Amazon’s Scorpion Problem

I expect most of you know the fable of the scorpion and the frog, but if you don’t, it goes like this (quoted from Wikipedia):

“A scorpion asks a frog to carry him over a river. The frog is afraid of being stung during the trip, but the scorpion argues that if it stung the frog, both would sink and the scorpion would drown. The frog agrees and begins carrying the scorpion, but midway across the river the scorpion does indeed sting the frog, dooming them both. When asked why, the scorpion points out that this is its nature. The fable is used to illustrate the position that no change can be made in the behaviour of the fundamentally vicious.”

Over the weekend, there was some commentary on AWS in fight of its life as customers like Dropbox ponder hybrid clouds and Google pricing. Amazon turned in slightly declining quarter-over-quarter revenue on AWS, although significant year-over-year quarterly growth, as explained in Sign of stress or just business as usual? AWS sales are off slightly.

“Could Amazon Web Services be feeling the heat from new public cloud competitors? Maybe. Maybe not. Second quarter net sales of AWS — or at least the category in which it is embedded– were off about 3 percent sequentially to $1.168 billion from $1.204 billion for the first quarter. But they were up 38 percent from $844 million for the second quarter last year. In the first quarter, growth in this category year over year was 60 percent. So make of that what you will.”

Could Amazon’s nature be catching up with it, or is it just operating in a more competitive market? A set of emails went around from some of the CEOs of our companies talking about this followed by a broader discussion on our Foundry Group EXEC email list. It contained, among other comments:

  1. AWS is not the low price provider.
  2. AWS is not the best product at anything – most of their features are mediocre knock offs of other products.
  3. AWS is unbelievably lousy at support.
  4. Once you are at $200k / month of spend, it’s cheaper and much more effective to build your own infrastructure.

While we are in the middle of a massive secular shift from owned data centers to outsourced data centers and hardware, anyone who remembers the emergence of outsourced data centers, shared web hosting, dedicated web hosting, co-location, and application service providers will recognize many of the dynamics going on. Predictably in the tech industry, what’s old is new again as all the infrastructure players roll out their public clouds and all the scaled companies start exploring ways to move off of AWS (and other cloud services) into much more cost effective configurations.

Let’s pick apart the four points above a little bit.

1. AWS is not the low price provider. When AWS came out, it was amazing, partly because you didn’t need to buy any hardware to get going, partly because it had a very fine grade variable pricing approach, and mostly because these two things added up to an extremely low cost for a startup relative to all other options. This is no longer the case as AWS, Microsoft, and Google bash each other over the head on pricing, with Microsoft and Google willing to charge extremely low prices to gain market share. And, more importantly, see point #4 below in a moment. Being low priced is in Amazon’s nature so this will be intensely challenging to them.

2. AWS is not the best product at anything – most of their features are mediocre knock offs of other products. We’ve watched as AWS has aggressively talked to every company we know doing things in the cloud infrastructure and application stack, and then rather than partner eventually roll out low-end versions of competitive products. We used to think of Amazon as a potential acquirer for these companies, or at least a powerful strategic partner. Now we know they are just using the bait of “we want to work more closely with you” as market and product intelligence. Ultimately, when they come out with what they view of as a feature, it’s a low-end, mediocre, and limited version of what these companies do. So, they commoditize elements of the low end of the market, but don’t impact anything that actually scales. In addition, they always end up competing on every front possible, hence the chatter about Dropbox moving away from AWS since AWS has now come out with a competitive product. It appears that it’s just not in Amazon’s nature to collaborate with others.

3. AWS is unbelievably lousy at support. While they’ve gotten better at paid support, including their premium offerings, these support contracts are expensive. Approaches to get around support issues and/or lower long term prices like reserved instances are stop gaps and often a negative benefit for a fast growing company. I’ve had several conversations over the years with friends at Amazon about this and I’ve given up. Support is just not in Amazon’s nature (as anyone who has ever tried to figure out why a package didn’t show up when expected) and when a company running production systems on AWS is having mission critical issues that are linked to AWS, it’s just painful. At low volumes, it doesn’t matter, but at high scale, it matters a huge amount.

4. Once you are at $200k / month of spend, it’s cheaper and much more effective to build your own infrastructure. I’ve now seen this over and over and over again. Once a company hits $200k / month of spend on AWS, the discussion starts about building out your own infrastructure on bare metal in a data center. This ultimately is a cost of capital discussion and I’ve found massive cost of capital leverage to move away from AWS onto bare metal. When you fully load the costs at scale, I’ve seen gross margin moves of over 20 points (or 2000 basis points – say from 65% to 85%). It’s just nuts when you factor in the extremely low cost of capital for hardware today against a fully loaded cost model at scale. Sure, the price declines from point #1 will impact this, but the operational effectiveness, especially given #3, is remarkable.

There are a number of things Amazon, and AWS, could do to address this if they wanted to. While not easy, I think they could do a massive turnaround on #2 and #3, which combined with intelligent pricing and better account management for the companies in #4, could result in meaningful change.

I love Amazon and think they have had amazing impact on our world. Whenever I’ve given them blunt feedback like this, I’ve always intended it to be constructive. I’m doubt it matters at all to their long term strategy whether they agree with, or even listen to, me. But given the chatter over the weekend, it felt like it was time to say this in the hope that it generated a conversation somewhere.

But I worry some of the things they need to be doing to maintain their dominance is just not in their nature. In a lot of ways, it’s suddenly a good time to be Microsoft or Google in the cloud computing wars.

  • I had never heard the Scorpion fable, but I will be using it from now on, whether it actually fits my point or not.

    • It’s one of my favorites.

    • “whether it actually fits my point or not.”

      “when in Rome…”

  • technosophos

    This is why I have often referred to AWS as the “Walmart of the Cloud”. You’re not paying for quality, your not paying the lowest cost, and your not paying for customer service. You’re just paying for the convenience of getting (most of) what you need from one vendor.

    But alternatives, like OpenStack, are stumbling along well behind AWS. Rackspace, once trumpeted as a leader in the space, very publicly put itself on the market… but could find no buyers. HP, IBM, and RedHat all very publicly touted their plans for public cloud. None are seeing fast adoption. MS Azure may at first seem like a surprise success story, but compared to Amazon’s poor Windows offerings, it should be no surprise that the legions of MS-oriented developers would trust Azure over AWS.

    To choose to not use AWS is to choose establishing multiple relationships with multiple vendors (or to choose building internal cloud infrastructure instead of using DBaaS, LBaaS, etc.). And going that route will entail no small amount of systems integration work, along with complicating accounting and systems administration.

    Until ops spend hits the $200k range, a private cloud is not really a tenable solution.

    So small cloud-centered orgs (and startups in particular) are stuck roaming the virtual Walmart aisles of AWS, hoping that we can scrape together enough mediocre infrastructure to “get by” until we can afford quality infrastructure. That is, until OpenStack (or Google, or Microsoft) manages to offer some modicum of competitive services. And a little competitive pressure would certainly spur AWS to address your #2 and #3.

    • kordless

      I’d hesitate to say OpenStack is ‘stumbling’, given its youth (4 years tomorrow, woot), Open Source heritage and a different delivery model than AWS. We could compare it to VMWare (16 years old), which is at a $5-6B/yr run rate. In comparison, AWS (8 years old) is doing about $5B/yr. OpenStack is the whippersnapper of the cloud.

      A recent OpenStack survey showed over 500 enterprise deployments by the same number of companies for OpenStack. That number is continuing to accelerate, especially with larger deployments, given the cost savings it represents to the companies using it. It’s also accelerating because the software actually works now and does (mostly) what it promises, given you have good people helping with it.

      I agree about the ops range for justification of deploying your own private cloud, but I also think there exist other models that make sense, given you could solve trust issues with some type of simple, lightweight federation.

