Ok, so not defending Apple here, but I'm trying to figure this out. If this data is correct, it seems that 2017 actually had less reported issues with the keyboard than 2014 and 2015 - and that other service events have reduced more, therefore the larger percentage of keyboard issues in the number of total repairs.
Again, this is not me saying "see, you see, the keyboard issue is blown out of proportion" - but the article is not clear to me.
Because what it says is that Apple actually made a keyboard (the 2017 revision) that is just as reliable (there were actually less keyboard service events for the latest MBP than with 2014 and 2015 models) while the number of every other issue with the laptop is now smaller (so the percentage of the keyboard issues is higher).
What I'm saying is this: imagine you get 8 throat pains and 2 headaches every year. Then, you find medicine that cuts those throat pains and now you have only 2 of them every year, so 2 throat pains and 2 headaches. You could say that while in the previous years, headaches were only 20% of your total medical issues, this last year they were HALF of your issues! Technically, you are right, but it sounds like things are getting worse, while they are, in fact, getting better.
This is what the article looks like to me. Am I missing something here?