Building a Reward System

Having a good reward system in place is critical to your dog’s progression in their training as a service dog. We want our dogs to absolutely thrive off of working with us. We want them to believe it is the greatest thing in the world. If your dog does not view the work as being rewarding you will struggle to get rock solid, reliable behaviors in the long run.

Most people who are not a fan of science based, positive, or reward based training typically don’t understand the concept of how to build a good reward system, or what "real life" reward are or how they work. They may sometimes call trainers who use these methods “cookie pushers.” But here’s the thing, while we may use a lot of treats in the beginning, no one wants to be a treat dispenser for life! And a dog that requires treats in order to do every little thing is not considered to be a fully trained dog in any training methodology. Our goal is to gain fluency, i.e. the dog will do what is asked, not because it is afraid of a punishment or other outcome if they don't, but because they genuinely want to, both because of what they get for doing it and because of the relationship that has been built between you.

Getting Started:

In the beginning stages of training, we may often hold a treat in our hand and show it to a dog to gain their interest or to help guide them into a desired position, however, it is critical that you get that treat out of your hand as soon as possible. I will typically try to do about 5 or 6 repetitions of something before I try encouraging the dog to do it without the treat. If the dog seems totally confused, I’ll do it a couple more times with a treat in my hand, then try again without and so on until the dog can respond to the cue without needing the treat in my hand.

Once a dog understands how to do the behavior we are teaching, we want to reward them for that behavior, not bribe them.

A reward is when you ask your dog to do a known behavior, they do it, and then a treat or other reward appears. There is no promise of reward when you ask your dog to do the behavior. A reward is simply a nice bonus that can come from doing what is asked. This will help your dog learn to eventually work regardless of if you have something to reward them with or not. You want your dog to be working primarily for rewards.
A bribe is when you show your dog a treat or other reward, ask your dog to do a known behavior, and then they do it. If you bribe your dog too much, they will never work without one. This is why it’s so important to get those treats out of your hand quickly! There is a time and a place for bribes. Sometimes our dogs can get so distracted or stimulated by something that we just can’t seem to get their attention back on us. Waving a treat in front of their face is an easy way to get them back to you and back into a working mindset. However, using bribes should be few and far between. If you find that you are having to use bribes too much, you likely waited too long to fade the treats out of your hand in the beginning or you are trying to push your dog too far too fast. In both cases, you will need to do some back tracking in your training and then work on building your dog up again.

Grading Scales:

We put dogs on a grading scale to help them improve their behavior. If your dog is getting the exact same reward, every time, regardless of how they do, your dog has no reason to improve! For example, if your dog gets 1 treat for a 10 second Stay and gets 1 treat for a 30 second Stay, then after 10 seconds, if your dog hasn’t gotten a treat, waiting longer simply isn’t going to be worth it to them and they are most likely going to break that Stay.
Here is an example of a basic grading scale:

Your dog performs better than normal. Jackpot! Give your dog 3-5 treats in a row or 30+seconds of play, affection, sniffy time, etc. If you are using treats, making sure you give the treats in a row is critical. 1 big treat and 1 little treat in a dog’s mind is still 1 treat. A handful of treats that your dog can just gobble up quickly will also seem like 1 treat. We really want our dogs to understand that they just got 1. . . 2. . . 3 treats! For a behavior. This will get the gears going in their head. Your dog is going to want to do the behavior that earns the better reward.
Your dog performs well. Give your dog 1 treat or 15-20 seconds of play, affection, sniffy time, etc.
You know your dog can do better. No reward, try again.

