1. Some of the proposed filters for the first spike don't make sense, for example devices for impressions. Think about how the filter affects all metrics and if the average rate needs to change when filtering.
2. In stead of using Scroll depth and time on page, how can we show the user how far and easy is their content to read.
3. How to let the user view where their customers come from and where they go after visiting the article.
Study and research about:
1. Filters. For the first approach we are just letting the user filter for period of time and date range.
2. Readability score in stead of using the scroll depth view to easily visualize how much someone has read.
3. Options to include a path to show the source and destination of visitors.
Medium uses some interesting approach to visualize just the important metrics for their audience: https://help.medium.com/hc/en-us/articles/215108608-Stats
As a result of this exploration we came out with new ideas and a new design for performance analytics which affects spike LPS-101121: