in position
• Breaking as you turn your back — distance was raised too quickly. Practise turning away for one second at close range
• Holds indoors, fails outdoors — normal. Behaviours do not transfer between contexts automatically; retrain briefly in each new place
• Fine alone, breaks with visitors — that is a distraction level you have not trained yet, not a lapse
• Gets up when you reach for a treat — your hand going to the pocket has become the release signal. Vary when you reach

What stay is not for

Do not use a stay to hold a frightened dog in place. A dog who is worried about something needs distance from it, and a stay in that situation teaches only that you ignore their discomfort — the anxiety stays and the cue gets poisoned.

The same applies to a dog who is guarding, over-aroused, or reacting to another dog. Stay is a cue for a dog who is comfortable and thinking clearly. If yours is not, the signals in our dog body language guide will tell you, and the answer is management rather than a longer stay.

Timeline

A week or two of short daily practice gets most dogs to a reliable thirty-second stay at a few metres in a quiet room. Distraction work is the long part — realistically two to three months to something that holds at the front door.

The dogs with the best stays are not the most patient. They are the ones whose owners raised one D at a time and always said the release word.