back to article SHOCK and AWS: The fall of Amazon's deflationary cloud

It's tough being a victim of your own business strategy, especially if you're Amazon founder and CEO Jeff Bezos. After spending twenty years applying a combination of price cuts, smart logistics, and ruthless efficiency to the publishing and music industries, Bezos' Amazon has grown into a vast company. But now the company's …

  1. Forget It
    Joke

    Ruthless cutting

    Bezos still has one more efficiency move to play though:

    Chop off one mammary gland.

    Like the mythical Amazon women.

    1. TitterYeNot
      Coat

      Re: Ruthless cutting

      "Chop off one mammary gland."

      Not sure that'll work - Microsoft lost one huge tit earlier this year, and it doesn't seem to be doing them much good...

      1. Anonymous Coward
        Anonymous Coward

        Re: Ruthless cutting

        "Microsoft lost one huge tit earlier this year, and it doesn't seem to be doing them much good"

        Apart from the highest share price in 15 years and growing market share and revenue in most key markets you mean?

    2. Stuart 22

      Re: Ruthless cutting

      I'm obviously not the target market - or the cutting is going to have to go way deeper.

      I've rented dedicated and VPS servers for over ten years - beginning with the price breaking RackShack @ $99/month a dedicated server. Indeed I'm still paying around that level today but getting more for my money (but not that much more!).

      I've looked into going cloudy but the prices are significantly higher yet offer no discernible advantage. The reliability on our servers (touch wood) has been exemplary - except from the VPS we have at AWS which doesn't fill me with confidence in using any more of their stuff.

      The obvious opportunity for us is to switch our backup servers to the cloud. We want to set up servers that can be turned on instantly when needed for production or updating. So most of the time the vendor is getting money for non-use and a premium when used. Overall that means they get more money for providing an active server and we don't pay much more than a nominal sum when it isn't being used. Less for us. Both win.

      But I've yet to see a pricing model that gives us that.

      1. Sir_Hops_A_Lot

        Re: Ruthless cutting

        I'm not the target market either: Google would have to get rid of the three jerks at the top, exterminate the part of their corporate culture that allows for rampant lying and cheating, and stop trying to force me to use the same log-in for all their products - so they can pin-point market the hell out of me - before I'd consider doing business with them ever again. If I'm reading you right, I think you want to look at Azure. Unless you did and found a problem.

  2. TheOtherHobbes

    AWS is a footnote in the Amazon business model - it's less than 10% of revenue - so this article is silly. AZ could walk away from AWS tomorrow and hardly notice.

    It's the Any Old Iron companies - IBM, HP, Microsoft, etc - who need to worry about commodity cloud pricing, because it's critical to their business models.

    If prices crash - and they will - they're going to be in a world of pain.

    1. smartypants

      The % of revenue isn't that relevant...

      What is relevant is the land-grab, which is why there was such quick and profound movement on price by Azure and AWS when Google moved its tanks on the lawn.

      People like me are putting a lot of effort into learning the technology stacks of what we consider to be the main players in this space. Over time, more of the IT world is going to sit in the datacentres of the successful players.

      Google knows it's playing catch-up, and it's even trying to make it easy to jump from AWS by offering 'AWS-compatible' APIs for some of its services.

      AWS can't complacently expect people to hang around if an alternative vendor wafts a far cheaper service their way, so their shareholders will just have to sit tight and watch AWS take it on the chin (for a change) to ensure they don't become the AOL of the cloud in 10 years.

      1. Anonymous Coward
        Anonymous Coward

        Re: The % of revenue isn't that relevant...

        "What is relevant is the land-grab, which is why there was such quick and profound movement on price by Azure and AWS when Google moved its tanks on the lawn."

        Google has managed to grab ~ zero share of the cloud sector it's just another 'me too' Linux vendor in that space. So it can only really compete on price - And the leaders in AWS and Azure can easily fight back as they have a much larger market share and benefit from the resulting economies of scale...

        "you mean like facebook have/will have/been/going to....."

        And Microsoft: http://arstechnica.com/information-technology/2014/01/microsoft-reveals-its-server-designs-and-releases-open-source-code/

    2. Anonymous Coward
      Anonymous Coward

      Amazon has a business model? Really?

      Whilst Amazon can scrap AWS at any point (they could do that with any of their divisions theoretically), the point is that Bezos has been playing essentially the shell game for nearly 2 decades. Growth is nothing without profits being realised at some point. It is all well and good controlling 90% of a market, but if that market is profit-neutral or loss making- you might as well control 0%- it would be cheaper.

      The halo of 'growth is king' is slipping, and Bezos has left himself with no where to go within his current portfolio, because he has engaged in what Michael Porter would term a 'race to bottom' in every vertical they've competed in. Rather than create a unique value chain, and protect that uniqueness and therefore command a premium on the price, they have always go 'it's books/music/compute but cheaper'. It cannot be sustained, and I suspect this decade will be the one where Amazon gets a sharp correction in its valuation and market reach as a consequence. Shareholders really don't care about market share- it's about dividends, and a zero profit company will never pay dividends.

      I admire the logistics that Amazon can pull off, but other than that there is little to admire. Between it's treatment of its staff as minions, to predatory market practice, to unethical tax loophole exploitation, it really couldn't happen to a nicer firm.

      1. Anonymous Coward
        Anonymous Coward

        Re: Amazon has a business model? Really?

        "Rather than create a unique value chain, and protect that uniqueness and therefore command a premium on the price, they have always go 'it's books/music/compute but cheaper'. It cannot be sustained..."

        And yet that is exactly what made Microsoft the biggest and most profitable IT company in the world. "It's software but cheaper".

        1. Anonymous Coward
          Anonymous Coward

          Re: Amazon has a business model? Really?

          Really late reply- but no, Microsoft didn't. They did actually innovate and offer uniqueness (then got lazy and stupid)- Windows had fewer barriers for others to develop on its platform (hence the HUGE back catalog of LOB applications that isn't on Unix or OSX), was pretty easy to use compared to the 'Nixs of the world, and didn't insist on you spending $100s just to put a floppy drive on your shiny new machine as an upgrade like Apple did when it scraped the ISA backplane (that it invented) and forced SCSI adoption because it didn't like 3rd parties manufacturing cheaper components.

          Microsoft were never the cheapest solution on the market- but they were never the most expensive. They also always had a quite generous markup- one they haven't really needed to drop the compete. Amazon has NEVER MADE A PROFIT.

          Compare that to Apple who since the Jobs Renaissance have always made a profit. Guess what Apple don't do? Compete in markets they can't make more than they spend in. That's why they are the big boys on the block today. The irony of that is, it's in no small part due to the cash injection Microsoft gave Apple in 1998- bet they wish they hadn't sold that stock in the early 00s now- well done again Ballmer. A cash injection based on the huge profits they were making on Windows OS & Microsoft Office sales.

    3. Anonymous Coward
      Anonymous Coward

      Price of iron?

      That doesn't make sense to me. If the market price of a given server is $X, that can't be affected by what anyone does with the servers after they buy them. On the contrary, the lower cloud service prices fall, the greater demand will presumably get - which means the cloud providers will need to buy MORE iron quickly, which if anything should drive up the prices manufacturers can ask.

      Unless people like Google have found out how to build thousands of powerful servers much more cheaply than experienced manufacturers like IBM. If so, they could indulge in vertical integration and make their own servers instead of buying them. But that doesn't sound a very safe strategy in the long run.

      1. Naughtyhorse

        Re: build thousands of powerful servers much more cheaply than experienced manufacturers like IBM.

        you mean like facebook have/will have/been/going to.....

        1. This post has been deleted by its author

      2. ratfox

        Re: Price of iron?

        Unless people like Google have found out how to build thousands of powerful servers much more cheaply than experienced manufacturers like IBM. If so, they could indulge in vertical integration and make their own servers instead of buying them.

        Unless I'm mistaken, Google (and Facebook, and Amazon, and Microsoft) have been doing that for years. This is not necessarly because they have more knowledge than IBM, but because they have very specific needs, and can take some very specific shortcuts.

        One example is that IBM servers, even when built for racks, seem to have graphic cards. I would be extremely surprised if cloud giants bothered with such a thing for machines in their data centres.

  3. Craig 2

    One of the points that always seem to be missed is that before all the price cuts, they were ALL making OVER 50% profits. (Assuming they can still turn a profit now) They were raking it in!

