John, The Puppy Guy was not always The Puppy Guy. In fact, long ago, in a galaxy far, far away, he spent an interminable amount of time on the corporate customer service desk for a large investment house. And not surprisingly, they wanted their phone jockeys to behave in very specific ways. John mostly just wanted a job in his field, and the paycheque. When the company and the employee have different goals, hilarity can ensue.
|Working on the telephone desk of a big corporation is a complex task that could be broken down into component elements, but often isn’t! Image credit: auremar / 123RF Stock Photo|
Let’s call the company Big Bank and Investing. BB & I hired a consulting team to come in and figure out what criteria they wanted the people answering the phones to use in order to keep their customers happy. They narrowed it down to a few relevant things. First was that the customer service rep was supposed to address the client by name four times in each call. Second, they were to ensure that each client was asked if they wanted their account information updated. Thirdly, they wanted the phones answered within the first two rings, so that the clients never had to wait to talk to someone. Finally, they wanted their reps to thank the customer for calling BB & I. The consulting company chose measurable, repeatable criteria, which is always a good idea when you want to change behaviours. They forgot a few minor details however.
The first of the details they forgot was that the clients were calling a discount brokerage desk. This meant that individuals were sitting by their computers waiting for the markets to move and then hitting the speed dial button to talk to John or any of a hundred other guys just like him. The phone calls would go something like this:
Hello, this is BB & I, and you are speaking with John. Who is calling please?
This is MikeBRaddock24352345234-L, I need to sell 700 of GoGolfing at 42.
Thank you Mike. Before we continue Mike, I see that you only have 600 of GoGolfing and GoGolfing is selling for 38, Mike. Mike, shall I put that trade through for you at 38, Mike?
Yes go ahead.
Now, Mike, is there anything I can do for you?
Before you go Mike, can I check your account details and make certain that we have your correct contact information?
No. Thanks. Bye.
Thank you for calling BB & I Mike.
Except that Mike would have hung up on you before you said good bye. And if you didn’t finish saying Thank you for calling BB & I to dead air, then you got a poor rating on that particular call. This was really depressing for the customer service reps, and in fact led to a suppression of the desired thanking behaviour. Not only that, Mike was annoyed because he talked to you fourteen times a day, every day and he was sick and tired of being asked to update his account. Some of the stories John told were quite funny about the responses they got to their queries about contact information. About the fourth time you are asked the same question in a given day, you start getting annoyed and you start giving smart ass answers, such as “Yes, please, update that I am in the den now instead of the office.” Or sometimes they would string along a rep in order to not change their address but to keep an agent on the line so they could make another trade without having to call back. And sometimes they would keep you on the line for a long time.
Not surprisingly, moral was not great amongst the employees, so the office management team was constantly challenged to come up with incentive activities. One such activity went like this. Each employee had four calls randomly selected each week to be evaluated by the management team. If you hit all the criteria that were laid out, you got the chance to kick a soccer ball into a net from a set distance. If you got the ball in the net (keeping in mind that everyone is in business attire), you got your name put in a box for a monthly draw to get a day off work. One day, John got all the criteria right, on all four calls and his manager proudly walked over to his desk, with the soccer ball under his arm. With great ceremony, he offered the ball to John.
“No thanks,” John replied, “I just want to answer my phone, and go home at the end of the day.” The manager was dismayed. “You earned this”, he said. “I don’t want it”, John replied. “You are not being a team player”, the manager replied. “I really don’t want to do this”, John came back. “I will write you up in your permanent file if you don’t kick the ball”, the manager pressed. So John stood up and took the ball and removed his suit jacket. He carefully placed the ball on the floor. He took two steps back. And he HOOFED the ball into the goal. And his shoe….flew up and lodged in a ceiling tile.
Two days later a VP of the company came and toured the department and for some unknown reason, the soccer game disappeared forever. I have always thought that the shoe in the ceiling tile is a wonderful example of some of the things we ask our dogs to do, for rewards they don’t want and the problems that ensue when the trainer and the trainee are not on the same page.