      I’m working toward that end by marrying Bitcoin/BitAuth to OpenStack: It is my belief the only way we will get true cloud computing is by having everyone cooperatively participate in sharing of infrastructure resources. Granted, there’s a lot to do to get there, but it’s a better promise for the world than running on a scorpion’s closed source code that could be doing bit knows what with your data.

      • technosophos

        I guess my choice of the word “stumbling” was poor. “Lumbering” is probably a better word. OpenStack is just a slowly moving project. It’s the networking layer that is still raw enough that OpenStack can’t quite compete in the same league as AWS. As someone’s who’s contributed to OpenStack, though, I can say that I do believe it is *poised* to be a threat to AWS. It’s just far enough behind that public cloud providers who rely on OpenStack will struggle to match AWS offerings on a significant scale.

  • IMHO it’s inevitable that Google will win this. Google is, ultimately, the best in the world at dev ops, information retrieval, and scalable architecture. And they can do it cheaper than anyone else to boot. Google has proven track record of winning without first mover advantage, and unless Amazon is willing to go all-in to win this war (hard to do while working on phones, tablets, logsitics, distribution, etc.), they wont win.

    When it comes to logistics, distribution, and fulfillment as a service then Amazon has the upper hand.

    But when it comes to winning the “platform as service” battle, my money’s on Google.

  • “In a lot of ways, it’s suddenly a good time to be Microsoft or Google in the cloud computing wars.”

    I like that takeaway. Microsoft with Nadella at the helm is exciting. I’m genuinely excited to follow along over the next few years as they take on Amazon, Google, Apple, and the rest.

    Part of it, I think, is that watching Yahoo under Marissa Mayer’s got me excited for another turnaround story. But another part of it is that it’s interesting to see MSFT thrust into the scrappy underdog position in many of it’s categories going forward.

    And – my scorpion instinct – it’s in my nature to always root for the underdog.

    • One of the things I’ve learned over and over and over again in this industry (since I started my first company in 1987) is that you should never, ever count out the underdog, whether it’s a big company entering a new market, an existing company re-inventing itself, or a company you’ve never heard of showing up at the scene.

  • Sarah O’Keefe

    What percentage of AWS customers are above the magic $200K/month? Or better, what percentage of AWS revenue comes from the behemoth customers?

  • On point. #1 I always think of the FB story of them hustling to find servers all over town. That has certainly changed. Nevertheless, it’s shocking comparing costs across providers to scale. There are a multitude of factors to consider / compare moving from shared to VPS or dedicated / cloud servers (migration support, number of concurrent users without crashing, storage, restrictions…) and each has its own limitations. However, even GoDaddy and iPage make way more sense from start-up to enterprise on a budget than Amazon. Both offer infrastructure, 24-7 support and specialists. That’s less required of your own IT team if you have one.

    #2 and the whole scorpion analogy made me think of Netflix and Amazon Prime’s movie service. Many hosts require using Amazon for video hosting / streaming (the aforementioned companies don’t). I certainly wouldn’t trust them with anything novel. #OnlyTheParanoidSurvive

  • Todd Vernon

    #5 public cloud providers including AWS have very poor performance for highly transactional businesses. Find an adtech business that hasn’t moved to bare metal. Just can’t do it in the public cloud yet.

    • This is correct for transnational businesses. I think the number is much less than $200k a month. More like $200k a year. 1 42U rack costs less than $20k a year. Add another $80k for servers (depreciated over three years) and another $20k for extra bandwidth, halve that for a disaster recover site, and add in the extra work of management and I think you are at right around $200k and you have lost a dependency.

    • Terminator

      All bull – no substance – you’d be surprised about the number of companies in the space on AWS.

