I think that you are almost-but-not-quite hitting another problem of binge consumption patterns that I think might be very central to the problem of Netflix show cancellations:
If I'm a long term subscriber and I'm used to bingeing finished shows: will I even want to start a show before it's concluded or would I rather wait a few years and then watch it completely? Inter-season gaps are bad enough in broadcast, but with season-bingeing they extend to >11 months and that's just not very attractive.
If Netflix isn't excellently careful with what subsets of their data they look at this effect might blindside them very hard: are people maybe not starting S2 because they loved S1 so much that they'll deliberately not switch on S2.01 upon availability but would rather start with rewatching S1 it in its entirety once S3 hits? This information is not in the tracking data, better focus on intra-season dropouts. Success metrics can be hard.