So what can dog trainers learn from this anecdote? Let’s start with reward programs. As humans we are inundated with ineffective reward programs. What was wrong with this part reflects the poor understanding that almost everyone has about rewards and how they affect behaviour. The fact is, from a behaviour analysis perspective, rewards don’t change behaviour, because they are not necessarily linked to the behaviour. In order to for a behaviour to increase, it must be reinforced, which is the outcome of an effective reward. We may USE rewards to reinforce behaviour, but we can also use rewards at difference times than the behaviour. A reward in short is something we want, but not necessarily something that will reinforce a behaviour. This subtle difference becomes important when you look at how people use rewards. You can think of rewards as the tool that CAN or MIGHT use to reinforce behaviour, but you may also misuse the tool and inadvertently reinforce behaviours you are not interested in having.
The reward that was being offered was a day off. The behavioural criteria were the number of times that the rep said the client’s name, checking for updated information, that the phone was answered in a timely manner and saying the company name when thanking the client for calling. The reward would be delivered at a MUCH later date than the behaviour, and in fact, behaviour needs to be reinforced at the time it occurs in order to assure that it will increase. Not only that, but the reward was not actually contingent on the behavioural criteria; the reward was contingent on your name being pulled from a hat, and your name getting into the hat was contingent on your soccer ball getting into the goal, and the soccer ball getting into the goal was contingent on your ability to kick accurately over a given distance as well having carried out all four of the behavioural criteria four times in a random sample of your work. Getting a day off would be a relief, but not a reinforcement!
I had a client at one time who decided that her dog would only get the exercise he needed if he was well behaved until four pm that day. Any transgression would result in a lack of a walk. Lack of exercise made her dog quite unsettled and uncomfortable; he was a border collie who was bred to be very active. The frequency of this dog’s walks decreased bit by bit until he was spending most of his day being destructive in the client’s home. When the reward comes much later than the behaviour, or when the contingencies are unclear to the learner, as in the situation John found himself within, or that the border collie experienced, you will never make progress with teaching the behaviours you want.
Let’s consider too that kicking a soccer ball in a suit and tie was uncomfortable for John. The reinforcement would need to be pretty immediate and very valuable to get him to do the task reliably. John in a suit and tie is a sight to see and dignity is something that he demonstrated when he wore his dress clothes. Kicking a soccer ball in the office just isn’t dignified. How often do we ask our dogs to do things that they might feel is undignified? Treat on the nose trick? Some dogs like it, but more do not. The reinforcer may be immediate and valuable, but an awful lot of dogs just don’t like doing it and demonstrate their dislike in a manner not unlike what John tried to do; they avoid the situation if they can. In John’s case, the act of kicking a soccer ball in the office was probably the opposite of reinforcing; it was embarrassing and upsetting and likely decreased the likelihood of his doing what his manager wanted in the first place. If using the client’s name less frequently would avoid soccer ball kicking, John and many other employees probably would intentionally stop using the client’s name. When we ask the dogs to do things they don’t want to do, no matter how pleasant the outcome, many dogs just won’t do it.
I know that the promise of behaviourism is that we can get dogs to do anything they are physically capable of doing through manipulation of reinforcers. This is a case of just because we can, doesn’t mean we should. Really. We don’t need to train everything we are able to teach a dog. Dogs have preferences and trainers should learn about those preferences and train within them. I am not saying that dogs should not have to learn to do behaviours that keep them safe, like not pulling on leash, but honestly, we don’t need to train our dogs to do things they don’t enjoy just to amuse ourselves. What does this mean to you and your dog? This means knowing what your dog likes and doesn’t like, what he is physically and mentally comfortable doing. Ask yourself if the behaviour is both necessary and kind. If you have a dog who hates water, there is little point in training him as a hunting retriever; that is un necessary and unkind. If that same dog lives on an island and has to take a boat to and from the mainland, teaching him HOW to swim and keeping up with that skill is necessary to keep him safe in the event of a mishap, even though it may be uncomfortable and to a certain extent, unkind.