    1. Paul Crawford Silver badge

      And people wonder why the likes of MS and Adobe are pushing on-line versions? Money, regular money, and no need to offer something new/better to keep getting paid.

      1. Naughtyhorse

        MS and Adobe are pushing on-line versions

        I'm beginning to come round to this idea TBH

        Anything that ends the Sisyphean 'upgrade' (what-the-fuck-grade) cycle of annual updates with no significant productivity gain (all more than offset by the bugs in the new bits), but different enough to require staff training or variation to your internal systems. Who knows someone might even come up with the innovation of software that just does what it's supposed to do, and wont go titsup every 10 minutes. and if it helps the vendors address piracy, than that's all good too.

        1. Paul Crawford Silver badge
          Unhappy

          Re: @Naughtyhorse

          If this brave new world of on-line services resulted in an end to stupid GUI changes and features no one wants, and instead bugs were fixed and actual improvements made I would support it.

          Sadly from my (limited) experience of MS and Google, it is the usual stupidity but on a regular payment plan.

      2. Anonymous Coward
        Anonymous Coward

        "And people wonder why the likes of MS and Adobe are pushing on-line versions? Money, regular money, and no need to offer something new/better to keep getting paid."

        I can't comment on Adobe, but with Microsoft one of the benefits of the online versions is far MORE frequent enhancements, new features and updates. And you don't have to do anything - the updates are all transparent / automatic.

        The 'regular money' is better for customers too - they can spread the costs of software over time, and can scale up / down as required - with no underlying software investment lock in. And incurring Opex. is usually more tax efficient than Capex.

        1. Michael Wojcik Silver badge

          with Microsoft one of the benefits of the online versions is far MORE frequent enhancements, new features and updates

          For the vast majority of software I use, that's not a benefit. It's an annoyance. More-frequent bug fixes that don't affect the user experience other than correcting an error would be a benefit. Once in a great while, an enhancement is a benefit; usually it's disruptive, because it changes the user experience or interrupts my workflow or reduces performance without doing anything useful for me.

          Even new features that are useful to me are temporarily disruptive, and I want to be able to roll them in on my schedule, not when the vendor decides to push them out.

  4. Charlie Clark Silver badge

    After spending twenty years applying a combination of price cuts, smart logistics, and ruthless efficiency

    You forgot: underpaying workers, underpaying suppliers and using tax numerous dodging strategies.

    Amazon's online shopping model is fundamentally flawed because individual delivery incurs very high costs. This is why margins have always been terrible.

    To his credit Bezos has always been clear about his aims and impressive with innovation. AWS was an interesting development due to Amazon's need for massive computing power to cope with peaks such as Thanksgiving in America and Christmas around the world. Kindle and LoveFilm are both attempts to get out of the ruinous business of storing and shipping physical goods individually. Publishing and investing in video production are attempts to move up the value chain.

    1. Steve Davies 3 Silver badge

      Hmmmmm

      Quote

      Amazon's online shopping model is fundamentally flawed because individual delivery incurs very high costs. This is why margins have always been terrible.

      I wonder if this is the reason why Amazon are going to start delivering their own stuff, much to the chagrin of the likes of ParcelForce etc?

      1. Warm Braw

        Re: Hmmmmm

        A lot of the home delivery market seems to rely on "self-employed" drivers who have to work crazy hours to make any sort of return at all.

        In other words, it's already a fundamentally loss-making activity, with the loss currently foisted on the people at the bottom of the pile. If Amazon want to add a genuinely loss-making activity to their portfolio of strategically loss-making activities, it won't be to anyone's chagrin but their own.

        1. Anonymous Coward
          Anonymous Coward

          Re: Hmmmmm

          "In other words, it's already a fundamentally loss-making activity, with the loss currently foisted on the people at the bottom of the pile..."

          Hello! Don't you people know ANYTHING about modern business strategy? That's called externalizing, and it's a crucial part of most successful corporate policies. Microsoft, to take the most glaring example, has made obscene fortunes by foisting the losses on its unfortunate users. (Well, they were unfortunate to be so dumb that they thought they could buy the cheapest software on the market and not pay in other ways - and so dumb they didn't even notice the other ways).

