The Allure of Sweet Fudge Versus a Lean Transformation Diet

In my role as an AWS Enterprise Strategist I am privileged to travel the globe working with amazing customers and their leadership teams. In nearly all of these encounters one of the questions, if not the first question is “What should my Target Operating Model be as we grow and use AWS Cloud?”. The TOM, as we have colloquially called it, is always a fascinating conversation. And I often refer to my own experience running Agile teams with AWS previously as Senior Director and CTO at Capital One UK and the many shared learnings of the customers that I visit from around the world.

I am incredibly lucky to work with a very diverse team at Amazon. And in conversation with John Enoch, Principal, Technical Business Development, I discovered he also had a fascinating way of thinking about it. So, here in his own words a guest blog from Mr Enoch.

Jonathan Allen
Enterprise Strategist & Evangelist, AWS

The Allure of Sweet Fudge Versus a Lean Transformation Diet by John Enoch

What do leaders of large, complex enterprises with huge technical debt think when looking over a cliff of urgency in their digital transformation projects? After all, “digital” infers binary ones and zeros processed by software to generate desired outcomes. Not all senior executives may have an in-depth knowledge of the importance of their organization’s ability to innovate, iterate, and reliably run software in a cost effective manner, let alone how to transform their operating model to be “digitally enabled.”

Such executives may be considering whether organizational change, a rethink of the role of technology, and what IT exists to deliver are decisions that really need to be taken today. Could hard choices be safely parked for a while, delayed or driven more slowly? The mathematician and satirist Tom Lehrer commented in the 1960s that looking at a world of great change and risk can make one feel either helpless or reliant on faith in higher powers. As he uniquely put it, “like a Christian Scientist with an appendicitis.” Faced with great risk and uncertainty, our belief systems can indeed create a sense of helplessness. CIOs faced with complexity and legacy debt could reasonably share that sentiment.

What if an executive team is constrained by a known lack of organizational budget or a low appetite for change? Wouldn’t it be compelling to have assurance that perceived urgency is an illusion and that change at scale to improve customer experience, governance, and cash-flow is not the only viable option available? What if, as in Clayton Christensen’s “The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail,” organizations could split IT operations into two and operate “new and disruptive” and “old and stable,” with both performing to plan in tandem? In 2014, Gartner started promoting this message as a viable approach to managing IT operations across the divide of a digital transformation and called it “bimodal IT.” Gartner created a valuable debate on whether cloud adoption could be successful, with innovation and agility benefits delivered, while legacy continued to run as usual, without a clear exit plan for those legacy assets and operating models.

Gartner’s assumption is that old world, legacy, technical debt is comparatively stable and can be run as “Mode 1,” managed safely and separately as-is and to be migrated to the cloud as-and-when. In parallel, “Mode 2” is cloud native, innovative, agile, more prone to failure, and to be driven by a separate team within the same organization. While this sounds intellectually elegant, Gartner was challenged on this two-tier IT operating model by Forbes (e.g., Bimodal IT: Gartner’s recipe for disaster”), Forrester (e.g., “Bimodal IT is past its due date”), and many others.

I have yet to see a body of quantified, empirical evidence that bi-modal is a viable choice that delivers the promised results of digital transformation. Furthermore, can it be proven that bi-modal IT delivers better results than a unified approach to executing a change program that is based on real-world experience, such as is outlined in the blog 12 Steps to Get Started with the Cloud?

To date, I have seen no such evidence. Gartner did conduct a CIO survey in 2017 that claims to validate their bi-modal theory. Such survey-based research methods are largely based on individual responses to interviews and questionnaires. Such methods can be questioned for being at best indicative based on the sample and at worst open to respondents’ bias and ambiguity, depending on the choice and wording of the questions.

I am surprised that bi-modal has gained such traction and entered language as a viable approach to cloud enabled change. I hear IT executives refer to their status in running “Mode 1” and progress in executing on their investments and plans in “Mode 2.” It is as if they are normalizing a health and fitness routine of eating “super-sized” burgers daily, while one day a week is dedicated to a separate diet of eating celery and doing five-mile runs. Bi-modal presents the idea that two incompatible approaches can be operated together, within the same organization. Christensen advocates the creation of a separate business entity with a new culture and operating model; not a new silo within the same business.

I think it best to leave the job of arguing the benefits of bi-modal to Gartner, as the authors of the theory. After all, in 2016, “bi-modal” ranked at number six of the top search terms on Gartner.com.

