Actually, I think their approach is about right. People tend to follow a reasonably consistent schedule. If you plug your phone in every night and more or less nothing happens until 7:00 next morning, then only change to 80% (which is pretty conservative) or 85% (still avoids the levels that cause most of the damage to the lithium mesh) until about 6:30 and then charge to 100% and does the same based on recorded battery use throughout multiple averaged days, you can have a greater capacity when it is needed, but not have the charge above 90% except for short intervals. This would significantly extend battery life while not significantly reducing the usage time of the phone. Better modeling based on not just daily used, but also looking at day of the week would do even better.
I've been wishing laptops would do this for years. IBM and now Lenovo have been doing a far simpler battery life extension technique for years, but, when I have suggested that it would be good if FreeBSD or Linux do this, I've been unable to get much interest from developers and generally been unable to convince them that this would even be a good thing even when providing pointers to papers on research into l-ion battery behavior showing the significant damage keeping a battery at 100% for long periods does.
I saw a video several years ago that showed a photomicrograph recording showing the physical movement of the lithium grid during charge and discharge at or near "full charge" and how over-charging or even fully charging and holding that charge would slowly break down the grid. I put "full charge" in quotation marks because that number is picked by the manufacturer/designer as a trade-off between capacity and life. Wish I had saved a pointer to this. It was pretty dramatic.