We know the optimal spend distribution will always be dependent on the time period we choose to optimize due to the effects of seasonality, strategy, etc. That means that we need to know the time period over which we are optimizing beforehand and we need to conduct an optimization dedicated to that time period.
This choice also allows us to maintain the performance of our dashboards. Because each time window comes with thousands of potential distributions, adding too many potential time windows will drastically increase the load times in dashboards. By providing a set number of useful time windows, we aim to strike a balance between dashboard performance and functionality