The most common objections to Gartner’s bi-modal construct include claims that:

• It creates a false sense of security that the pace of external disruption can be managed within the constraints of internal processes and capabilities.
• It ignores the risk of Mode 1’s increased fragility and incongruence over time.
• Mode 2 systems likely tie into and are dependent on Mode 1 processes from day one.
• The transition state of “hybrid” can easily become static and permanent.
• It duplicates and complicates by creating two systems, processes, and even cultures.
• Bi-modal relies on splitting the IT organization into two and creating siloes. Best practice aligns IT budgets with lines of business serving customers, not technology.
• It creates a good vs. bad technology team, with those on the “bad” team feeling disenfranchised, without a career path, and more likely to disengage and leave.
• It creates a political dogfight between the two siloes over IT budgets and priorities.

Shortly before I joined Amazon, I spent time with an insurance company that invested in operating bi-modal IT. They had a large IT Infrastructure team, huge legacy debt, and had bought a smaller, agile-based IT company to drive change from within. The new agile silo in the organization was totally dependent on IT’s old waterfall-based process, and the team had to grab a ticket and wait in line for whatever needed to be delivered, the same as any anyone else. The status quo of the IT operating model was not challenged because it was running to plan as Mode 1. The morale in the new team deteriorated as timelines slipped and frustration grew. Interventions from the CEO did little to change things, since some senior executives were heavily invested in the status quo, a few years from retirement, and did not want to embark on an enterprise-wide change initiative.

More recently, I have seen banks tempted by offers of IT outsourcing (ITO) of their data centers so that legacy skills can be siloed externally and the organization can focus on innovation in-house. This is another form of bi-modal IT: the ITO promises “keeping lights on,” smooth running of existing assets while taking those assets off the books. Looking more closely at the market value of a typical data center versus its book value and the cost and risk of recovering the difference, it would require somewhat onerous terms for the ITO provider to deliver a commercially profitable contract. The delays in successful execution of customer relevant IT projects once the urgency of executing clear change plans may be even more costly than just biting the bullet and executing a change program.

How people see problems tends to dictate their choice of solutions. If the problem being addressed is the cost and risk of change, then change will likely be constrained by cost and risk. If the issue is selecting IT choices that drive up customer experience scores, tighten controls, and increase free cash-flow, then change is more likely to be accepted as a tough but necessary decision since the status quo is unsustainable. There may be no available decision fudge of a middle way.

Standish Group’s annual chaos report from 2017 found that only 36% of major IT projects deliver on expectations, budget, and timelines; 19% are abject failures; and the rest are “challenged.” The odds of delivering an IT-enabled transformation project successfully are low. The 2018 Standish Group report gave insights into how to improve these odds. It highlights the impact of “decision latency.” Organizations that adopt an Agile approach to project delivery will be able to cycle through decisions more effectively, and achieve 58% success and 9% failure rates. In contrast, those that maintain a more complex waterfall approach only have an 18% probability of success, with 32% deemed failures and 50% challenged.

By maintaining a bi-modal approach to running IT, there is no impetus to change the process and method that has delivered such statistically poor outcomes to date. Rather than focusing on the cost of change, perhaps organizations should ask how much money is spent on IT projects, what percentage of projects fail, and how much, for example, a 40% rise in project success rates would add to cash flow. IT should be focused on where there is a need to do more and faster to serve customers better than competitors, regardless of technology and process. There is no substitute to a well-structured change program, and the hard choices and trade-offs it entails, starting with the initial 12 steps.

Is there a body of objective, empirical evidence that proves that there really is a trade-off between running IT as and Agile, DevOps based cloud versus on-premise that equates to “fast and innovative” versus “slow and stable” according to the bi-modal IT construct? If so, I have yet to see anything that evidences measurable impact on cash flow, customer experience, risk, and their combined effect on company valuations over time. The evidence I do see is that disruptive, successful organizations, with high market values such as Amazon, AirBnB, Netflix, Salesforce, Pinterest and many others are adopters of cloud at speed and at scale, and are not attributing their competitive success to running bi-modal IT.

John Enoch
Principal, Technical Business Development, AWS

LinkedIn

Forbes

John Enoch

John joined AWS in 2016 and specialises in Cloud Economics and risk. He works with commercial and technology executives to help quantify the impact of IT choices on cashflow, customer experience and risk. Prior to joining AWS, John was a Fintech CEO, a Director in Deloitte’s Audit Advisory practice and an IT Strategy Principal at PricewaterhouseCoopers. He has a patent on a compliance analytics technology and an MSc in Decision, Cost and Policy Analysis from the Engineering faculty at the University of Gävle in Sweden.