Why Many Private Cloud Projects are Failing

private cloud failWhile survey after survey show that enterprises claim to prefer private cloud models, IT is struggling to deliver.  Off the record many in the IT consulting and SI business will say the majority of private cloud pilots they’re seeing aren’t going  well.  In fact some believe that up to 90% of private cloud pilots in the enterprise will fail to move forward.

In most cases the reasons these pilots are failing has little to do with technology.  This isn’t to say that things like installation and setup can’t be improved for OpenStack and other platforms – they can.  But the major factors that are causing private cloud projects to sputter and die have more to do with other, non-technical factors.  Here are four reasons:

  1. Success metrics – most private cloud pilots and proof-of-concepts are focused almost exclusively on technical validation. While the IT team proves to itself that it can deploy a private cloud, users and developers are measuring it against services already available in the market.  Failing to demonstrate a better services value proposition isn’t compelling, and certainly won’t build internal momentum.  And does the organization really need an existence proof that IaaS environments can be successfully deployed?  Pilots need to include business success metrics as well as technical.
  2. Inertia – so why aren’t the metrics defined correctly?  Because in many cases enterprise IT just sees cloud in a fundamentally different light.  As noted in posts by James Staten at Forrester and Giri Fox at Rightscale, enterprise IT tends to see cloud as the next logical step in the evolution of infrastructure and virtualization.  The focus is on demonstrating robustness, stability and change control, the enterprise IT comfort zone.  The factors that business users really care about like usability and provisioning time are left to the wayside.
  3. Lack of vision – this inertia exists because CIOs are failing to answer the basic long-term question of where they’re best positioned to compete for business budget dollars, and whether it really makes sense for IT to be in the IaaS business.  Without a well articulated ITaaS vision, it’s hard for business users and developers to get excited about internally provided services they believe they can get from Amazon, Rackspace, Google or Microsoft today.  While the new default answer is that IT’s vision is to provide hybrid or “multi-cloud” environments, few can actually articulate what that means and why business users should care.
  4. Motivation and commitment – so what are we left with?  In many cases private clouds seem like a good way for IT to come up the cloud learning curve while leveraging existing hardware and infrastructure.  For others, part of the reason is to show that “we’re doing something with cloud”.  Regardless, in some cases enterprise IT isn’t yet fully bought in to the paradigm shift being brought about by cloud, and seem to view private cloud projects more as science projects.

Net, net many CIOs still haven’t fully grasped the fact that their private cloud will basically be competing in the market against AWS and others.  While security, compliance and performance factors will enable IT in some cases to maintain their internal “monopoly” through mandated use of private clouds, the clock will be ticking…

UPDATE: Several readers have objected to the statement that “some believe that up to 90% of private cloud pilots in the enterprise will fail to move forward.”  These readers believe that because the statistic isn’t cited, then it must not be real.  Please note that we did in fact have a conversation with an executive at a leading IT consulting services provider that stated exactly that figure.  As the conversation was off the record, we cannot provide a direct citation.  The observation was also consistent with other conversations we’ve had in the enterprise cloud market, and while the numbers they stated may not be 90%, they say the failure rate is still high.  As enterprise IT project failure rates tend to be fairly high in general, it probably shouldn’t be shocking if it’s also high for a new technology platform.


  1. I would also say that many Public cloud offerings fail as well. They fail in the sales and delivery to the end client, as many of these “cloud Providers” sales pitch is we want to move your business to the cloud. These providers don’t even understand what the clients environments look like or what their needs are, they just assume that the cloud is right for everyone and that is truly not the case.

  2. I disagree that private clouds will compete with Amazon as the notion of hybrids makes plenty of sense, especially for companies that have infrastructure that simply will never make it into a public cloud. It is not an either/or but quite possibly a both/and.

    Use of Eucalyptus, OpenStack, CloudStack, et al, can benefit some private cloud instantiations because those that share common APIs, e.g. Eucalyptus does with AWS, can make hybrid-ing clouds straightforward.