  • josh

    AWS was going to revert to the mean at some point. Amazon is a middleman. Their differentiator is selling 200 m skus against bby or wmts’ < 1million. Not in building focused products with deep moats. They know this, that's why the few things they do make are sold near cost. I don't think their shareholders get this. Bezos has given people the impression that because they use a lot of tech they have a business like google or facebook. He doesn't point out they're just reselling everything and they have no pricing power. It doesn't matter how much or how fast he increases sales, resellers can't set prices so they'll never have dramatically better margins than WMT, much less google. Their innovation with AWS was to treat unused server capacity like a commodity that they could then turn around faster. It was always going to be hard for them to compete with companies like microsoft and google, who have 50%+ gross margins and can give away service just to figure out how to sell higher up the stack at a premium.

  • Orando Labs

    We (Orando Labs, have experienced Amazon’s competitive nature with our product EnduroSync. Compare EnduroSync with Amazon’s Coginto. Notice any similarities? Like the services, name, and even pricing? Cognito is a very scaled down version of EnduroSync, which is in line with #2 in your post.

    We think it’s a very dangerous move on their part. Instead of encouraging innovation on their platform, they are actively discouraging it. Do you think we intend to build anything else on AWS? Of course not. Anything we do on AWS just gives them more information. In fact, we are moving off of AWS. EnduroSync was using DynamoDb as a core component, but no more. Fortunately for us, the db layer in our products is abstracted, so it’s easy for us to move. Hello Google Compute Engine.

    • Pow. Great example. @joelklee:disqus – here’s one for you per #2.

  • I find the statement “AWS is not the best product at anything” misleading. The services that AWS provides as a technology and hosting platform are best-in-class. The Dropbox-like product I view as a peripheral product that Amazon built on top of AWS services, not necessarily a product of AWS itself (you don’t see this product anywhere in the AWS console).

    Take for example DynamoDB, S3, SQS, Redshift… these are all extremely powerful products that are better than, or at least on par with market equivalents.

    • Good point. The core IAAS services are still best in class. It’s the endless list of incremental products, which I think AWS views as features, which are mediocre at best.

  • janon

    Lots of hidden bias in here, sorry (not the least of which is that the VC community is pissed off that AMZN isn’t feeding them cash for quasi-useful to useless startups the way they’re used to). Maybe the pace of house upgrading in the Bay Area will have to slow a bit to every 14 months I guess.

    But the final conclusion is the most funny… So in each of these cardinal sin points above, the expectation is that >MSFT< will somehow be "better"? OK good luck with that. MSFT has a great history when it comes to delivering the best possible product at the lowest price and caring and feeding the partner ecosystem (and treating acquisition partners well), right? Oh. That's right. No, they don't.

    As for $200k a month… Every CIO I've ever met believes that until they are proven to be not tracking the vast majority of costs and pushed out by the CFO. If $200k/mth were the point where, for must customers, it was "cheaper to build", then cloud wouldn't be gaining momentum. Most say "$200k/mth and we can cure cancer!" and reality they are spending $500k/mth and barely providing the baseline they need.

    The startup business is a bit different, but honestly when I hear people saying things like "I can get a rack for $20k a YEAR!", that just tells me they have a demo and a dream and no real scale, SLA, production operational process, etc.

    • 1. I don’t live in the bay area so that’s not in the mix.
      2. I didn’t assert that Microsoft will do better. However, the opportunity exists because Amazon hasn’t built a wide-enough mote.
      3. Neither Amazon, Microsoft, or anyone else has a responsibility to buy anything from us. And we don’t ever invest with that expectation. But when they call with the notion that they are considering buying, and then suck the brains out of people’s heads, and then come out with a competitive offering, well, that doesn’t feel so good. It’s our responsibility however not to be long term suckers.
      4. $200k / month. I don’t know your background, but the cost curve on scaling for a fast growth startup is important. And the CIO doesn’t make the decision, often because there is no CIO in the mix – the last thing these companies need is a CIO (now a CTO is another matter – and the CTO is always in the mix.) Furthermore, if they can’t provide “the baseline” there is no business!
      5. “A rack for $20k a year” – I couldn’t give less of a shit about “a rack” – none of the companies that are scaling in my world need “a rack” – it’s a totally different discussion about infrastructure.

      • mbyrne

        I assume you’ve just read The Mote in God’s Eye…

        • One of the best scifi books ever.

          • mbyrne

            Yeah, just making a bad pun re “wide-enough mote” but I knew I’d get a response by dropping that book title!

    • In regards to MS, I would rank Azure better than Amazon for just about everything, with the exception that there’s no PostgreSQL option on Azure. Support is a BILLION times better on Azure than AWS. We have everything on AWS and I personally find it slow compared to lower specs on Azure which seem to be faster and cheaper. The Azure Add On’s are growing and has a good selection. Swapping between environments seamlessly is great.

      Stop living in the past with your MS hate.

  • joelklee


    Barb Darrow’s article is not my favorite in this round of hand wringing about Amazon’s quarterly earnings. Instead, I think Lydia Leong makes a much stronger argument in . Its worth the read IMO.

    To the specific points
    #1 Cost – At the moment Google sets the price but Amazon and Microsoft follow. So we have a troika of low cost providers. I expect prices will continue to fall by 50% every three years until physical infrastructure is “sold” at cost. I don’t see anyone on the horizon who can compete with these three on economies of scale – giving them a price advantage to competitors.

    #2 Interesting assertion – is there a list I can look at of the 32 services that are offered and the superior alternatives? I can think of a few (superior services (which are built on top of AWS)) but from a basket of services point of view, aws offers a compelling offering where there’s a non-trivial cost to building and maintaining a relationship with vendors. I would argue that their product suite on the whole is superior to any other offering on the market.

    #3 That’s not my experience but in fairness I don’t have to work with massive production workloads at this time. I’d be curious to know what Netflix thinks of AWS support. How would you characterize the type of companies that are dissatisfied? Young? Old? Big? Small? And is the support paid or not?

    #4 That’s a much longer conversation but I would argue that there are many application types where even at 200K a month it’s not going to be cheaper to build it yourself. An application might have global distribution requirements, a customer might need to burn that 200K on only one day a month. Depending on the requirements the TCO might not pencil out or an agile cloud model might always make more sense.

    I think it’s too early to worry about AWS. Instead I would welcome the arrival of Microsoft and perhaps google. Having a few dominant players (as opposed to one) is the sign of a healthy market.


    • Your last point is the one I unabashedly agree with! The competitive pressure will make everyone much stronger / better and will broadly benefit the entire ecosystem.

      • joelklee

        BTW – thanks for a fantastic discussion. I’ll be referring to all the content created here in future cloud strategy work.

  • Very interesting and thanks for posting. If companies want to build out their own infrastructure, do they contract directly to a company like Cisco-or would they try to use a company like


      If they don’t already have the skills in-house that are needed to complete the task. Then they must go outside the organization.

  • I think there’s a few things worth considering with the points you highlighted:

    1. Looking purely at the “compute per hour”, Google is obviously cheaper but you really have to factor everything in. You mentioned the “reserved instances” as a stopgap for Amazon but I think along with “spot instances” really a key feature for controlling costs. I doubt anyone is really running 25 ECs for 3 months and then 0 for 3 months. Buy some reserved instances, save some money.

    2. This is certainly true but it’s worth looking at the AWS ecosystem as a whole works together. Sure, everything is only “OK” and Amazon’s Elastic Transcoder is just a knock off of Zencoder but using that as a specific example, you aren’t charged for I/O from S3 to Elastic Transcoder for example.

    3. Agreed – the forum support is horrible but if you want a proper SLA at say Rackspace you’re paying big bucks anyway

    4. No argument there


    Anytime you compete on price your quality will eventually suffer!
    Remember “the cloud” was really suppose to be “we cloud” with everyone on the planet being able to integrate “a cloud” into “the cloud.” Also everyone wants to have their own dedicated hardware and software setup. There are many important reasons for this.
    When I designed my EHR system it was cloud based. You could pull down local apps that I called “rain drops”. No one seemed to understand how that was possible.