        2. Michael Wojcik Silver badge

          Re: Hmmmmm

          A lot of the home delivery market seems to rely on "self-employed" drivers who have to work crazy hours to make any sort of return at all.

          A lot of it, but not all. In dense urban areas, with good logistics, it can be profitable.

          If Amazon want to add a genuinely loss-making activity to their portfolio of strategically loss-making activities

          I suspect what they want to do is cherry-pick the profitable parts.

          We've seen this in the US with delivery companies and the USPS. There were profitable post offices in the US, such as Omaha, which handled a lot of commercial mail (catalogs and things) and local packages. The delivery companies could undercut USPS in the profitable markets by offering lower prices there, while charging higher prices for delivery to remote areas. USPS, meanwhile, was constrained to offer universal service and flat-rate delivery (by category) for many items.

      2. Anonymous Coward
        Anonymous Coward

        Re: Hmmmmm

        "Amazon's online shopping model is fundamentally flawed because individual delivery incurs very high costs"

        I can ship reasonably sized parcels for £1 each plus VAT by courier, so I imagine the cost to Amazon is even less. That 'flawed' model seems to be working surprisingly well for billions of pounds of sales every year.

    2. Anonymous Coward
      Anonymous Coward

      "You forgot: underpaying workers, underpaying suppliers and using tax numerous [sic] dodging strategies".

      You seem to forget "a penny saved is a penny earned".

      Which morally beautiful altruistic corporations do you have in mind as counter-examples? As far as I can see the American (and hence "Western") corporate philosophy is crisply summed up by Phineas T Barnum's maxims:

      "Never give a sucker an even break" (whether customer, employee, partner, stockholder, or government)

      and

      "There's one born every minute" (so you needn't worry about alienating customers).

  5. bazza Silver badge

    Fundamental limits

    The fundamental limit on the costs of running a cloud are hardware, bandwidth and engergy. Once they've all thinned down to those costs alone then there's not going to be much to choose between them. Energy ain't getting any cheaper... Are we going to see Google Nuclear before too long? Amazon Atomics anyone?

    1. Anonymous Coward
      Anonymous Coward

      Re: Fundamental limits

      The Thorium reactor, coming to a in a neighbourhood near you. ;)

    2. ROlsen

      Re: Fundamental limits

      There is room for differentiation. There are workloads that benefit from different types of computing, (e.g. search and FPGA's), that the cloud providers could offer.

      1. bazza Silver badge

        Re: Fundamental limits

        ”There is room for differentiation. There are workloads that benefit from different types of computing, (e.g. search and FPGA's), that the cloud providers could offer.”

        Not really. Someone somewhere else will be offering the same blend of chips (FPGA, GPU, CPU). If a cloud provider is buying the 'best' chips in the open market then the others can also buy the very same ones.

        The only way to truly differentiate is to make their own chips and be better at that than anyone else. It's the only way to get exclusivity. But getting into the chip design business is a huge challenge and fraught with pitfalls. Being 'compatible' means a struggle to be better than the original. Being incompatible means no one will have any software or firmware to run on your cloud and any efficiencies you managed to achieve don't count for anything.

    3. Anonymous Coward
      Anonymous Coward

      Re: Fundamental limits

      Maybe they can move into the mass heating business...

    4. Turtle

      @bazza

      "Are we going to see Google Nuclear before too long? Amazon Atomics anyone?"

      Google did at one time have a specific project that they, in their arrogance and ignorance, thought would be able to solve the problems of nuclear fusion.

      They didn't get too far with that. So now they are going to solve everyone's health problems.

      Don't expect too much from that either.

  6. Anonymous Coward
    Anonymous Coward

    Both Google and Amazon vulnerable

    If these companies are forced to pay tax and bring profits onshore, their balance sheets will suffer just as much as through price wars.

    The biggest growth market is China, and through US Govt ineptitude over a long period, these companies are mostly excluded from it. Amazon has a stake, but can't compete with local players on goods for the China market. Alibaba or Baidu could enter the US market simply to hang these companies by their own rope, as they say in China.

    Article omitted Amazon TV and Google TV, which surely are growth areas in the short term among the global couch potatoes.

