The Costs of the Cloud: Double-Check Me on This, Would You?

by Andrew McAfee on October 5, 2012

For the past few months, I’ve been working with Google and Analysis Group on a project to understand the comparative costs of cloud vs. on-premise computing. As I wrote in Harvard Business Review last fall, I believe that the cloud will eventually become the status quo for most types of corporate computing, in part because of cost advantages stemming from massive scale efficiencies.

But that doesn’t mean that the cloud is cheaper now, or that it’s currently the cost-effective choice for young and small companies. We wanted to find out if it was, so we built a relatively straightforward model of the computing costs facing different such companies, and used it to project cloud vs. on-premise costs under various assumptions and scenarios.

We don’t know of any other model that’s built from the bottom up like this one is, and that tries to be comprehensive — to incorporate all significant IT cost categories. We don’t try to model any productivity or revenue increases as a result of moving to the cloud here; we’re just after comparative costs.

Now it’s time to subject that model to some good old-fashioned peer review. I’d really appreciate it if you and/or the sharp-eyed geeks you know would take a look at our model and let us know if we’ve missed anything important, or if you find any of our parameters or assumptions hugely off base.

The model itself is here. Please feel free to play with it, examine its inputs and assumptions, and poke at it in any and all ways you’d like. The ‘Start Page‘ tab explains a bit about how it works, the ‘Summary‘ tab lets you enter your own parameters and compare total Cloud vs. on-premise costs, and other tabs contain further explanations, cost assumptions, and so on.

A form to collect your feedback is here; please use it to tell us what you think, and how the model can be improved (you can also leave a comment here, of course).

Thanks in advance for taking a look and letting us know what you think.

  • http://blogstu.wordpress.com stu

    Hi Andy – I’m an analyst at Wikibon (an open source research firm, HQ in Massachusetts) and we did a pubic vs private analysis back in 2010 that you might find useful. Here’s the article: http://wikibon.org/wiki/v/Private_Cloud_is_more_Cost_Effective_than_Public_Cloud_for_Organizations_over_$1B and we also published this in a fun infographic: http://wikibon.org/blog/private-cloud-computing/
    You’re welcome to leverage our figures.
    Cheers,
    Stu Miniman
    Twitter: @nohype:disqus

  • Mike Burroughs

    This is one of the most complete models I have seen but, for a very small company like the one that I manage, this model doesn’t tell me much about the impact on cash flow. Many small businesses are far more impacted by cash decisions than by cost decisions. A $4000 server may be cheaper in the long run than a cloud based solution but I either have to commit $4000 up front to its purchase or incur financing costs (and obligations) in order to manage available cash. I wouldn’t say this is a knock on the model as presented as it is amazing complete and powerful, but it would be interesting to see how the inputs to this model could be used to develop a cash flow impact summary.

  • http://twitter.com/JFTAXI John Fitzpatrick

    As Mike Burroughs says, the cashflow considerations can usually outweigh other beneficial parameters. E.G. a server requires the inputs of an IT expert, which may or may not be available in house, but either way ends up influencing decisions that are impacted by available cash. Depending on the perspective you are coming from, this exercise can be positive or negative and have similar impacts on the viability of the business. The exercise is coming from the perspective that the business is adequately funded and thus would not reflect reality for the vast majority of start ups.

    On the other hand if one was adopting a culture of developing enterprise, then clearly some form of IT management systems can only improve inefficiencies providing the scale of the business is sufficiently high to warrant such inputs.

    Thus smaller companies could not absorb the costs of either cloud or on site IT management systems.

    I don,t mean to be negative, but on my side of the world the scale of indigenous business development is microscopic compared to the figures used in the exercise, nevertheless I bleilev that with state provided training all companies can benefit from the Cloud shared services, but only over the long term.

    John Fitzpatrick
    Dublin,Ireland.

  • Alex

    This is a great article because cloud computing is begging to become extremely popular. I have just recently began using cloud based app’s which have been really help. I have also been using Hojoki which really saves me a lot of time by having all my cloud app’s in one feed, its really cool. I’m looking forward to more articles about cloud computing.

  • kaili

    On the other hand if one was adopting a culture of developing
    enterpriseworld of warcraft gold, then clearly some form of IT management systems can only
    improve buy world of warcraft gold
    inefficiencies providing the scale of the business is
    sufficiently high to warrant such inputs.

  • Faiyaz Haider

    Excellent work, one of the most detailed calculations I’ve seen so far. One point – I do see the Sq.Ft cost for the facility and electricity costs as well. Is data center cooling cost included?

  • Janice Abel

    It obviously looks like cloud computing is gaining more and more positions. I thinks that it really is a solution for small and medium-sized businesses and I agree that cloud hosted project management software can be a relief for many of those seeking for a way to increase their efficiency.

Previous post:

Next post: