Continuous Delivery

The Flaw at the Heart of Bimodal IT

Gartner’s “Bimodal” IT model has taken the world of enterprise IT by storm. Gartner defines Bimodal IT as “the practice of managing two separate, coherent modes of IT delivery, one focused on stability and the other on agility. Mode 1 is traditional and sequential, emphasizing safety and accuracy. Mode 2 is exploratory and nonlinear, emphasizing agility and speed.” In this post I will describe three flaws in Gartner’s model, the most fundamental of which is that high performing organizations do not in fact trade off agility for safety. In fact, high performance is characterized by consistent improvements in levels of both agility and safety. While the Bimodal model provides a valuable guide to enterprises beginning their agile journey, I will argue it is misleading and, I believe, dangerous when viewed as a destination in its own right.

The essence of Gartner’s model is this: We can divide a big chunk of enterprise IT into two kinds of systems: systems of record, which manage the sensitive data that is most valuable to our organizations (think bank account information), and systems of engagement, a set of public-facing systems by which our customers access our services¹. These types of services present two quite different sets of risks. The risks inherent in building and evolving systems of record are, say Gartner, better managed through waterfall, while agile methods are more suited to building and managing systems of engagement. It’s a nice, simple model, and it appeals to the experience of many of us in the IT world: Making changes to the gnarly systems at the heart of many enterprises, usually decades-old COBOL software running on mainframes or packaged software built by vendors, is painful, expensive and risky.

There are three serious problems with the Bimodal model which, when taken together, mean that leaders that fail to move beyond Gartner’s advice will end up falling further and further behind the competition. They will continue to invest ever more money to maintain systems that will become increasingly complex and fragile over time, while failing to gain the expected return on investment from adopting agile methods.

The first problem is that the model is overly reductionist. In Gartner’s world, we’ve moved from a one-size-fits-all model to a two-sizes-fit-all model. This is progress of a very crude sort for organizations stuck in the 90s, but the leading digital enterprises are making much more fine-grained risk management decisions at the level of individual products and services. Market leaders such as Google and Amazon are managing each of their thousands of services according to its unique risk profile, with continuous delivery as the default approach. Rob England has a good discussion of this shortcoming of the Bimodal model.

The second flaw in the Bimodal model is that those fast-moving user-facing services are almost always coupled to systems of record. Gartner acknowledges that synchronization between the modes is required at defined points. However the reality is that unless product owners of Mode 2 (“Agile”) systems are collaborating throughout the delivery lifecycle with the product owners of the systems of record they integrate with, the rate of evolution of any Mode 2 system will be constrained by the rate of change of the slowest system of record it talks to. In reality, the situation is even worse: the majority of enterprises are still forced to perform infrequent, big-bang orchestrated deployments even for their agile products and services, due to the tight nature of this coupling. Much of the benefit of using agile processes is thus wasted.

Finally, and most important of all, Gartner’s model rests on a false assumption that is still pervasive in our industry: that we must trade off responsiveness against reliability. The conventional wisdom is that if we make changes to our products and services faster and more frequently, we will reduce their stability, increase our costs, and compromise on quality.

This assumption is wrong.

Companies like Amazon and Google haven’t beaten all-comers by building flaky, insecure systems. What’s special about the DevOps movement is that it represents a game changer, a true paradigm shift. The people who built the DevOps movement had to solve a wicked problem: how to build reliable, secure distributed systems at an unprecedented scale, while enabling a rate of change orders of magnitude faster than the industry had ever achieved. The fruits of their approach can be seen in the original presentation that kicked off the DevOps movement—John Allspaw and Paul Hammond’s presentation from 2009, “Dev and Ops Co-operation: 10 deploys a day.” Amazon is now achieving change rates three orders of magnitude higher than this². One of the key results from the 2015 State of Devops Report is that high performers are able to increase the number of deployments per developer per day as they scale.

This kind of paradigm shift is not unprecedented. It strongly echoes the way that Toyota changed the game in manufacturing. Toyota didn’t win by just getting cars to market faster than the competition. It won by building cars faster and cheaper and at higher quality than the competition, and then accelerating away from other companies by working to continuously improve their capabilities along all of these axes.

I have been part of a research program for the past three years studying IT performance in companies of all sizes, in all domains (including highly regulated sectors such as financial services), worldwide. Our published, peer-reviewed research is unambiguous: high performers achieve both higher levels of throughput and higher levels of stability than their low performing peers. They achieve this by implementing a high performance, lean culture and adopting continuous delivery practices across all of their products and services.

While Gartner is correct that a Mode 1 approach cannot be used to build modern, responsive, user-centric services, the converse is not true. Agile methods and continuous delivery have been successfully applied to everything from mainframe systems in large financial services companies to embedded systems in consumer electronics, consistently delivering higher quality, faster delivery, greater business responsiveness, and reduced costs over the product lifecycle.

Any company that hopes to survive in the digital age must move beyond zero sum thinking. The recipe is easy to understand, but hard to implement: leaders must set and communicate clear business goals in terms of time to market, quality, and cost. They must then invest the necessary resources for everyone in the organization to collaborate so they can solve the problems that prevent them from achieving these goals. Nothing should be out of scope—enterprise architecture, process, budgeting, and governance, risk and compliance. Most important, lasting change must come from inside the company. It can’t be hired or bought or consulted in.

Gartner’s work on Bimodal provides valuable advice on how to kick-start an enterprise digital transformation. But leaders wishing to create a significant, lasting impact on IT performance should look to move beyond the Bimodal paradigm in the space of months rather than years. The most important enabler here is also, as John Kotter writes in Leading Change, the hardest to achieve—a sense of urgency at all levels of the company that significant change is absolutely necessary. Given this sense of urgency, a clear set of goals, and a culture of collaboration and problem-solving throughout the organization, anything is possible.