The DevOps Disruption
Author: Jerry Overton and Gene Kim
Why invest in DevOps? Can a DevOps transformation lead to competitive advantage? Can it lead to market disruption? These are the questions we sought to answer by simulating DevOps transformation scenarios and comparing them to known business innovation patterns.
By doing this, we can deliberately design strategies for the future we want. Our key finding was that DevOps transformation can not only lead to competitive advantage, but under the right conditions, it can create market disruptions. In other words, when we can convert our systems of innovation into systems of disruption, the value of adopting DevOps may be far higher than we previously thought. Download The DevOps Disruption.
This report — part of ongoing research designed to make business innovation more of a (data) science — is based on the results of simulation modeling. We’ve published the model in the form of an interactive web application to allow this report’s audience to run their own simulations and provide us with feedback.
Firms announcing transformational IT investments tend to experience positive, abnormal changes in market value. Simulation gives us a way to explore this pattern in more detail by simulating the specific effects of DevOps transformations. We can represent the performance of a business model based on the money invested into the business and the revenue generated as a result.
Business model simulator
We modeled the influence of a handful of performance factors using a set of simple one-variable mathematical functions. We added random jitter to each function to represent uncertainty and chance within the simulation — unforeseen problems with suppliers, chance improvements in efficiency, unexpected jumps in buyer purchases, etc. The simulation predicts a final return on investment (ROI, revenue as a function of costs). Parameters allow us to tune the simulation so that it represents a specific business model. The simulation output can be thought of as theoretical propositions or hypotheses.
In this research, we explain how the business model simulator leads us to the following hypothesis: Companies get a higher ROI and competitive advantage from their DevOps investment — but when the basis of competition is innovation, the effect tips from an advantage to a disruption.
What's important to customers?
We start by simulating two business models, both under the same set of controlled conditions. The coefficient of innovation describes the importance customers place on new features when adopting a product: The higher the value, the more consumers are influenced to purchase. For example, Intel’s 80486 PC released in 1991 had a coefficient of innovation of 0.0160 — nearly twice that of IBM’s 1984 G4 mainframe, which had a coefficient of innovation of 0.0089.2 Using a coefficient of innovation allows us to define innovation based on how product features drive adoption, rather than on any intrinsic property of the product features themselves. Sultan, Farley and Lehmann’s analysis of 213 product diffusion models found the average coefficient of innovation to be 0.03 and the average coefficient of imitation to be 0.38.3
For most consumer goods and services, price elasticity tends to be between 0.5 and 1.5.4 We assume a price elasticity of demand of 1.1, which makes our simulated business models approximately unitary elastic — every percentage decrease in price leads to a percentage decrease in demand. Our model assumes an output elasticity of 0.5, which is reasonable given that Gau estimates the long-term output elasticity of the U.S. labor market to be 0.41.5 The average improvement rate for a new product in the United States is 0.25.6 For our control conditions, we assume a more conservative improvement rate of 0.15.
Measuring the effect of DevOps transformation
What effect does a DevOps transformation have on the ability to compete in the marketplace? Companies that deploy IT automation generally do so to improve the efficiency of existing business processes.7 The IT development benefit from DevOps means that the business experiences shorter development cycle times. In the model, partner efficiency is simulated using the microeconomics notion of a learning curve. Efficiency in execution is based on the time it takes to get through a production cycle and how much is learned in each cycle. The smaller the cycle times, the faster that learning occurs. This kind of execution efficiency is very important in scenarios like DevOps transformations where there are short development cycles and continuous learning feedback.
To learn more, read the The Devops Disruption.
About the Authors
Jerry Overton is a data scientist and Distinguished Engineer with DXC’s ResearchNetwork.
Gene Kim is co-author of "The Phoenix Project: A Novel About IT, DevOps, and Helping Your Business Win" and the upcoming “DevOps Cookbook.”