Blog post

Problems Encountered by 95% of Private Clouds

By Tom Bittman | February 05, 2015 | 17 Comments

private cloudFuture of InfrastructureCloud

In my last blog post, I identified ten reasons private clouds were failing. I consolidated that list to six items (below), and polled attendees at Gartner’s Datacenter Conference in Las Vegas in December. I asked the question “What is going wrong with your private cloud?” I was a little surprised that 95% of the 140 respondents (who had private clouds in place) said something was wrong with their private cloud.

Priv Vendors 2011 2014

  1. Focusing on the wrong benefits: Internal, bottom-line, or not putting the right metrics in place. (Usually, this is focusing on cost-savings, not agility).
  2. Doing too little: Is this really cloud? Or just virtualization? And what about the stuff running inside the VMs?
  3. Defending I&O — and doing too much: Optimizing for everything means optimizing for nothing.
  4. Failure to change the operational model: Agile clouds need agile processes — and people are your biggest supporters, or your biggest roadblocks.
  5. Failure to change the funding model: When you build a drive-thru service model, you better get paid first.
  6. Using the wrong technologies: What’s tactically right might be strategically wrong.

The Gartner Blog Network provides an opportunity for Gartner analysts to test ideas and move research forward. Because the content posted by Gartner analysts on this site does not undergo our standard editorial review, all comments or opinions expressed hereunder are those of the individual contributors and do not represent the views of Gartner, Inc. or its management.

Leave a Comment


  • I would like to see how many of these reasons can be adopted to public cloud scenarios.

    “Doing too little” is quite interesting. Maybe we should start with two questions: “What is private cloud?” and “Does private cloud exsist?”

  • Tom Bittman says:

    Miroslaw, on the “what is” question, the answer is it is the cloud computing style delivered with isolation. Fully private would be fully isolated. It doesn’t need to be owned and managed on-premises, but today it often is (I’d say, 90-95% of the time). A growing percentage is moving to third-party offerings, where isolation is a variable – sometimes fully isolated (an outsourced private cloud), sometimes “virtually” isolated. Now, does this exist? Of course it does, because the cloud style should not only be applied where security and compliance requirements allow sharing. The better question is – does it need to be owned and managed on-premises? Until third-party offerings were available, it was the only choice. Also note that many things described as private cloud aren’t private, or aren’t cloud (which goes to “doing too little”). The reality is that cloudiness is gray, a spectrum, not absolutes, and the same is true with privacy. Choices need to be made not based on definitions or getting checkmarks, but in the value needed and derived.

  • PC is having trouble differentiating itself from plane old virtualization or VDC. That is mostly because vendors are cloud washing every product that has a hypervisor embedded in it.

  • HaroldV says:

    Private clouds is the tool to perform your business. Every design is different in how they are implemented. However, most have similar conglomerate use of technologies: Cisco for networks, Microsoft for servers, VMware for virtualization, etc. How they are uniquely applied is the secret sauce made by your IT team.
    I’d like to add another category to the mix as to why they fail: Using technology incorrectly. Technology needs to be properly aligned with the business needs, and a tool to provide value-add. We as technology implementers need to understand the business to apply technology efficiently and build the private clouds that support the business processes, and forecast future growth supporting new capabilities. Of course, having an understanding of frameworks such as Project Management, ITIL, COBIT, Compliance, and Security helps developing an infrastructure with best practices the first time around..

  • Cloud != saving money, Cloud = saving time and increasing the production value of IT…the latter outcomes can often save/generate money or real ROI. Neither can be achieved without appropriate attention to service-oriented delivery and automation. Without these facets, you may just end up with well-coordinated infrastructure, not “Cloud”.

  • Chuck Hollis says:

    Hi Thomas

    I’m seriously confused here.

    The title of your post is “Why Are 95% of Public Cloud Failing?”. But the survey asks respondents to share what is wrong with their private clouds.

    “Failing” is not equal to “has one or more problems”. There certainly can be problems with an initiative, but that doesn’t necessarily mean it is a failure.

    The evidence presented does not support the title of the post.


  • Hi Mr Bittman

    Firstly, I will advice you to take a research and statistics courses

    Because your questioning style is simply adversely effecting subjects. To see this point, I offer you to take this poll with the following question

    “What is going correct/successful with your private cloud”

    I am sure you will get (much) more than 100 percent change compared to this one.

    Moreover, it is wrong to arrive failure result. Failing is something addressing total picture. But respondents might simply indicate efficency/usage/performance problems.

  • Jim says:

    Not without value, but I also have to agree with Chuck and Mehmet about the misleading title.

  • Nicloadianis Kurolas says:

    Hello all
    I have noticed a lot of technicalities above. But the reason we are all missleaded with words like “Failing” is pure marketing and commercial, covered by so called technical reasons.

    There is no failure in Private clouds as there is no failure in the Public clouds. Both DELIVERY model for IT services produced in a Data Center have pro and cons. Challenges are on both sides and are more related to the type and scope of each company business operations.
    Across the world there are regulatory authorities that enforce utilization of the private cloud solutions due to security reasons.
    The trick resides in the company itself how the business operations might run more efficiency supported by IT services and solutions that can be delivered from a mixture of Public, Private or Hybrid Cloud.

