Teaching a dog to be alone

First, which problem is this?

Three different things look similar from the doorway and need completely different plans. Getting this wrong is the most expensive mistake on the page, because the training below is the right answer to only one of them.

• The dog is bored. Damage is spread around the house, arrives later in the absence, targets bins and furniture, and the dog eats and sleeps normally while you are out. This is an occupation problem. Start with enrichment and play, not with departure training.
• The dog struggles with the absence itself. Signs start within the first few minutes, damage concentrates at doors and windows, vocalising comes in bouts, and a food item the dog normally loves goes untouched. This is what the rest of this page is about.
• The dog panics. Self-injury, breaking through doors or windows, sustained distress that does not settle, or signs that appeared suddenly in a dog who was previously fine. This is not a training project. It is a veterinary appointment, and the section on when to stop and get help is the only part of this page that applies.

You cannot tell these apart by what the room looks like when you get home. Film one full absence before deciding anything. The full diagnostic version of this fork, including why the guilty greeting at the door is not evidence of anything, is on your dog left alone.

Find the real starting point

Almost everyone starts too high. The number you need is not how long you have to be out for work. It is how long this dog is currently fine, which is often much shorter than the answer people expect and is occasionally measured in seconds.

Set a camera where the dog settles and where the door is visible, then leave normally and stay away for longer than you think you need. Watch it back and mark the moment the dog stops coping: the first sustained pacing, the first bout of vocalising, the point where they get up and do not settle again. That timestamp is your starting point, and you work below it, not at it.

How the training works, mechanically

The method is systematic desensitisation, which is a formal way of saying that the dog meets the thing they find difficult at an intensity that does not frighten them, often enough that it stops predicting anything bad. The whole craft is in keeping every single repetition below the line you found above.

• The criterion is the dog's behaviour, not the clock. A ninety-second absence the dog slept through is a success. A forty-second absence that ended with pacing was too long, whatever the plan said. You are training a state, and the timer is only a way of describing it afterwards.
• End it before it goes wrong, every time. Come back while the dog is still settled. The repetitions that build tolerance are the ones the dog finished calmly; a repetition that ends in distress has taught the opposite of what you wanted, and it costs more than the successful one gained.
• Vary the duration instead of climbing it. A staircase of steadily longer absences teaches the dog that every departure is worse than the last, which is a reason to worry from the first second. Mix short and slightly longer within a session, so a departure stops predicting a duration at all.
• Change one thing at a time. Duration, which door, whether your coat is on, whether the car starts, time of day. Move one and hold the rest still, or you will not know which variable caused a bad rep.
• Decouple the departure routine. For many dogs the fear starts with the keys and the shoes, well before the door. Pick up the keys and sit back down. Put your coat on and make tea. Do it until the routine stops announcing anything, and only then attach it to a real absence.

A session is short. Ten to fifteen minutes of this, a few times a week, moves faster than one long grim afternoon of it, because the dog is only learning while they are under threshold and there is a limit to how long anyone stays there.

What a fortnight actually looks like

Slower than you want, and not in a straight line. A realistic pattern is a few days spent below a minute, a stretch where progress feels quick, and then a plateau or a slide backwards after an unrelated bad week. Backward steps are normal and are not evidence the plan has failed; they usually mean the last few repetitions crept over the line.

Keep a note of each session: the duration, whether the dog settled, and anything unusual about the day. It takes seconds and it is the only way to notice a plateau that is really a pattern, such as every failure happening on the days the bins are collected.

If six weeks of consistent work has produced nothing, stop adjusting the plan yourself and read the next section. Persisting with training that is not working is the most common way owners lose months.

The training mistakes specifically

• Testing rather than training. Leaving for an hour to see whether the dog can manage it yet is not a training repetition, it is an exam, and a failed one sets you back further than the pass would have moved you forward.
• Letting it run because the dog will settle eventually. Absences longer than the dog can cope with tend to sensitise rather than habituate, so the fear is practised daily instead of fading. This is the single most common way a manageable problem becomes an entrenched one.
• Punishing what you come home to. It does nothing to the panic, which happened hours earlier, and it adds a second thing to dread about your return. Aversive methods are associated with worse welfare and behavioural outcomes without better results, and every relevant professional body advises against them.
• Assuming a crate is the answer. Some dogs settle better in one and some injure themselves against it. Never introduce confinement blind, and never leave it untested on camera.
• Getting a second dog as treatment. Many affected dogs are attached to a person rather than to company, and still panic with another dog in the room.

