I wrote several times about the need for any professional campaign (defined as a campaign with a budget of more than half a million USD) to start with a baseline survey. This being said, I now want to discuss the need and benefit of a tracking survey. First of all, the purpose of a tracking survey is to assess progress. Sure, candidates with a political instinct can sense if they resonate with people or whether the voters they talk to are just being polite. But how do you know you are making enough progress in view of your efforts and spending?
Also, many elections tighten when election day is approaching and it is good to have your own internal reading of public opinion comparing apples with apples.
Secondly, a tracking survey allows to recalibrate efforts and especially to adjust spending and targeting for the next part of the campaign. Anything you do in a campaign should either consolidate something you are about to lose, or give you something that you do not have yet. In that sense, a tracking survey defines low-hanging fruits.
I feel that candidates sometimes also conduct tracking surveys to check on their own people on the ground. It is not wrong in itself, but it is important to have a sample size that is big enough, and that thereby allows an accurate reading of results at the desired level. In unstable and/or new democracies, surveys are also conducted close to election day to prevent maneuvering on election day or on election night. I once had a client who wanted to time the fieldwork of the last tracking survey in a way that it would end on the evening before election day. Results were then delivered after the voting. You may think that this is moot and academic, but it is in fact a great way to assess the effectiveness of a get out the vote operation.