    Looking forward …. 🙂

  • David Moore says:

    +1 for misleading title

  • Tom Bittman says:

    Peace on the title. In my defense, clients that I’m talking to have been self-describing how their private clouds are failing to meet their goals, failing to meet requirements, failing in comparison to public clouds on price or speed. We’ve had the same discussions in workshops, where attendees tell us how much they are struggling to solve these problems, or how inappropriate goals started them in the wrong direction.

    I suppose a more accurate title would have been “only 5% of private clouds are problem-free.”

    But perhaps this new title is a little more to the point – “Problems Encountered by 95% of Private Clouds.”

    I do appreciate the feedback.

  • Heath Muchmore says:

    Tom, I can completely understand why customers are saying this. In the old aspect of people, process and technology, customers are still buying technology to compensate for the lack of process and people improvement. You can throw ITIL, CMMI, Agile, TOGAF or any other acronym into the mix and it does not fix anything if the people do not believe in it. Private cloud implementation is no different than all the failures over time in the ERP and CRM world. Per Mr Hollis, were they complete failures? No. But to your point Tom, they certainly failed to deliver on many of the promises and the customer perception of that is the ultimate deciding factor. Per VMware, I won’t reveal one of the most monumental failures in private cloud, that I know of. Why was it a failure? It certainly was not VMware’s fault, as the technology did exactly what it was supposed to do. The problem was the fact that the customer never changed how they operated. This caused a lot of internal problems between the old guard and the new cloud team, which lead to limited use of the private cloud by the lines of business for which it was built.

    So for me, I whole heartedly can believe this survey to ring true.

  • Tom –

    Great piece – and good insight. I actually interpreted your data a little differently… After years of being in the enterprise cloud space – on the big and small vendor side, and now at an SP – I think there are certain functions that a private cloud simply cannot perform….

    But that’s too much to write up here – so, I wrote a reply at our blog. I’d welcome your feedback!


  • Tom,
    Putting aside the semantics of success vs. failure, I think you are right on the money. Without getting too philosophical, I believe the root cause is the common misconception that cloud is a technology trend and as such requires technology to address the changes that it brings. While technology is definitely a key ingredient, it is not the main one. Cloud is a delivery model. Really, it should be called “everything as a service”, but that’s not catchy :), so we invented cloud.
    The same way that digital fundamentally changed the CME market, cloud is changing the IT market. And what we are seeing in IT parallels what has happened and continues to happen in the CME market. The incumbents look at it as a threat and are playing defense, rather than look at it as an opportunity, try and understand what customers want and build to for them. So there are a lot of knee jerk reactions, “build it and they will come” or slapping “regulations” on what can and can’t be done.
    IT needs to stop thinking like a tech provider and start behaving like a business provider. So yes – it means having product managers, sales & marketing functions, doing service/product design, etc. Because that is what the competition is doing. In particular, and where the biggest challenges will lie, is in changing the economic model of IT from cost to value center. This means taking performance management seriously, manage by KPIs, etc. It also means that – on the flip side – LOBs need to understand that while running around IT to the public cloud looks very attractive right now, the long term implications are such that it is in everyone’s interest to keep IT as a preferred provider.

  • PiroNet says:

    @Heath and @Miron you’re absolutely spot on!

    Cloud Computing is NOT a technology per se but a business opportunity to achieve cost reduction, better quality of service and increase Business agility.

    Also for any project out there you can’t escape the mantra that is People and Process first and Technology eventually!

    I have witnessed those failures and I have summarised them in a blog post few months ago ->

  • Tom Bittman says:

    The problem with a blog post like this is it is out of context – out of context of the presentation in which the poll was given (“Building Successful Private and Hybrid Cloud Services”), and out of context of the research I deliver to clients. I agree, on its own it sounds like I’m saying the sky has fallen. In reality, I use this information to help focus my advice to clients. For example, I published a research note on the subject to my clients that gave 26 specific action items to avoid or mitigate these problems. No, I don’t publish that to my blog.

    But still, we shouldn’t ignore the implications of these problems. To use an analogy of a house with “problems”, my clients aren’t complaining about minor issues with their houses. They’re saying the builder forgot to put a door on the house, or a foundation, there are no bathrooms, no one will buy the house, the house was built in a swamp – very basic, “foundational” problems. These are conversations I’m having every day. In my research on private cloud for nearly seven years, I’ve consistently told clients that cloud computing was primarily a process, cultural, business relationship/model, people transformation. You can’t buy or install cloud in a box and call it good. The majority of problems being reported by my clients are day zero, transformational issues.

    I spend no more than one hour on each blog post, skimming from my research, and without any actionable advice. This post in particular has drawn a lot of attention. I realize that maybe a little more context will help avoid being misinterpreted, and I’ll do more next time. I hope this gives more context on this one in particular.