The benefits of tracking are in 3 main points
1. Reproducibility
As a data scientist I want to re-produce the same solution (results) again. so I will use the info I stored previously to get the same solution again.
2. Organization
You are not working alone, you team need to find where to find what. So they can compare their own results with yours and also to not trying to develop the same solution again while you did it already.
Even you are working alone like on a Kaggle competition you need to keep track of your progress so far.
3. Optimisation
Using the data you stored from the previous trials to solve the problem, you can optimize your solution.
The process of hyper parameter tuning still needs lots of manual work and not fully automated so you need to have a idea of what HyperParameters can make the solution better depending on the previously used ones.