  • markmacauley

    I am having ongoing conversations with colleagues about this and have for the past few months. Your synopsis is a crisp way of presenting the question we keep coming back to – why hasn’t Amazon optimized its own cloud?

    I have 2-3 conversations a week with Fortune 50 companies about cloud vs. colo costs. Our findings are that at $50K/month you hit one inflection point, and the other is at $200K. Cost of capital, cost of operations, and risk are the three legs of the stool. Full disclosure – we move a lot of companies off of Amazon every year to other providers whose scorecard is 85% or better across almost a dozen dimensions that define ‘market’ for cost, performance and risk. Amazon may be popular, but so what?

    On a call last night with a streaming company, we asked and discussed among other things – if Amazon is the dominant player, why don’t they appear like it under the veneer of their brand? The initial pass was – if they (AMZN) are profitable enough then does it really matter, because profits are more important than the scorpions of ‘stickiness’ and ‘eyeballs’ of eras past.

    My personal take is that cloud – which ever flavor you prefer – is another road and iteration to the destination of computing as a true utility. Only it’s a smart utility where AI and predictive analytics are fused with electrons for an integrated high value digital human experience – whether you are a frog or a scorpion.

  • Paul

    Could you provide some links for point number 4? I’m not surprised that people are doing this analysis (they really really should be!), and even that some are making their analysis public, but I just haven’t bumped into any yet.

  • Our experience at both SugarSync and was that support actually was good.

    • That’s good! What support tier do you have? How much a month are you paying? I’d love to have data points that counter my experience.

  • Stefano Bellasio

    What I see from our is that more and more people are asking for Google and Azure learning materials and some of them are also looking for specific content on “how-to migrate”, comparisons and so on. AWS is a great product, the quality is not something we need to discuss, except for the support where I clearly agree with you.

    I spent 5 years in the “dedicated server / rack / private data center” world, and I still don’t see how some companies can build everything on their own when it comes to something like DynamoDB or SQS and so on.

    BTW, great post and conversation. Thank you


    • You’d be surprised. For companies that have this built into their infrastructure (as many of the companies in our Glue and Adhesive themes do), they actually need dramatically more flexibility and power than AWS services like DynamoDB and SQS provide.

      • alimoeeny

        With all due respect Sir, I think this comes from not understanding the weaknesses and strengths of dynamodb and sqs. You need to set aside the older way of doing thing to be able to benefit from these. If you want to bend them to the older ways, they will resist, and you will get tired.

        • RBC

          Ah snapppp – challenging bfeld to get technical. You’re about to set it off!!!!

          But seriously great post, and great comments. You have to love the incumbent advantage, you really need to p*ss people off as the infrastructure, training, existing skills, are all in your favor. That being said it seems there has been an exponential increase in free online education such as cloudacademy and stackoverflow that reduces some (but not all) of that advantage.

  • A provocative article was on TechCrunch a week ago on the broader topic of data centers. After 8 yrs. in stealth, MaidSafe is betting AWS and its brethren will be disrupted via a new peer-to-peer approach to the Internet.

    I have neither the expertise to judge their approach’s credibility nor a dog in the fight, but it’s certainly interesting. If it goes, AWS will have far bigger problems.

    How credible is their concept?

    • No real clue yet on my part.

      • Not an internet creepo

        I read the post about MaidSafe and it appears to be addressing the whole eves-dropping and data collection issues of today’s internet.
        However the problem is not the internet. It’s the people who use the information to harm others that are the problem. It’s always people who are the problem. A stick can be sharpened and used to stab someone. But the stick isn’t the problem. The person who does the stabbing is the problem! This whole topic is fought over again and again in the gun control debate.