Next, you’re going to use Jean Donaldson's Rule of 4/3/2 to let you know when your dog is ready to adjust your grading scale. A first grader will get an A in school for knowing the answer to 2+2. That’s not going to cut it for a high school student! Using this system will help you know when your dog is ready to “go up a grade.” Here is a breakdown of how the Rule of 4/3/2 works;

If your dog succeeds 4/5 or 5/5 times- it’s time to push your dog and make what you are dong harder! Let’s use Stays as an example again. Let's say you’ve been working on getting a longer Stay. Your dog has been getting a Jackpot reward (several treats in a row) for being able to hold a 20 second Stay. Your dog just managed to hold a 20 second Stay 4 times in a row. Now when you do a 20 second Stay your dog is only going to get a regular reward (1 treat). Your dog has to hold it past that 20 second mark in order to get that Jackpot reward. If your dog doesn’t make it to 20 seconds, you know they can do better, so they get no treat and have to try again.
If your dog succeeds 3/5 times- stay where you’re at and keep building on your foundation. Your dog is starting to understand the concept, but they’re not quite solid in it yet.
If your dog succeeds 2/5 times or fails 3 times in a row- what you are doing is too hard! You need to figure out how to make it easier to help your dog succeed. If your dog keeps failing over and over they are going to get frustrated and will stop wanting to work with you. We always want to try and make sure that our dog’s success rate is much higher than our dog’s fail rate.

When applying your grading scale, always remember that every little change in the situation is going to make things harder on your dog. Just because your dog is really good at doing a Wait at home doesn’t mean that they are going to be able to do it at the store. Dog’s are not good generalizers and every little change in a situation or environment can throw them off. Most dogs take an average of 12 different scenario shifts before they will finally generalize something. If your dog is really struggling to do something that you KNOW they can do, take a step back, look at the situation, and figure out what is different that is causing your dog to struggle. Then, just go back a few steps to make what you are working on easier for your dog and build back up again.
Almost every time I find that someone is stuck somewhere in training or struggling to get their dog to improve, it is because they are not using their grading scale properly. Either they are not using Jackpot rewards when they should be, they are trying to push their dog too fast, or sometimes they aren’t pushing their dog when their dog is ready to progress and their dog has gotten bored with the training. So, any time you find yourself getting stuck, or feel like training just isn’t progressing as you’d like, go back and look over the grading scale and make sure there isn’t something you’re forgetting.

Moving on to Real Life:

Just by using the grading scale above, you will already be on track to no longer needing to be that constant treat dispenser. However, we want our dogs to be able to always do what we ask them, in any given situation, regardless of whether we have treats on us or not. Once our dogs have hit a level where we are happy with the behavior, we start applying an intermittent reward system or everyday life rewards.

An intermittent reward system is when you start varying how often your dog receives a reward once they are solid on a cue. Your dog may do a behavior once and get a treat, then five times and get a treat, then twelve times, then three, then seven, etc. It’s basically the same concept as a slot machine. People will keep pulling that lever over and over, even when they aren’t getting anything, because they know that eventually it will pay out. Using this system can be incredibly effective at creating extremely solid behaviors because it basically turns our dogs into little gambling addicts.
Using everyday "Real Life" rewards is the concept that anything your dog likes, wants, or needs can be used as a reward. For example, the reward for sitting calmly and politely is getting to greet the new friend that came to the house. If the dog can’t sit politely, they don’t get to say “hi.” The reward for not pulling on the leash is getting to go sniff that really cool bush over there. Etc. The obedience behaviors that we teach our dogs are not tricks, so our dogs should not need to be given a treat every single time they do them. My general rule is that a dog should always have to at least Sit before getting anything they like, want, or need. I view this as my dogs’ way of saying “please.” They do this because it is polite, good manners.

Different Types of Rewards:

Let’s dive a little more deeply into the different types of rewards and when is the best time to use each:

