For best results in positive reinforcement long term, you should be providing rewards of variable value from high value to medium value to low value to no reward. Initially, you want to use high value rewards when the behavior (or positive reinforcement in general) is new. Then you should switch to randomly providing high value or low value rewards, gradually moving to no or usually a low value reward for established behaviors. It should be like a slot machine, usually you get nothing or just a few coins back but every once in a while, you get a huge jackpot and that keeps it from getting boring.
I can't say where on the reward scale 'good dog' fits for your dog. I had one dog for whom that was reliably one of the highest value rewards and another for whom it was merely the marker that there might be a reward coming (ie like the slot wheels spinning.) But the important thing is the variable reinforcement.
There is a lot of good information about positive training on the Dog Scouts of America Training Page