  • alimoeeny

    I have used AWS, for a while and even their free twitter support is very good, don’t know what you are talking about in #3

  • williamhertling

    I see your points, and I think they have truth to them, but they don’t feel like the whole truth.

    We’ve used AWS and two other cloud vendors. My experience is that AWS provides pretty good tools for a large number of different needs. Other vendors either provided worse tools or didn’t provide the tools we needed.

    We don’t expect AWS to be the cheapest, but cheapness comes at a cost when it requires having the right Ops people to build yourself what you get with AWS. Good ops talent is hard to get. Especially for larger companies with inflexible HR policies, it can be almost impossible to hire what really talented Ops people demand. By providing pretty good solutions across a very large solution space, I think AWS decreases the Ops burden placed on a company.

    So yes, they aren’t the cheapest, or the best at everything, and at a certain point, if you can get the right talent, you need to invest in building it yourself. But they are a pretty good one-stop shop that’s going to let a lot of companies accomplish what they need at a fast pace and with less Ops investment.

    • I’ve gotten some feedback that I fumbled part of the message, which is that Amazon will behave in a way that is it’s nature. One of the elements of this is to be the low cost provider. And when they aren’t, that causes interesting / complicated behavior.


        What do you mean by interesting/complicated behavior? Competing on price has been done many times by many different companies over the years. It almost always causes a loss of quality! If I remember correctly. Competing on price was always suppose to be a short-term approach to gaining market share. It wasn’t suppose to be part of the long-term business model.

        • Actually, competing on price is one of the fundamental strategy options. Michael Porter first really nailed / clarified it in the 1980s in his book Competitive Strategies and then built on it in his book Competitive Advantage.


            Cool. Thanks. I’ll check into it. With the way the price based giants are always needing to put their emps on assistance or needing some kind of gov help. Do you think the price based business model is still good as a long term strategy?

          • Anyone who needs government help is playing a losing game.

  • EllenO

    Brad, All good points though I’d argue that the AWS API being the de facto cloud standard could outweigh the issues you raise. As you are aware from time to time software technology/platforms emerge as informal software standards that are part of an entire software ecosystems – software developers, books, third party tool, developer certification and so on. And these standards continue even when better and/or lower cost alternatives exist because the cost of switching is too high. Software development costs are still the major cost for a software or Internet company. Porting applications is expensive as is the cost of finding and employing developers who understand a non-standard cloud platform.

    Clearly AWS is such an ecosystem or platform. Currently there are attempts to develop open source software alternatives compatible with the AWS API, see for example Eucalyptus, but as yet there is still nothing that even approaches the breadth of the AWS offering.

    My guess is that despite your valid points AWS will thunder on gaining users as the ecosystem broadens.

    For the record I have had good (though not stellar) AWS support from Amazon.

    • There is no doubt in my mind that Amazon will continue to be successful here in terms of overall market share. The deeper question is whether they will be able to maintain a clear leadership position and, over time, actually generate net income (or even free cash flow) from AWS.

  • michael

    I would argue that point #4 is a little high – I have seen companies that have cost savings well below the $200k/mo mark – even a modest business at $5k – $20k/mo expenses on Amazon can have huge cost savings over 3 years with buying their own servers, sure it is a larger outlay and some maintenance, but overall you get more performance and more reliability (if you have good operations).

    Overall, AWS is a great at allowing someone to start a project quickly and then move to physical servers or other providers. Few companies take the advantage of spinning up and down servers at AWS as they need, more often they just keep them running. It is also a great ad hoc developer environment, if you have automated all your deploys and environments.

    AWS is not a low cost provider, it just provides convenience, which can be invaluable at times.

  • James Cole

    You forgot one thing… they are among the absolute worst at partnering.

    • Dmitri

      Any facts why?

  • Dmitri

    Using as my Backend, extremely generous free package, great response via Forum, why would I ever need to use AWS?