    I will agree that deploying clouds requires IT leadership to engage beyond a maintenance role. Thinking that Amazon is the same as locally hosted except cheaper misses out on the tremendous value that can be gotten from clouds … of course, companies need to rethink their business and technology architectures to realize the grander goals …

    • Tom – good points, though I do I think that CIOs need to be aware that the value propositions for their private clouds will need to be competitive with public clouds, especially for developers. Private clouds may in fact be the appropriate answer for an organization, especially if security or compliance issues apply. But developers who can spin a vm up in 15 minutes (or less) in AWS will need to be made aware of why a private cloud environment will be more attractive for them. At the end of the day I think developers care less about how the infrastructure service is delivered (private vs public or hybrid), and more about the characteristics of the service itself.

  3. Emma Huang says:

    I think companies are planning or already applying cloud technique should take supply chain perspective into consideration. A clear vision with measurable metrics is essential and should be visible to the users of the cloud technique, not just the IT leadership or the IT team. Also, the transition from the incumbent system to cloud would inevitably create some hidden cost that should be identified, defined and monitored with caution.

  4. The 90% failure figure does seem high – though I wonder what percentage of cloud projects are considered pilots? Cloud computing in certain segments are fairly matured and well-defined, and it’s difficult to imagine enterprises these days are willing to undertake more pilot projects than not.

    • Sam – I would agree that segments like SaaS are more mature and accepted in the enterprise today than IaaS or PaaS. The figure in the post refers purely to private cloud IaaS pilots and projects…

  5. There are many technologies that can be deployed for cloud.
    IT professionals seem to have no issues in creatively developing solutions that for the most part work.
    However , the basic of business questions that i have visibility to are;
    What is the risk to my data?
    Will I have the same speed, performance, access and security that my end user enjoys right now? Or better whilst saving money?
    will the service provider solution meet my RTO and my RPO requirements as required by my business?
    Where is my data physically located and what happens if I cannot get access to it? Ie. is the data going to be stored in SE Asia – the USA – EMEA and then governed by the local rules within that country.
    As a business in Australia. How do I deal with that if something goes wrong.
    Then of course there is the myriad of questions on commercials, risk, legal and the unknown issues that are yet to be realized as organizations learn by trial and error.
    IMO – Most of these issues will become hot topics of discussion, and over time will attract enough attention and thinking to resolve.
    Exciting times ahead.

  6. It is an interesting article, and certainly provides food for thought. As with any new technology, extreme caution should be taken when considering moving to a cloud environment.
    Being in the Health IT space, a private cloud holds tremendous benefits in the market that I understand may not be the case in other sectors. I believe each business sector must clearly define its individual needs and expectations, and have a compelling business case before moving to the cloud.

  7. Most of the organizations are not yet ready for Private Cloud as IT organizations themselves are unclear on the ROI and benefits of private , they ask for pilots and don’t even use private cloud as a part of the large ITO deals, mainly because either their existing data center offerings are good enough and optimization and virtualization gives them required benefits. Most of the companies are now asking for more efficient data center with their Apps running on them OR will have direct access to Public cloud offerings as well.. Private cloud though advertised mainly by Hardware vendors like EMC, Microsoft , Cisco and NetApp to sell their properierty hardware.. all of them are happy selling their data center gear goal..

  8. Moving to a cloud environment when performed correctly should be painless and of course does require the correct team of experts. There are many other variables to consider when making a generalization about cloud projects failing. Furthermore, if this information is based on a single source…

    Cloud configurations and migrations should be easier than ever, and has taken off greatly in the last couple of months. Most experts will agree you should stay away from any proprietary strategies, which may cause you problems in the future when you need to move your data.

  9. Cloud is nothing more that a resurrection of offline storage that existed before the internet. That failed and so will the so called cloud.
    To me it is one BIG con. Buy an external drive, backup your files and take then home. Cheaper, Safer and more efficient

Speak Your Mind