3 Comments

Summary:

Here’s five attributes for you to remember that make up the cloud — and an acronym to boot!

Joe Weinman, VP, Cloud Services and Strategy, Telx Structure Europe 2013
photo: Anna Gordon/GigaOM


Session Name: Is Cloud A Technology, An Operations Model Or A Business Strategy?

Chris Albrecht
Joe Weinman

Chris Albrecht 00:02

Thank you Joe. Al right, it is with great pleasure that I bring out next guess. You might remember him from such launch pad as yesterday where he was a judge. He is Mr. Joe Weinman. He is an SVP with Telx and author of the book Cloudonomics and he’s going to be speaking on “Is Cloud Computing a Technology, an Operations Model, or a Business strategy?” Please welcome Joe Weinman to the stage.

Joe Weinman 00:27

Good morning everyone. Okay, I only have about an hour so I’m going to try and speed through this relatively quickly, but what I want to cover is actually some thoughts I’ve had around Cloud Computing Economics and kind of implications around business models and cost structures and strategy. And so what I’m going to do is I’m going to start of pretty easy then I’m going to get like irritatingly nasty in terms of some equations and then I’m going to pop it back up to a strategy level so just you have sense of the dynamics of this morning’s flow. By law, you’ve got to start with a [niece?] definition which you’ve all seen 12,000 times and I’m not going to go through this, I just want to point out that it’s actually a great definition, we have the co-author of it right here. It’s just a little too long and in fact it’s so long that it won’t even fit on one slide as you can see. So what I did is I reinvented the definition and to me a Cloud, just looked at abstractly, has five major characteristics.

Joe Weinman 01:30

First of all, it’s common in the sense of common to a variety of customers where those customers have workloads that are statistically multiplexed dynamically into a shared pool of resources. Secondly, it’s location independent. It shouldn’t matter where you are or where the service is actually delivered from as long as response time objectives were met. Thirdly, it’s online. In other words, the network is essential or as necessary as broad network access, is critical to making it all happen. Fourthly, we can call it a utility. In other words, paper use pricing or usage sensitive pricing or paper use whatever you want to call it. And lastly, it provides on demand resources whether shouldn’t there be a delay between a variation on the workload and the fulfillment of those resources. Now there’s our five key attributes, you can tell they spell a handy acronym and the one comment I want to make about this now is that the reverse characteristics are also possible and there’s nothing inherently better about Cloud except for certain types of workloads and certain types of business context. Just as one example, we can think about flat rate pricing actually being very beneficial, that’s the whole model of saying medical insurance where you translate variable potentially catastrophic medical expenses into a flat rate premium.

Joe Weinman 02:56

This as a business model has nothing to do with technology perse and so for example a London taxi cab obviously is a paper use measured service type of service. Another good example is a hotel which offers again paper use on demand resources namely rooms. And then even unusual context like banks offer the same type of thing. Their pool is their capital base and the paper use rate is called an interest rate. Now there is a little bit of a basis in technology. For example, you can see in the taxi cab, the driver has been loaded into the physical hardware and the case of the hotel, you’ve got one server on the right running two instances of Java in the system tray whereas the other one is dedicated to housekeeping task and on the right, you can see the multi-level cache, but I don’t want to get into the technology underpinning of that. Thank you for those of you that are awake enough to laugh at those jokes.

Joe Weinman 03:55

With those five attributes, this is the part that you’re not really going to like the story in the morning but I just want to point out that there’s actually pretty rigorous quantification that can be provided behind each of these attributes and there are implications of this quantification. Quick example, if you define a penalty cost function based upon either over capacity that’s extra and unnecessary and under-utilized or under capacity and therefore insufficient resources to meet customer demand, it’s revenue impacting. That 1 over square of mwa indicates that it actually doesn’t take very many workloads to reach a pretty close to theoretically optimum penalty cost function level and therefore, it means that there’s plenty of rules for mid-sized providers. Location independence basically shows that there are diminishing returns to building out service nodes which says that for most applications depending on their degree of interactivity, it actually make sense to have probably about a few dozens service nodes laid out across the planet. More than that is mostly overkill although there’s also network-caused implications not just latency but I’m not going to get into that. The only point is that you can in fact do all this all kind of analysis that I have. So that’s sort of more the straight forward cost quantitative analysis and it’s great as far as it goes but what I want to share with you is some logic around the strategic value of both digital native cloud-centric business models as well as how they can actually apply to virtually every industry.

