Behaviors and Biases: Understanding Cloud Economics

You can tell when an idea becomes mainstream when it appears in multiple, primetime, weekend, broadcast TV Ads.

From an IT perspective, cloud computing has clearly mainstreamed, with promotions from Amazon Web Services, Adobe, and IBM. Cloud computing has become a big deal not as a result of hype, but simply because the economics of cloud make sense.

Cloud economics make sense for two principal reasons.

Economies of Scale: Organizations such as Amazon Web Services can purchase computing resources in greater quantities and at a lower total cost than IT departments of individual companies.

Global Reach: Cloud computing resources, when combined with the pervasive connectivity of the Internet, abstracts the location of resources. As a result, the labor costs for creating, deploying, and maintaining IT resources can be optimized as well.

The Economics of Cloud Economics

In order to frame the discussion, it’s important to break down the phrase “cloud economics”. “Cloud” as mentioned above is well understood. But the “economics” portion does not merely mean that “We can save you money”. There is a larger notion behind it.

“Economics”, as broadly understood, is the study of human decision making, most frequently in the area of commerce. There are two fundamental models of economics. First, there is the “Traditional model” of economics.

Traditional Economics

Traditional economics has two fundamental precepts: First, that humans, or in our case Homo Economicus are rational, always attempting to maximize their “utility”, or value as a consumer. Utility can be thought of as a function, where all choices have a real-number value, and the choice with the greatest value always is the rational choice. The second precept is that Homo Economicus updates opinions and beliefs correctly, based on new information. This model, developed in the late 18th and early 19th centuries by the likes of Adam Smith, Thomas Malthus, and John Stuart Mill, can be thought of as “the theory of decisions people should make.”

1.Behavioral Economics vs. Traditional Economics: What is the Difference?”, Dr. Vicky L. Bogan, Hartford Funds Website.

Behavioral Economics

The second model of economics is known as the behavioral model. It has, as its base, a study of actual human psychology around decision making. This model has two core ideas as well. First, that humans tend to have “blind spots” in their thinking that cause them to make mistakes, sometimes repeatedly and systematically. This often is a result of cognitive biases, or flawed ways of thinking. The second core idea is that context, or the framing of the problem, matters a lot when making a decision. The model of economics was developed starting in 1979 by Kahneman and Tversky and formalized by Richard Thaler.


2.“Prospect Theory: An Analysis of Decision Under Risk”, Kahneman, D., and Tversky, A., Econometrica, March 1979, v47, ii, p.263-291. Accessed from JSTOR


It should be noted that it was a secular change for a field that was popularized by Adam Smith in 1776 with The Wealth of Nations. Secular is an interesting word, and it is included here because it is frequently used by VMware CEO Pat Gelsinger. It derives from the Latin word saeculum, which means “age”, or “generation”. The implication is that this kind of change is permanent and long-lasting, and represents a new “age”.

These concepts revolutionized and broadened the field of economics by being able to build a model of economics based not only on rational behavior, but by the actual behavior of people. It represented a fundamental change, and a new age in economics.

The advent of cloud computing (and storage and networking) is a similar, secular change. There are elements of this change that mirror the expansion of economics beyond the traditional model with the addition of the behavioral model.

Traditional Cloud Computing: Money is Smarter than People

From a narrow perspective, the value of cloud computing comes from traditional economic concepts like “economies of scale”. Customers are able to save money because large cloud vendors, like Amazon, can purchase and set up hardware in bulk, and offer it at a substantially lower cost than if the customer had purchased it on their own.

Similarly, cloud providers can rent or lease computing power on an on-demand basis, essentially reducing the friction of getting a commodity (in this case CPU cycles, etc.) to the customer. Providing this kind of flexible consumption is clearly more efficient: using only what is needed, when it is needed.

There is a saying in the financial service industry, “Money is smarter than people.” This is meant to indicate the traditional economic notion of market efficiencies – that capital always will gravitate to its best (most profitable) use. Cloud automation technologies, like those in VMware’s vRealize product line, enable users to do arbitrage. That is to allow cloud tasks to be done on the most efficient platform, allowing customers to take advantage of the most appropriate cloud platform for their workload.