    • Frank

      But you are — Parse is built on AWS 🙂

      • Dmitri

        That’s different – I only deal with Parse, so don’t see anything from AWS including their enormous documentation 🙂

  • Ilja

    Interesting article.
    #1: There are providers with lower cost in some scenarios to be sure. However for things like Redshift it is beyond most companies to get anything near like the price. Additionally a lot internal calculations about we can do this cheaper simply ignore the engineering tradeoff of using your good guys to build something that actually differentiates.

    #2 Point is being good enough so you can choose to use the service provided. You can still roll your own. Good enough solutions again enable you to use your smartest engineers in something not provided by cloud providers. A lot of companies I know that run in AWS do not use all the services but roll their own for flexibility and features.

    #3 unpaid support is not comparable to the paid one for sure.

    #4 Optical illusion. If you run an single app in large scale and also completely ignore the non-zero exit cost risk (and the misplaced engineering effort) yes you might be able to save money. There are also some specific use cases that don’t fit the AWS pricing model at high performance and short run testing etc. However majority of the so called enterprise IT will never be able implement the internal stuff anywhere near the efficiency as the single app guys. LAN networking in AWS is about 30X cheaper to operate than any large corporate IT network (hence the SDN rush). Point here that for non internet sector opex dominates capex.

    So the dilemma is for AWS is that for the large internet properties such as Dropbox the cost alternative to migrate to other cloud or on premise is significantly less than enterprise customer. So these guys can get the best price from cloud providers and they spend a lot of money on AWS. Cloud providers could choose to ‘buy’ AWS customers.

    For the enterprise IT market the picture is totally different. For competition, right now looks like only Azure and GCE have the resources to catch up with AWS but it is early days.

  • AWS is not the low price provider.
    Eh. Not sure why this is relevant and also not sure it’s true for what you are getting… It’s like saying “there are books cheaper than that book you just bought.” Well sure there are, but do they have the information I want in them?

    AWS is not the best product at anything – most of their features are mediocre knock offs of other products.
    This misses the point – their features are SIMPLER knockoffs of other products. That’s why it’s an accelerator. Dropbox and Salesforce and all the successful cloud entities have said “you know, some enterprise user left to their own devices is going to generate a list of 1000 requirements they don’t really need. Forget that. Let’s make the actual core functionality they need and leave off the rest so it’ll actually get used.” This is why they dominate the IaaS business. Many of their products are named to match. “SIMPLE email service.” “SIMPLE queue service.” “SIMPLE notification service.” This drives a new wave of architectural thought – instead of complicated services packed with stuff, what if instead I integrate simple, well-designed microservices? After doing a lot of cloud architecture work, those attributes are positives, not negatives.

    AWS is unbelievably lousy at support.
    I’m not sure I’d want to be in a race with Amazon, Microsoft, and Google to see who supports customers worse. I’m not sure I’ve ever been part of an enterprise happy with its Google support, and all experiences I’ve had with Microsoft support have been some Brazil-esque “you can’t actually ask them questions, only some VP is a designated contact on the corporate contract…”. Amazon is positioning themselves more like a hardware vendor, you don’t bother getting much support from them besides parts replacement, you get support from the managed hosting provider or whatnot that’s a MSP on top of them if you need it.

    Once you are at $200k / month of spend, it’s cheaper and much more effective to build your own infrastructure.
    This is frequently untrue and based on people not understanding the full costs of getting stuck in the infrastructure business. What’s your cost of delay? Average enterprise “wait for servers” time is about 6 weeks; assuming you’re not just using them for nothing, your ROI is delayed by that amount. And what about all the operation of those complex systems? You can’t just stick in the salary of the developers and sysadmins you’d need – stick in your revenue per employee instead, because that headcount could be doing something useful for your company instead of plumbing. Not to mention the cascading percentage of each layer of management’s time spent worrying ab out the plumbing and the plumbers instead of conducting the core business of the company. Cost of delay from lost agility and opportunity cost are never taken into account but definitely should be.