On the surface the statements seems to make sense and the arguments are well put across, but unless someone actually A/B tests it and proves that it makes a difference to readership, or conversions (or whatever your goal is) then it's currently just unsubstantiated theory.
One little known tip that I've been using for years to reduce bounce rates is to increase the text size by a few pixels.
I've successfully reduced the bounce rates on a few sites with that hack alone. On others it made little difference. I guess the older the audience, the more pronounced the effect.
Literally a guess. I've worked with very few sites that target older consumers, so my data points are limited, but in one of these the bounce rate dropped by more than 20% (home decor industry).
Unlike more recent experimentations, I did not conduct a formal AB/multivariate test, but the font size had been the only internal change at the time. Also, I have no idea whether these lower bounce rates translated into higher revenue.
Also, all of the sites mentioned above were female-oriented.
