Read on Twitter
In my Amazon shopping basket, I have one "Fresh" item, one "Wardrobe" item, and one "Pantry" item, and then the normal shopping cart. Amazon federating its shopping cart (with different price limits to check out in each), is the result of Amazon doing autonomy poorly.
You need to keep the user and their needs first. A lot of companies have this problem... showing their "organizational underwear." This is exacerbated in large lean-style test-and-learn companies, where you have a lot of orthogonal experiments being presented to the user.
Amazon is one of the paragons of the autonomous team models and an early exemplar, but they are still constrained by having one pipe to push everything through. We had the same problem at Spotify... only so much real estate to present things to the user.
Google seems to handle this better by having early established lots of different products and being ruthless about protecting the main product from complicated UX.
Facebook has tried to fracture its' experience into multiple apps as a way to do this as well, but the results were a bit underwhelming because it hadn't done a great job differentiating the experiences when they were in the same app.
Adobe is another company that does this fairly well, but it is because it emerged as a collection of different purpose-built products that are each run as their own business (mainly because most came in via acquisition).
The teams that lead these products at Adobe are hyper-protective of their user base and generally don't want others messing it up. Of course, that sometimes leads to those teams falling prey to the same problem as bigger companies (see 3D being added to Photoshop, for example).
I don't have an easy twitter-able answer to this problem. I think starting early in the product definition phase to understand how you want to evolve your vision across multiple experiences and resist the temptation to keep adding things into whichever app is driving the revenue.
