A few key problems with this source.
1. It says very little about how the data is collected. Just "Geotab developed the tool based on an analysis of 6,300 fleet and consumer electric vehicles." that's not a lot of transparency.
2. It appears to be based largely on commercial vehicles (although see point 1, it's hard to tell). Consumers drive way less than commercial, which makes analysis by year less than useful.
3. It has a sample size of 6000 vehicles. That might sound like a lot, but since it is broken down into 21 different types of vehicle with about 3 models per type, that's only 100 vehicles per model+year. This would be less of a problem if they disclosed how many vehicles of each type they had, but they don't.
4. They also don't disclose how they have averaged vehicles.
5. The data they have that is easy to cross check (i.e. vs Tesla) shows that their numbers are way different from other sources https://electrek.co/2018/04/14/tesla-battery-degradation-dat... and https://www.evchargerreviews.net/should-you-worry-about-ev-b... that are much more transparent about their data.
All in all, it's very unclear with that source where the data is coming from, how it is analyzed, and it's conclusions differ dramatically from much better analysis.