Food Rewards: The reason so many trainers like to use food rewards is because food has intrinsic value to it. You do not have to teach a dog that food is rewarding. A dog can’t survive if they don’t eat. Because of this, even the least food motivated dog in the world can become food motivated. If your dog is truly, 100% not food motivated, it’s time to take your dog to the vet, because your dog is going to die!
- Food rewards are useful when we want to promote a thinking brain. When we are working on more complex behaviors and need a lot of focus from our dogs we use treats. This is because treats can reward our dog almost instantaneously. This makes it so that we can do a lot of repetitions, all in succession, in a short amount of time, making the behavior stick in our dog’s heads much more quickly.
- Food rewards are great because they can be put into a grading scale all by themselves, for example, most of the time when you work , you might use basic treats or even your dogs kibble. For things that are super hard for your dog, you can pull out the big guns, like bits of chicken, cheese, hot dog, tuna fish, etc. The general rule is; the slimier and smellier the better! Suddenly, this makes the hard thing much more worth while to your dog. If this is the only chance they’re going to get this amazing treat, they’re going to want to take advantage of it. As your dog gets better you start only giving them that higher value treat when they perform better and the rest of the time they get their regular treats until eventually you’ve weaned them off of needing the high value treats all together.
- My favorite time to use food rewards is during my dog's meal times. Before my dogs are allowed to eat, I have them Sit (the foundation for good manners), Wait (the foundation for self control), and make eye contact with me (foundational to trust and relationship building). Once they've done this, their reward IS their meal! That gives me twice a day that I can help my dogs practice some key, foundational behaviors and earn a "real life" reward.
Play Rewards: These are used when we want to build motivation for a behavior. If you want your dog to think that something is the funnest, most amazing thing they can do, reward them with play.
- There are lots of different kinds of play rewards, these can be anything from throwing a ball or frisbee, playing tug, keep away, flirt poles, bubbles, etc.
- When it comes to play, you are typically going to reward with how long you play, rather than what you play. For example, a Jackpot could be 30+ seconds of tug or 2 or 3 throws of a ball. An average reward could be 15-20 seconds of tug or just 1 throw of a ball. Because it requires time to reward with play behaviors, it is not ideal to use when teaching a new behavior.

Social Rewards: This is anything involving people or other dogs.
- The most common social reward is teaching a dog to behave politely when greeting. We don’t need to use treats because the thing the dog wants is to say “hi.” So we teach our dog to use the behaviors we have taught them in order to get that reward.
- Other common social rewards can be getting permission to go play with another dog, getting their ears scratched in a certain way, getting their belly rubbed, getting a doggy massage, etc.
- Puppy Parties can also be really effective social rewards. This is when you get way, super, over the top excited. Clap your hands, jump up and down, act like you just won the lottery! If you act like your dog just did the most amazing thing in the world, your dog is going to think that they just did the most amazing thing in the world!

Environmental Rewards: This is anything in the environment around you that your dog may like or that catches their interest.
- Examples can be getting permission to go sniff something, getting to go play in some water, getting to go outside, getting to go for a ride in the car, getting permission to get up on a couch or bed to snuggle with you, getting to go investigate something that made a strange sound, and more!
- Environmental rewards often get overlooked. Rather than viewing them as rewards they are more often than not viewed as distractions. By just adjusting the way you think and view the world a little bit, these things can actually become very powerful rewards. Think about it: the whole reason that dogs tend to stop paying attention to us when we are out working is typically because they have found something in the environment that they find to be more interesting than we are! We can use this to our advantage, though. In short, we want to teach our dogs that we are the key to their environment. So long as your dog will check in and make eye contact with you, and not pull you to the interesting thing, we can get closer and possibly even go investigate the thing. If the dog starts pulling us or hits a point that they can no longer make eye contact with us, we get farther away from the thing. Eventually you will find your dog starting to actually let you know when there is something that has caught their interest, rather than just writing you out of the scenario and focusing all of their attention on the thing. Now, obviously we can’t actually let our dogs go and investigate every little thing that catches their attention, however, we can still acknowledge that there is an interesting thing and then use a different kind of reward to just keep our dog’s attention and focus on us.
Behavioral Rewards: These are natural behaviors your dog enjoys engaging in that you have put on cue. These kinds of rewards aren’t used as commonly as the other kinds of rewards, but can still be useful to know about.
- Some examples are teaching a dog to bark or howl on cue, or teaching a dog to jump on cue. My Great Dane’s favorite reward is when I give him permission to jump up and put his paws on my shoulders. These are behaviors that often times are “forbidden” unless your dog has been given permission to do so. These are things that your dog may naturally enjoy doing, but that can be very problematic in the human world. So giving your dog permission to do these things on special occasion can be hugely rewarding to your dog.