UPS helped move Amazon towards its own network when it refused to deliver on Christmas a few years ago and sent everybody home, leading to massive refunds and embarrassment by Amazon.
Until now, the reality is that no outside entity was going ultimately to be able to handle Amazon's exploding volume. The numbers are so vast, and the growth was so vast, that massive capital investments were going to be needed by either carrier, and they didn't want that. Amazon's growth will slow eventually, but for now what Amazon takes in-house isn't per se necessarily coming out of somebody else's volume. Certainly at the beginning of the Amazon in-house network (and remember that a huge component of that was ground line-haul, which was well established before the air line-haul was even an idea), every bit of what was going into it was coming out of growth, and FX/UPS were also seeing growth from Amazon.
We are going through a national (actually, global) transition to e-commerce, which was always going to cause changes for two reasons: (1) massive additional volume (2) It's dominantly B2C rather than B2B - but UPS and Fedex were always more B2B.
So, UPS started to fail under the strain of peak deliveries, but more broadly, when you're Amazon, you have these quarter-billion+ assets known as FCs (specifically, AR sortable warehouses) and those only make sense when they are run 24/7/365, which requires seven day a week deliveries.
Further, Amazon has grabbed more and more share over time by offering selection and convenience, and convenience means getting your stuff ASAP. Which means if someone finds their star-spangled widget on Friday morning on the Amazon website, they're far more likely to buy it if it arrives Saturday rather than Monday. Consumers don't care about weekends - in fact, weekends is when many consumers used to do their shopping, so you better be able to get stuff to consumers on the weekends just like any other day.
So Amazon has financial and consumer satisfaction reasons for wanting its stuff delivered ASAP. The heck with weekends, the heck with traditional UPS and Fedex five-day-a-week delivery cadence. It needs seven day a week delivery, period, end-of-story.
And UPS and Fedex were not set up for that.
Further, Amazon wants to get inside the economics of a delivery network. A huge percent of the cost of delivery is the last mile. If two packages are delivered to the same address on the same day, that's money in the pocket of UPS. Amazon wants that.
Some of that, Amazon can do itself. At least four years ago, Amazon software started allowing consolidation of packages. If you made two orders in the same day, Amazon had the ability to figure that out and consolidate stuff within the same box when possible. Amazon has started encouraging this kind of thing by "subscribe and save" programs whereby if you subscribe for certain stuff on a certain schedule (every month, every two months, every week, every X weeks) it allows Amazon to group those deliveries as well.
But then think about, say, an assisted living center for older folks with a single mailroom. If UPS delivers dozens of packages a day there, that's a huge score for UPS, because it's a single last-mile trip for dozens of packages. Amazon wants that too. Same thing for a big apartment complex with a single mailroom. Or a NYC doorman building. Amazon wants the savings from that single last-mile trip that covers dozens, maybe even over 100 packages in a single day.
It's true more broadly too. The more dense the deliveries (i.e. more packages per mile) the more efficient is the last mile. It would be interesting to know what fraction of all addresses Amazon's own delivery vehicles pass by in a given city on a typical day. It must be pretty dang high by now. And the higher it is, the more cost effective is every additional package delivered to that city that day. Amazon wants those economics.
My guess is that as Amazon figures out what addresses are most economic, you'll start seeing targeted promotions. Special programs for folks living in assisted living facilities. Frankly, delivery of stuff like dietary supplements, food, vitamins, etc (on large-print version of websites) can be a great thing for older folks - and it's fantastic for Amazon if it's going en-mass to a single mailroom.
I could see rural delivery version of Amazon - Amazon cafes/convenience stores near the remaining business centers of rural counties. Pick up your Amazon packages at a central location and get a free cup of coffee, low-priced essentials like milk, and chat with your neighbors at the same time. Or maybe there will be some kind of benefit where if you pick stuff up at the central location X times, you get a free something else - a reduction your Prime membership fee, perhaps. A loyalty program of some kind to induce people to pick up centrally. That tradeoff could be absolutely worth it for Amazon if it allows them to reduce the cost of the last mile. If they do it right, they'll get credit for encouraging a sense of community at the same time.
There's a ton of room for creativity on the part of Amazon going forward, that's for sure.