Echo is a consequence of exploring digital appliances as marketing tools. If you're going to build a device which can voice-interface someone ordering 6 bottles of Tide with extreme ease, doesn't take much more to turn it into a nice music speaker and tell you weather & sports & jokes on request. Echo, Fire, etc are just extensions to feeding the retailing monster.
Do notice how Amazon's innovative foray into cell phones crashed & burned hard - despite being lauded as rather a well-built device. That shows how straying too far doesn't work, and that "far" isn't very.
While the hardware division was indeed shut down, Amazon did manage to salvage as much as they can.
The cloud service didn't take a life on its own - in the very early days there was just EC2 and S3, nothing else. It took a lot of focused investment, commitment and trust from the management to give the AWS experiment more time so that it could succeed and eventually it did.
Fire phone was one of Amazon's many experiments. It failed sure, but hey you miss 100% of the changes you don't take.
Just a minor point of correction: Actually SQS was first.
https://aws.amazon.com/blogs/aws/aws-blog-the-first-five-yea...
It's easy to point to one failed product as proof that "straying too far doesn't work", if you can do the mental gymnastics required to classify all their successful products as somehow more connected to their original goal.
"a logistics business, both information and physical"? Really? Calling their massive retail business, original content, and physical devices an "accessory" to their warehousing is an extremely myopic view of Amazon.