Amazon’s holiday success and UPS’ holiday fail highlight the internet economy’s problems


Shipping giant UPS failed millions of customers this holiday season, missing the delivery of “a small percentage of its packages” on the Christmas Eve, according to a statement it released on Tuesday. Meanwhile on the day after the Christmas Day, e-tailing giant Amazon is crowing about signing up more than one million Amazon Prime members last week and that it registered record number of orders. Later Amazon said it would offer shipping refunds on packages affected by the UPS delays. Both events are linked, and here is why.

The reality check

Amazon’s great success doesn’t have to be UPS’ failure, but in this case the culture and expectations of the web met the real world, and the real world experienced what the web kids call a “fail.” There are some obvious reasons for this, such as UPS’ decision to let its workers take Christmas day off or people’s general tendency to wait until the last minute to order a gift online. The physical world is no different, as anyone who hits a mall on Dec. 23 or 24 could see.

Perhaps sometime in the near future, with drone delivery Amazon can solve this problem, but the situation illustrates two big problems we’re going to keep bumping up against as we transact more of our business and lives online. The internet has turned us into slaves of instant gratification. When we want to listen to a song, we click and stream. When we want to read the latest book, there’s another click and it’s on our tablet. However, this is creating an expectation that is challenged by the physicality of the real world.

So the first problem is that the gap between the online expectations where everything moves pretty much instantly (or at least within tens of milliseconds) and the real world, where crossing thousands of miles means actually crossing thousands of miles over increasingly congested and crumbling infrastructure is going to seem ever larger.

Making the real world elastic involves tradeoffs

The second is that you can’t prepare the real world people or infrastructure for peak demand. Most of our roads are literally set in stone. Our workforce is not as fluid as a flexible internet-like network needs to be and it’s not clear if that’s the society we want or should want to build. We are shifting to the web economy faster than the real world can keep up.

Yet, the internet and digital mediums reward and even encourage peak demand — be it a viral video or hundred of thousands of people downloading Beyoncé’s new album overnight. To bridge the gap between real-world limits and internet demands we need to have a more accurate understanding of the real world, perhaps through sensors and data for incredibly accurate predictions.

There will be a need for intelligence — even in the dumbest of machines. Until then, we need businesses and governments to consider how to manage the internet expectations in the real world. For example during last week’s peak demand Amazon cut off some people’s ability to sign up for Prime to avoid crushing the system for its existing members.

On the UPS side, more data and insights about the health of its fleet and workers might have helped it establish its own limits and set a stopping point after which it would have to declare the system overloaded. At that point, its management could make a cost benefit analysis associated with staffing up or buying more trucks to meet the peak. Unfortunately for employees, this sort of elasticity tends to lead to contract work, unless we radically rethink how we employ people.

Also helping “solve” the problem of employees and peak demand will be robots. Amazon’s delivery drones or warehouse robots are an example of how these can help. Unlike UPS employees, robots don’t need Christmas Day off.

The combination of the expectation gap and the conflict between the internet’s encouragement of sudden peak demand and the physical world’s inability to deal with that demand economically will be one of the defining business and technical challenges of the next decade. Robotics, data and the internet of things will help, but we’re going to have to adapt societal institutions to make it work.

Or we can just accept a few late Christmas presents.



Let me solve this problem for you UPS/FedEX/USPS. If you are going to give guaranteed delivery dates around Christmas then give yourselves another day or two. Otherwise just say it is a “best effort” Christmas delivery, but not guaranteed.


What we are witnessing is the clash of two different economic rules: As Amazon fights to attract every last bit of retail it cannot afford to discourage customers from thinking – read feeling – that they can expect last minute gratification from Amazon just as they would if they walked into the mall on the last minute to pick a gift. As a result Amazon does not in any way price away last minute shoppers. Quite to the contrary it has to attract them with all kinds of last minute deals. Amazon can do this because it is sitting on load of cash that it is using to expand it fulfillment capacity, capacity which WILL NOT site idle the rest of the year. Why? Because Amazon’s business is expanding rapidly, so any spare capacity now will be fully in use soon, and nearly every package will be processed by these systems regardless of the shipping option.
Why can’t UPS do the same (invest in massive delivery capacity)? Because the moment a package leaves Amazon’s fulfillment centers, it DOES matter what kind of shipping method is chosen. UPS cannot profitably build a delivery system geared for a massive surge of next day and 2 day deliveries knowing that after the holidays most shipments will revert back to the lowest cost option.
For more see:


This is a solvable problem, and not by hordes of contract workers. Amazon need only (with a great deal of effort) match what they can promise with what they can deliver. I am certain UPS would give them guaranteed delivery in return for a guaranteed shipment size, but it would come at a price. This would require them to reserve spots, and tell people when those spots are full.

Mike Harms

I wonder if Amazon saw this coming. I ordered something on Friday the 20th around 2pm Central time. At that time I was told to expect delivery on the 24th using you Prime 2nd Day delivery. So I was surprised to see it arrive on Monday the 23rd. The package said it was sent next day air. Thought it was weird and my order clearly said 2 day shipping. I wasn’t really expecting it until Thursday as I had heard UPS was way behind earlier in the week.

Scott Ellsworth

This is a solvable problem, and not by hordes of contract workers. Amazon need only (with a great deal of effort) match what they can promise with what they can deliver. I am certain UPS would give them guaranteed delivery in return for a guaranteed shipment size, but it would come at a price. This would require them to reserve spots, and tell people when those spots are full.

This is a special case of the general principle – build in slack, or suffer. Tom DeMarco wrote a great book on this – slack looks like waste until you need it. Then it looks like planning.


Fail is not a noun in this context. Also, shop ahead of time if you can’t assume the blame.


Bull crap. The companies that guarantee delivery are at fault. I refuse to take the blame if I was guaranteed something that didn’t happen. Period.

dave parker

The holiday’s were great for Amazon and lousy for UPS. Apostrophe’s. And they’re use’s. Thank’s.

Don Goguen

Total BS story. Highlights a lack of knowledge on how the shipping industry works and how modern expectations for instant everything should work. It is all unrealistic. To expect any shipper will gear up for a one or two day demand cycle goes so far beyond the realm of reality as to be incredulous. Simply put, anyone who thinks UPS is going to be 100% on Dec 24th also believes in the Easter Bunny

Comments are closed.