Short answer - as already implied in the piece - inadequate end-user testing.
My own experience in relation to this is two-fold. In a diverse career I have had experience in R&D, design and manufacturing of both hardware and software. In the former case I started a company that both designed and manufactured, on a modest scale volume-wise, professional (analogue) video signal processing equipment. Part of our distribution was through a major Japanese manufacturer for whom we were a niche supplier of equipment that wasn't in direct competition with their core products; I say this by way of validation that we were modestly successful - technically at least.*
From my own experience and my observation of my contemporaries' similar businesses I can say that the cost of fully testing first production-run examples of new products - at least at our scale of manufacturing - was completely prohibitive. So, unfortunately, the early adopters ended up doing the testing for us. This entailed some additional costs to keep our clients happy, but not nearly as much as full-blown production-prototype testing. I suspect, in this instance, that the camera manufacturers' financial strategy is similar.
At a later stage in my life I was responsible for testing and implementation of custom production software in the newspaper industry. This is an environment where you don't last long if the software fails unpredictably in an environment ruled by incredibly tight serial deadlines. Most of my working life was spent testing, frequently in parallel runs, and feeding back the results to programmers and production managers. Even so we still found that only when in the hands of the end-users in a live context would some additional quirks and bugs turn up.
I don't know how much this experience can be related to the photographic industry but I think it's not far out. For that matter I've found innumerable instances of domestic appliances that exhibit ridiculously deficient design. I have a juicer that belongs in a design museum's chamber of horrors. That said, some of the recent atrocities committed by camera manufacturers would only seem explicable by deliberate, wilful, crippling of products in order to maintain and feed the "upgrade" cycle. Some of the recent basic ergonomic blunders defy any other interpretation. Insert your own examples.
* My company was vandalised by banksters at the very moment it had started manufacturing a range of products that had had a 3-year development schedule: 24 years later this still baffles me, even if it's mildly gratifying to have my homicidal opinions endorsed by recent events
. Never mind; I have a product leaflet headed "Hitachi" to show for 7 years' work... In the aftermath it took me 10 years to just be flat broke again. Ho hum.