Skip to content
Animiyo
Get started
Behaviour and training7 min readUpdated on August 5, 2026

Positive reinforcement: using treats and the marker with the right timing

How positive reinforcement really works, why marker timing matters more than the treat, and how to move from food to a behaviour that lasts without it.

Audience
Pet owners
Species
Dog, Cat
Scope
Valid everywhere

Almost everyone knows that rewarding works better than punishing, but few use positive reinforcement effectively, and the reason is nearly always the same: timing. A treat given half a second too late does not reinforce the behaviour we wanted, but whatever the animal was doing in that half second, which is often standing up, turning or looking away. Dogs and cats do not understand our intentions: they read the coincidence between what they were doing and what happened straight afterwards. That is why modern training uses a signal, the marker, that tells the animal precisely which instant will earn the reward. This article explains how positive reinforcement works, how to build and use a marker, which mistakes make it ineffective, and how to reach behaviours that hold even when food is absent. For serious behavioural problems, such as aggression or marked anxiety, the reference remains a behaviour professional.

What a treat actually reinforces

A behaviour followed by something pleasant tends to repeat. That is the principle of positive reinforcement, and it holds for food as for a stroke, a game or access to something the animal wants. The key word, though, is followed: what gets reinforced is not the intention but the behaviour present in the instant before the reward. If the dog sits and gets the treat three seconds later, when it has already stood up again, we have reinforced standing, not sitting.

This explains most of the failures of people who use treats in good faith. It is not the number of treats that teaches, but the precision with which they are linked to the right moment. And since it is almost impossible to deliver a treat at the exact instant, an intermediate signal is used that marks the moment and promises the reward straight after.

The marker: building the signal that promises the reward

The marker is a short, always identical signal, a word like yes said sharply or the sound of a clicker, that means only one thing: what you are doing now earns the reward. For it to work, the animal must first learn that the signal always announces something good. This initial phase, in which the marker is charged, does not require the animal to do anything in particular: it only builds the association.

  1. Charge the marker in a quiet setting

    Give the signal or use the clicker and immediately follow with a small reward. Repeat many times until the signal makes the animal turn to you expecting food.

  2. Mark easy behaviours the animal offers on its own

    Wait for it to sit or look at you spontaneously, mark the exact instant and reward. Ask for nothing yet: you are reinforcing what happens.

  3. Always mark one precise instant only

    The signal coincides with the behaviour, the reward comes after. If you mark late, you teach the wrong action.

  4. Add the cue only when the behaviour is reliable

    The word that asks for the action comes later, when the animal offers the behaviour consistently, not at the start.

  5. Always reward after marking, even if you were wrong

    If you gave the signal, the reward must follow: the marker loses value if it sometimes fails to keep the promise.

  6. Keep sessions short and frequent

    A few repetitions at a time, several times a day, are worth more than one long session that tires and frustrates the animal.

The marker works with cats as well as dogs, and in many cases it is even more useful with cats, because it makes clear a communication that otherwise stays vague. The rewards and attention spans change, not the principle.

The mistakes that switch reinforcement off

Positive reinforcement rarely fails because it is wrong in itself; it fails through execution errors that cancel its effect. Recognising them in advance avoids months of unproductive work.

  • Marking or rewarding too late, thereby reinforcing the next movement instead of the intended behaviour.
  • Repeating the cue several times before rewarding, teaching that it need only be obeyed on the third or fourth ask.
  • Using treats that are too big, which fill the animal quickly and cut useful session time.
  • Showing food first as a lure, creating an animal that responds only when it sees the reward in hand.
  • Rewards of always equal value, when a difficult task calls for something tastier.
  • Long sessions that pass the attention threshold, after which the animal errs more and learns less.
  • Not counting treats in the daily ration, ending with an animal that learns well but gains weight.

From a treat every time to a behaviour that lasts

Rewarding every single time is right when teaching something new, but it is not the goal: an animal that obeys only when it knows a treat is coming has not learned solidly. The turning point comes when you move to an intermittent, unpredictable reinforcement, and when you widen the range of rewards beyond food.

  1. In the learning phase reward every correct behaviour, to build the association quickly.
  2. When the behaviour is reliable, move to rewarding unpredictably, not on every repetition.
  3. Make the reward variable: sometimes the treat, sometimes a game, sometimes access to what the animal wants at that moment.
  4. Use real life as a reward: opening the door, throwing the ball, allowing a sniff are powerful rewards.
  5. Keep an occasional surprise reward even on established behaviours, to keep them strong over time.
  6. Keep marking difficult behaviours or ones done in new situations, where distraction is greater.

A reinforcement that becomes unpredictable is stronger, not weaker: it is the same mechanism that makes it hard to stop checking something that sometimes, not always, gives a reward. The final goal is not an animal without treats, but an animal for which cooperating with you is worthwhile in itself, with food as one tool among many.

Frequently asked questions

Do I have to use a clicker?
No. The clicker is handy because it makes an always identical, very crisp sound, but a short word said always the same way works just as well, and it is always with you. What matters is not the tool but consistency and timing: the signal must coincide with the instant of the behaviour and must always be followed by the reward. Many people start with the clicker to train their own timing and then switch to a word.
If I use treats, will the dog obey only when hungry or when it sees food?
That happens when food is used as a lure, shown first to persuade, rather than as a surprise consequence after the behaviour. If the reward comes after, and if over time you move to unpredictable reinforcement and widen rewards beyond food, the animal learns to cooperate regardless of a visible treat. The key is never to show the reward first and not to reward every single time once the behaviour is solid.
Does positive reinforcement work with cats too?
Yes, and often better than expected. The cat responds well to a marker followed by a very tasty reward, provided sessions are very short and in a quiet setting. With cats it is even more important to respect the short attention span and to choose the right moment, when the animal is willing and not annoyed. The principle is identical to the dog's: mark the exact instant and reward straight after.
Is positive reinforcement enough even for problems like aggression?
For everyday behaviours positive reinforcement is the main tool, but faced with aggression, marked fear or significant anxiety it is not something to handle alone with a few treats. These cases call for assessment by a behaviour professional, often together with the vet to rule out a medical component, such as pain, that can lie behind many sudden behaviours. Positive reinforcement remains part of the plan, but within a guided path.

What to do next

Charge a marker in a quiet setting, then mark the exact instant of the behaviour and reward straight after, never showing food first as a lure. Keep sessions short, add the cue only when the behaviour is reliable, and gradually move to unpredictable reinforcement while widening rewards beyond food. Count treats in the ration and, for aggression or marked anxiety, turn to a behaviour professional.

Related content

Positive reinforcement: using treats and the marker with the right timing · Animiyo