Treats and chews
How often should I use food rewards in training?
Last updated 23 September 2026 · How we write and check this
Short answer
For a brand new behavior, rewarding every single correct attempt for the first week or two builds the fastest, most reliable learning. Once a dog performs a cue correctly about 8 or 9 times out of 10, gradually reducing to rewarding only some correct responses, known as intermittent reinforcement, keeps the behavior strong while using noticeably fewer treats over time.
Why constant rewards matter early on
In the first stage of teaching any new cue, a dog has not yet connected the action to the outcome, so rewarding every correct repetition, even loosely correct ones, speeds up that connection. Skipping rewards too early in this stage, hoping to save on treats, usually slows learning and can cause the dog to lose interest in the exercise.
Shifting to intermittent rewards
Once a behavior is reliable, moving to rewarding roughly every second or third correct response, rather than every one, actually strengthens the behavior further in many dogs, since unpredictability tends to sustain motivation similarly to how it works with other reward based learning. This shift usually happens gradually over two to four weeks, not overnight.
Keeping some food rewards even for well known cues
Dropping food rewards to zero for a long mastered behavior like sit can cause the response to fade slowly over months, particularly in distracting environments, so occasional reinforcement, even just once every 10 or 20 repetitions, helps maintain the behavior without requiring treats at anywhere near the original training rate.
Key takeaways
- Reward every correct attempt for the first one to two weeks of teaching a new behavior.
- Shift to rewarding roughly every second or third correct response once accuracy reaches about 80 to 90 percent.
- This transition usually takes two to four weeks to complete gradually.
- Keep occasional food rewards even for well known cues to prevent slow fading over months.