Cover, while you are training

There is an obvious contradiction in this work: you are teaching a dog to tolerate short absences while your job requires long ones, and every long absence undoes some of it. The way through is to arrange cover so the dog is not left beyond what they can currently manage, and to treat that as part of the plan rather than an admission of defeat.

• Free. A neighbour, a friend, a family member, or someone in the household staggering their hours. Ask specifically for company rather than a walk, which is the part that matters here.
• Cheap. Splitting the day, so the dog is alone twice for short periods rather than once for a long one, is often the difference between under and over threshold, and it costs nothing but logistics.
• Paid. A sitter, a dog walker or daycare. What you are buying is a shorter maximum absence, so book it around the longest gap rather than at whatever time is convenient.

The paid options are a real line in a household budget and are usually underestimated. They are itemised, with published figures, on what a dog costs.

When this is not a training project

The first call is your veterinarian, not a trainer. Physical discomfort lowers tolerance for everything, pain is suspected in a substantial share of dogs referred for behaviour problems, and a new difficulty with being left in a middle-aged or older dog is a medical question until somebody has ruled it out. Sudden change in particular is a reason to be examined before any plan is written.

After that, this is behaviourist work rather than something to solve from a web page. Severe separation-related distress is a recognised clinical presentation, medication is sometimes a legitimate part of the plan rather than an admission of failure, and any prescription comes from a vet who has examined your dog. How to find someone properly qualified, and what to ask before booking, is set out on choosing a professional.

Breed is close to useless as a guide here in either direction. A reputation for clinginess neither confirms the problem nor excuses it, and the largest genomic study of dog behaviour found breed to be a weak predictor of how an individual dog behaves.

Common questions

The answers below draw on the same sources cited in the sections above.

How long can a dog be left alone?

There is no species-wide number, and anyone quoting you one confidently is guessing. It depends on the individual dog, their age, whether they can relieve themselves, and what they have learned about departures so far. The only number that matters for training is the one you measured from your own dog on camera. For an adult dog with no difficulty being left, the practical limits are usually about toileting and boredom rather than distress.

Should I tire my dog out before I leave?

A sniffing walk or a foraging game before a departure is worth doing, because a dog who has had something to do is more likely to settle. Be clear about what it is: it makes the absence easier to sit through, and it does not treat the underlying fear. Hard physical exercise immediately before leaving can leave a dog wound up rather than settled, so favour the sniffing and searching kind. The options are on enrichment and play.

My dog is fine as long as another person is home. Is that separation anxiety?

It may be attachment to one particular person rather than distress at being on their own, which is a recognised pattern and matters because the plan is different. Test it deliberately: film an absence where the usual person leaves and someone else stays, and compare it against an absence where the house is empty. If the dog is settled with anyone at all, you are training tolerance of one person leaving, which is often easier.

Does leaving the radio or television on help?

It may make a house feel less abruptly empty and it masks outside noise, which is genuinely useful for a dog who reacts to sounds in the street. Nobody should expect it to resolve separation distress on its own. Try it as one variable, held constant, while you do the actual work, and watch the recording to see whether it changed anything rather than assuming.

The full entry on separation distress → Is it distress or boredom?

Related: howling, whining, displacement behaviour, and trigger stacking.

Practical guides: enrichment and play, reward-based training, feeding and routine and what a dog costs.

Behavioural claims on this page are graded against the source register. Where the practical mechanics are professional convention rather than trial evidence, the page says so in place rather than borrowing the authority of the citations around it. Separation-related behaviour is a description rather than a single diagnosis, most household evidence is owner-reported and briefly filmed, and the differences between individual dogs are large. Think we got this wrong?