Cloud Economics: Switching Cost

In classical economics, when Homo Economicus sees a net penalty in not choosing a new option, or sees a net profit over switching, the switch is made. The cost of switching from one choice to another is not only monetary, but also involves psychological, labor, and time costs as well. When switching from one system to another, the cost is not only the cost of the new option, but the cost and the time of the work needed to make the switch. This is especially true in terms of deciding to choose between an on-premises (traditional) computing environment and the cloud.

Behavioral Aspects of Cloud Computing

Among Kahneman and Tversky’s contribution was to catalog a series of blind spots that occur when traditional economics indicates that “irrational” decisions are made. These blind spots are useful to keep in mind when choosing to move part or all of an IT environment to the cloud. These blinds spots, the list of which is substantial, include the following, along with a few questions to ask yourself when considering a move to the cloud.

Overconfidence Blind Spot: Are you overconfident that you understand the costs and timeline of the project? Have you adequately determined what your “Rework Tax” will be when moving to a new architecture?

Recency Blind Spot: Are you making the right choice in platforms? Are you responding to the newest and hottest technology? Have you taken a clear-eyed look at which existing technologies will help you achieve your goals the most rapidly and at the lowest total cost?

Confirmation Blind Spot: Have you designed your solution based on pre-existing notions, or have you full reviewed the data to determine if the value of new technology plus some skill acquisition will put you in a better future? Are you playing the long game, or making short-term assessments?

Blind Spots for Cloud Infrastructure

These blind spots can affect your choices. Some common cloud blind spots are listed below:

 

The Refactoring and Rework blind spot: As discussed earlier, refactoring applications to run on the public cloud requires significant investment in development, test, and deployment. Development time for a refactored application for native cloud images can consume months or years. Combined with the number of applications running in a data center, such an approach can consume innovation budget better used elsewhere. Since VMware Cloud on AWS provides a common platform across cloud providers and on-premises infrastructure alike, the Rework Tax is fixed.

The Talent Reskilling Blind Spot: IT organizations have invested billions of dollars in management and operations solutions. When incorporating native cloud platforms, much of that investment is not transferable. Organizations often find they need to maintain multiple operations team, adding costs to a hybrid or multi-cloud agenda. The common platform for VMware Cloud on AWS eliminates the need for platform retraining.

The Lock-in Blind Spot: IT organizations assume significant rework costs to move to a native cloud architecture. Organizations desire options to take advantage of the multi-cloud landscape. Hybrid architectures benefit from portability of applications. The costs of reworking apply both ways, and can create barriers to portability, cause lock-in, and limit innovation. Using any provider-specific API can immediately create lock-in, so care must be taken when architecting both the platform and the application.

Operational Costs Blind Spot: On-going operational costs are often myopically viewed as a summation of the entire cost structure. Moreover, not paying attention to the cloud cost structure can result in some nasty surprises. All the major cloud providers charge for data egress. Data egress is when data leaves the providers network, going somewhere else. Depending on your data flow, these charges could easily overwhelm your budget.

Conclusion

The ancient Greeks had an aphorism, “gnothi saueton”, or “know thyself” in English. This is particularly useful when considering what and how to move apps and workloads to the cloud. While it is important to understand the basics of the value provided by cloud platforms, it is also useful to understand the mental model of the person or group architecting a new solution. Understanding cognitive biases, or blind spots, will yield a better, more complete and longer-term overall solution for your business.

Similarly, architectural solutions do not exist in a vacuum, context matters a great deal, and it is important to factor this into any final solution. Cloud computing is clearly popular and mainstream, and it can provide substantial value to a business, but to get the most value out of a cloud decision you must consider the implications of Cloud Economics – both classical and behavioral.

About the Authors

William Roth

Director, Product Marketing at VMware

Bill Roth is Director of WW Cloud Economics at VMware. He is a 30 year veteran of the enterprise software industry. He has been a CMO for startups, an evangelist, run product management organizations, an operating executive, and a journalist. He has worked for such companies as BEA, Sun, Morgan Stanley, and startups like LogLogic, and Nexenta. He lives in San Jose, CA.

Leave a Reply

Your email address will not be published. Required fields are marked *