Well, first off, decide which you want to do - is it to go to a marker, or to run around it? Do you want it to be a post that your dog targets with their nose, or a mat they sit on?
With this sort of thing, it's often easier to start with the end behaviour and then work backwards from there. This is called "back chaining". I heard Donna Hill talking about an application that made me go "oh yeah!". She said that if you teach someone a song one verse at a time, but starting with the last verse and then adding them on one by one in front, then as the person learns it, they practice the stuff at the end more and more so they become more and more confident as they progress through the song. That is, for a song with four verses, they'd learn:
4
34
234
1234
Compare this to the normal way of learning songs from the start to the end, and how often we peter out as we've had more practice with the first bit but not the end.
Anyway, I digress, but it struck a chord.
So, let's go with the idea that you want your dog to run out and touch a post with his nose, this is a very simple chain with only two links. But we still want to start at the end; touching the post with his nose. You can shape that with clicking him for looking at the post, any stepping towards the post etc. Normal shaping protocol, where you increase your criteria until you have the confident nose target behaviour you're after. Once you have that, you can start adding the first link in the chain - the run to the post. You do this by gradually increasing the distance. Once your dog understands the game is "touch the post with my nose" and has fluency of movement, you can take a step away from it so he comes to you for his reward and then has to move that step towards it to touch it again. Gradually increase the distance. Put it on cue. Tada, you have a dog who will run to a post, touch it with his nose and then (because you've been rewarding out of your hand), come back to you for his reward.
