Makes me wonder how much of this is based on experiment data. I’ve seen teams spend weeks on well crafted, thought-out welcome messages and walkthroughs, only to find (through analytics) that most users just X out of the walkthrough right away.
This is actually something that I'm missing in this analysis: walkthroughs and greeting messages are presented as a "must have", but unlike other conclusions in the report, I didn't see any data supporting these assumptions.