I often see clients coming to me and saying that they have tried everything and nothing works.
By "everything" they usually mean: rewarding with food for good things; ignoring the bad things; crossing fingers and hoping the bad things will go away because the first two steps are being taken.
One problem is that bad things do not tend to go away unless you tend to these bad things accordingly.
Another problem is that some dogs tend to get into a headspace where they do not care anymore about your food, your toys, or your affection. What do you do then?
It is heavily ingrained in dog owners' heads that a dog has to be "positively reinforced" for doing what was asked of it. But what happens when the value of doing something you do not want your dog to do overpowers any reward you can provide?
I will tell you what happens: people tend to get lost because they believe they do not have leverage to motivate the dog.
A very simple and blunt reality is that you do not need to have "positive reinforcement" leverage to stop the behaviour. You just need to make the rehearsal of this behaviour impossible, unpleasant, and unrewarding. You need to have "negative" leverage.
Think about this: as a member of society, you have learned that yelling at people is not acceptable. Have you been encouraged for NOT yelling at people? Probably not.
You have either been lectured about why this is bad (and you stopped to make the lectures end) or you have been taught by some negative event that yelling does not result in anything good.
In other words, you have been either negatively reinforced to not yell or you have been punished for yelling. Of course, it is also possible that you never had a tendency to yell in the first place or that there was no lecturing but a simple explanation, but please entertain the idea.
Even though dogs are not humans, this specific learning mechanism works all the same, with the exception that, instead of reasoning with a dog like you would with a human, you would use your tools (your leash, your collar) to demonstrate to your dog that a behaviour is unacceptable and not to be repeated.
In this scenario, when your sole focus is on figuring out how to deliver this message of undesirability to your dog in the most efficient way, any thoughts about the rewards are on the back burner. It does not matter, in the context of this behaviour, whether you can reward your dog or not. Your goal is to stop the behaviour.
Once you realize that, it all becomes clear. You stop depending on treats and hoping that your dog will choose them over something in its environment. You stop pretending that you are progressing while you have been at it for thirty-eight months ("he is doing so much better, but there is still work to do"), because your delusion and your pretending do not serve either your dog or yourself.
Remove the crutch of having to reward for everything, because not every dog has enough food drive or toy drive to appreciate that reward. Every dog, however, understands the power of negative consequences, and absolutely every single dog, once presented with these consequences, will think twice next time.