Frequency data is vital for understanding how frequency can drive an impact on your results.
The first step is understanding how the frequency data was collected, especially if you are combining multiple data sources. You need to make sure that you are combining and analyzing the right things.
We call this, having a good data foundation.
Then take apart your data and sort it into clusters, what was the impact at each level of frequency. Was there an ideal number of impressions which had an impact. Was that output statistically significant?