Joe Weinman 05:41

Okay. So, there’s a famous book out there called ‘Does IT Matter?’. It’s sort of a business bestseller and for those of you who haven’t seen it, I’ll give you the executive summary of the book. ‘No’. And Carr’s argument is actually very articulate and insightful. I just don’t agree with it but the argument is because IT has become ubiquitous that makes IT a commodity and even it’s commodity is non-strategic. Anything that is non-strategic by definition, the correct investment approach is basically to minimize cost, right? You don’t get gold-plated paper clips to bind your printed documents, if anyone uses those anymore, because there’s no benefit to that so minimizing spend would be appropriate. Now, the counter example to that is this server rack, does anyone know who this is? Or was? Yes. This was Google before it was Google when it was the back rub project at Stanford and the point is as of this morning at least, Google’s market cap is back over 300 billion and the question is ‘If IT doesn’t matter how exactly did that happen?’ Google had no first mover advantage, had no brand at that time, had no preferential access to capital, had no preferential access to resources. They were the number 17th search provider if you remember back to Altavista and Lycos and God knows what else has been in the midst of time. In Rover, this IT-based advantage both in terms of a superior algorithm implemented in IT as well as superior execution of IT at skill has proven to be highly sustainable. So if you look at Google’s search market share, it’s basically health study exactly two thirds of the market. So the– What I would submit to you is, if IT doesn’t matter, how exactly do you explain the success and this is just one example, we can think of others. At the other extreme, you’ve got things like Encyclopedia Britannica relative to Wikipedia. These are logs scales shown in the chart so as you can tell and I’m sure you know, sad to say because I guess it say local product but Encyclopedia Britannica had exponential decline while Wikipedia was rising exponentially, again, I would claim that that was superiority in IT. And I don’t need to go through all these, it’s pretty clear that this isn’t restricted just to search and advertising engines in order to encyclopedias, every dimension of our society is as Mary Meeker says, ‘Being re-imagined and re-imagined is code for the destruction of all that is old and it’s replacement by IT enabled capabilities and business models.’

Joe Weinman 08:39

So I think maybe to kind of give the Carr argument it’s due but putting in a broader context, it’s perhaps safe to say that there’s a range of IT value from tactical and operational to strategic and even existential. And the range of benefits that one man can ascribe to this different layers of IT value, includes everything from the base functionality of the system that may enable the use of software as a service type of capability total cost reduction through optimization. One of my proofs that’s been around for about five years shows that hybrid architectures are always optimal in the presence of variable demand, subject to a few caveats. Business agility which is different to quantify exactly as opposed to IT agility which is, but is certainly a benefit to operational level deployments but then there’s this whole range of higher level value that can include the survival of the firm and a good example of that is borders which is no longer around thanks to competitor over there that was able to leverage information technology in the Cloud and the web appropriately. So, my book, I’m sort of covering one of the chapters in there but I have the exact counter argument to the Carr IT doesn’t matter argument which briefly is, we live in an increasingly digital world given all that, it’s probably obvious to many of us but not always obvious for those of you who that are CIO’s that report to a CFO, or in let’s say, more mainstream business that isn’t as with it necessarily around the ability leverage digital technologies. Because we are in an increasingly digital world, of course, IT and the Cloud is an imperative and of course it’s therefore a strategic and existential and the strategy rather than cost reduction should in fact be focused investment.

Joe Weinman 10:43

Now if I just stop there and said, well IT and the Cloud are good, you may be have some useful arguments but I want to just quickly review some of the quantitative data. This is from [Eric Branielson?] in MIT. He basically shows a direct correlation between the use big data and decision making and profitability and return on equity and the usual business matrix that we would want to consider as well as another study from [Siniel Methos?] said, ‘The university of Marilyn, the 1.912 up there is a critical number, it shows that after adjusting four than that present value of money and different things like various statistically facts and not confusing correlation with causality. A dollar invested in IT actually returns $2 to the company, and that’s a better return actually than advertising or RND which is a little bit surprising perhaps but it shows that IT is a great investment. Now, the numbers are good but what I wanted to try and do was explore the strategic nature of IT and so what I did is I used this framework from Tracy and Wiersema called the Value Discipline Framework. Some of you may remember the original HP article that was customer intimacy and other value disciplines where they argued that there are three main ways in which companies can strategically differentiate themselves, one is operational excellence that are processes. Another is product leadership, better products or services. And the third is customer intimacy or better relationships with customers. So even if you’re processes or terrible and you’re products are me too, if you have a great customer relationships that can actually be just as effective in means of competing in the market.

Joe Weinman 12:28

To the remain three, I added a fourth which seems to be pretty important which is accelerated innovation. Now, I’m going to quickly go through then some examples of IT and the Cloud and big data and mobility applied to these strategies. It’s pretty clear to us how IT is important to Cloud native or digital businesses. But I think what’s interesting is applying it to more mainstream, less quote, high tech businesses. So one example of operational excellence is a supply chain package delivery, field support type of company where the idea would be that being able to leverage real time routing information playing that against some algorithms that look at both our road congestion, packing strategies on the truck as a ways for platinum customers versus gold customers being able to do optimal routing to basically minimize a slay violations, maximize customer satisfaction, maximize customer satisfaction, maximize customer lifetime value as well as increase sustainability and labor productivity and things like that. Another good example is product leadership or something like the iPads, Smartphone, Phablets, or all great devices but wouldn’t be much without the Cloud and that includes things obviously like apps stores but also hybrid functionality where the actual app running in the mobile device endpoint is hybridize with some capabilities running in the clouds. And actually I went to the store apps yesterday to address that but something as simple as Siri or search or Facebook, obviously, you won’t be able to do social networking just on your device on an uncharted basis. It’s not just high tech devices like iPads low, it’s also cars, yesterday we heard Mario from BMW talk about basically how the car is just an end point. What I think is really fascinating is how you can differentiate an athletic clothing brand with these type of advance technologies as well. Think something like a Nike fuel or a Nike plus.

Joe Weinman 14:37

So the last pace is customer intimacy where the notion of being able to build better relationships and you can just think about something like an amazon.com or what have you, or a Pandora that really understands better what’s going on with customers, and then the last thing is the ability to do accelerated innovation through Cloud source contest. The last thing that I’ll close with because I’m out of time is the idea that we’re moving to a new world where rather than a vertically integrated firm that basically has all its IT within the four walls. We’re really getting to a point of not just a network firm with overlay organizations or even or a networked virtual corporation but a dynamic network company. A good example of that actually is Lean Fang which I just close with, where the ideas that they actually dynamically, we have too many builds here for being overtime, where what they do is they select their suppliers each week and alter them so it might be a fabric from Thailand today, buttons from Vietnam, next week it’s something else and as we think about this new world with global sourcing and that degree of dynamism having something like an inter-Cloud type of capability will be the essential IT underpinning to that. So, with that, I’m done. Thank you very much for your time. I know I cover a lot of material but hopefully it was of interest. Thank you.

firstpage of 2
  1. Paper use? Or pay-per use?

    Share
    1. Pay per use

      Share
  2. You can also grow and shrink your capacity (processing power, storage, network) very quickly, in minutes or hours. So there would be a factor of rapid elasticity.

    Thanks for your insight.

    Share

Comments have been disabled for this post