|This little guy doesn’t look thrilled about getting into the water; asking him to retrieve ducks would be completely un necessary and unkind, not to mention unrealistic. Image credit: cynoclub / 123RF Stock Photo|
The next thing to consider is the meaningfulness of the reward, the tool you use to increase the likelihood of the behaviour you want. In John’s case, the ultimate reward was a day off, something he wanted very much. The problem was that the reward was deeply buried in a complex and convoluted reward cycle that included an element of chance. Everything that happens between the time that you mark or identify the correct iteration of the target behaviour and the end reward is part of the reinforcement cycle. If you click and then walk your dog towards a treat station, both the walking and the treat station, as well as the treat itself are part of the reinforcement cycle. When the reinforcement cycles get too long, then the behaviour that you targeted gets lost in the process. If the ultimate reinforcer is something that the dog wants but the rest of the reinforcement cycle is onerous, then there are a good many learners who had the same reaction that John did; “thanks, no, I don’t need that badly enough.” Making John kick the soccer ball was like offering D’fer the chance to sleep in his crate overnight so that he can have a special breakfast in the morning; the reward is too far away, and the behaviour of sleeping in his crate is something that he doesn’t want to do in order to get it.
Finally let’s look at the target criteria. I have intentionally placed them out of sequence, because that is something that trainers often do. When you want your dog to come in straight on the recall and sit in front of you and wait till you tell him what to do next, and then you decide that sitting must be straight you cannot just tack that criterion onto the list willy nilly. When we are training, we have to suss out a single criterion, train that, and then suss out a second criterion and train that, and then put those two together. To make more complex behaviours we train each criterion individually, and then when we have achieved fluency we can put those behaviours together. If we get halfway through our process and then add something in the middle as an afterthought, the learner just gets frustrated. In John’s case they wanted the rep to say the client’s name four times, check for updated information, and say the company name when thanking the client for calling, and somewhere along the line someone added in the need to answer the phone within a specific number of rings. Instead of training each criterion as an individual behaviour, the managers tried to lump everything together and without actually reinforcing successful iterations of the target behaviour, they worked on all the behaviours at once. I see this problem so often in training classes that it is often amazing to me that the dogs get trained at all.
One of the most common places I see this is in teaching the recall. If pressed, the handler will state his target behaviour as the dog coming when called even if distracted. This is a really vague criteria, and it gets worse when the handler asks the dog to stay and then walks across a room with nothing between them and then calls the dog out of his stay, but only rewards the dog if he also sits when he arrives. I break coming when called into several discrete behaviours. I want the dog accept handling when he comes, so I teach a collar grab as a first step. Grab, feed and release; lather, rinse and repeat. When the dog is happy about being grabbed, we do a restrained recall where the dog is held back and the handler runs away and then stops and calls the dog. When the dog is good at that we put the two behaviours together; the collar grab AND the restrained recall. Then we do this with toys in between the handler and the person restraining the dog. Then we add people and dogs between the dog and the handler. Then we play a game where the dog is free and we call him and click and place a treat between our feet and move away while the dog is eating. Then we play a game where we click when the dog comes in close and then we throw the treat away to get the dog to move away from us. THEN we start calling the dog out of play. There really isn’t a link between coming when called in a distraction free environment and coming when called out of play. By working on each of the component elements of the come when called out of distractions and rewarding for the smallest successes, we reinforce the behaviour we want.
When you examine training programs it is tempting to use results as your only measure. From the perspective of the young managers who were trying to get better customer service, they thought they had a solid plan. They didn’t understand the difference between a reward (the tool) and a reinforcer (the effect), and they didn’t understand anything about the reinforcement cycle. They didn’t understand criteria, and the process of taking behaviours and deconstructing them into their component elements. In short, they didn’t really understand what they were doing. This is why we need to have instructors who really understand the process of training; so we can help our students to avoid the pitfalls of the shoe in the ceiling.