    1. DavCrav

      Re: Both Google and Amazon vulnerable

      "The biggest growth market is China, and through US Govt ineptitude over a long period, these companies are mostly excluded from it."

      It's probably more to do with the Chinese government than the US government, to be fair.

    2. Notas Badoff

      Such a trick we've done!

      The Chinese blaming US government ineptitude for the 'necessity' of Chinese protectionism to aid domestic industry and defeat the Western imperialist dogs who scheme to overrun the motherland with their corrupting products and turn the minds of our dear youth to ruin.

      It's so much worse if it is a foreign company than an internal company.

  7. Arctic fox
    Headmaster

    Amazon is a (brain dead) predator:

    In the same sense that at vampire is a predator. Their strategy in practice (in the longer term) destroys the market they are exploiting*. They are a bubble company. The moment they are required (by circumstances - shareholders, the authorities and the like) to show a (sustained) profit then they are in the shit up to their necks.**

    *The only world a vampire understands is one where he is last of his kind left standing.

    **Oh and yes, I am perfectly well aware that there are plenty of other examples of BigCorp out there who deserve the same critisism.

    1. Anonymous Coward
      Anonymous Coward

      Re: Amazon is a (brain dead) predator:

      "They are a bubble company. The moment they are required (by circumstances - shareholders, the authorities and the like) to show a (sustained) profit then they are in the shit up to their necks".

      Just like the USA, then.

  8. bigtimehustler

    All of this misses one point, the decision of who to use is not just a price based one. As far as features are concerned, AWS is a still light years ahead of the others. Granted it what you need is provided by them all, your fine, but if you use something only AWS provides, it isn't going to matter how much cheaper others are.

    1. bazza Silver badge

      ”All of this misses one point, the decision of who to use is not just a price based one. As far as features are concerned, AWS is a still light years ahead of the others. ”

      However one of the most fundamental points of marketing is that cleverness and 'quality' doesn't sell. Look at iPhone; didn't have multi tasking, didn't even have copy and paste, fairly rubbish power hungry OS, put it in a nice case with a large screen and sell by the millions. Android is still a pretty low quality piece of software with many disastrous flaws, sells by the bucket load. BlackBerry's BB10 is a fantastic OS, well thought out design that actually allows you to do many things easily, quite complicated ideas, no one bought it. VHS vs Betamax; VHS won because it was cheap and no one cared for Betamax's quality. Itanium was quite a good chip, no one cared.

      History shows us that if your customers need to be geniuses to see why your offering is better than anyone else's then you customer base is at best 5% of the overall market. The other 95% are either too lazy or stupid to work out what the best solution is for themselves and will resort to judgements on price and trivial differentials such as looks, feel, etc. AWS may well be light years ahead of anyone else's cloud, most of the market won't care or understand why.

      1. Mark #255

        This old canard...

        VHS vs Betamax; VHS won because it was cheap and no one cared for Betamax's quality.

        Also, the (original) maximum recording length for NTSC Betamax tapes was 60 minutes. Far fewer people cared for an intermission in their movie than the putative better quality that Betamax's shorter tapes offered.

        1. Anonymous Coward
          Anonymous Coward

          Re: This old canard...

          "VHS won because it was cheap and no one cared for Betamax's quality. "

          No, No, No. VHS won because it had porn for rent.

  9. Anonymous Coward
    Anonymous Coward

    Nothing to see here

    "Its price-to-earnings ratio was still over 500." (http://www.nytimes.com/2014/04/26/business/amazons-shrinking-profit-sets-off-a-seismic-shock-to-its-shares.html?_r=0) and the average PE is around 25. Pop !

  10. kmcowan

    The Caveat:

    Google's APIs are quite lame, poorly documented and wonky. Amazon's AWS API is solid, and much easier to utilize.

  11. Matt Bryant Silver badge
    Alert

    What about the smaller players?

    This article is only looking at the Big Three and doesn't consider what this price war will do to the smaller cloud companies. They don't have alternative business lines to supply money for a cloud price war, in essence they are already competing at a disadvantage because their smaller size means they do not get the advantages of scale as AWS, Google and Azure do. This is like arguing over who will win in the out-of-town supermarket wars whilst not noticing the little family shops on the highstreets are all struggling and dying.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like