its just easy to under stand....
once you draw the state diagram of the convolutional encoder...
you'll be able to draw the trellis diagram of it... then you decide upon the best path that the encoder should have passed through... from which you decode the input.....