An Irresistible Pitch

I worked at a Fortune 500 company at the dawn of the cloud revolution, when every CTO was racing to move everything onto cloud providers. The pitch was irresistible: run the business online for a fraction of the cost, with none of the hassle of managing data centers around the world. So we committed. We spent years migrating, and the migration became the strategy. We missed opportunity after opportunity. We left half-finished projects scattered behind us as teams moved on to the next migration. The system got more brittle, more convoluted, and more loaded with technical debt than it had ever been before.

Setting oneself on a predetermined course in unknown waters is the perfect way to sail straight into an iceberg.

McChrystal's line is the diagnosis. We had set the course in waters we did not understand. The pitch made the future look like calm seas. The future was nothing of the kind. Every quarter that passed without naming the course as a hypothesis turned the course into the strategy, and the strategy into the only thing the team knew how to defend.

The punchline is that the cloud now costs more than the data centers it replaced, once you add up every service the providers bill for on top of basic hosting. We lost our flexibility and our in-house expertise, so the provider can raise prices and we have no recourse, while competitors run faster and more modern. It was a huge, expensive mistake.

The Same Mistake, Across the Industry

Accumulated reactive decisions are expensive. Product launches often fail on poor change management, and many of those failures trace back to organizations that committed to a single industry-blessed direction and then lost the ability to consider alternatives.

What the Framework Would Have Surfaced

The hardest part is that the cost trajectory was foreseeable. If the team had dug into the data, run experiments, hedged its bets, and explored the scenarios, we would have seen it coming. Assumption Challenge would have surfaced the belief that cloud cost would stay below data center cost as the highest-consequence assumption under the entire strategy, and would have proposed experiments to test it before committing. Scenario Preparation would have drafted at least one scenario in which the provider raises prices and the company has no recourse. Constraint Setting would have set a reversibility budget so that no migration commitment ran longer than two quarters without an explicit decision to continue. Pivot-vs-Persevere would have given the team a structured place to ask, three years in, whether the migration was still the right strategy.

In the beginner's mind there are many possibilities, but in the expert's mind there are few.

Suzuki names the second failure mode that compounded the first. By the third year of migration we had become experts in a particular kind of work: moving things from one place to another. The expertise was real and the work was hard. But the same expertise that made us competent at the migration also made it almost impossible for us to see what we were giving up by continuing. A reversibility budget would have forced beginner's mind back into the room on a fixed schedule, whether we wanted it there or not.

How to Avoid Making the Migration Your Strategy

• Name the assumption that the industry-blessed move will keep saving you money, explicitly, with a confidence rating. Then design an experiment to test it within ninety days.
• Draft a scenario in which the vendor raises prices, sunsets your favorite service, or consolidates with your competitor. Decide ahead of time what you would do in each case.
• Set a reversibility budget. No multi-year migration should run without a forcing function that requires the team to decide again, in public, whether the strategy is still right.
• Audit which strategic decisions are migrations of capability versus migrations of vendor. Most failed cloud migrations conflated the two.
• Build in-house judgment alongside vendor capability. The skills you outsource permanently are the skills the vendor knows you cannot reclaim cheaply.

SharedForesight exists because this kind of mistake is what happens when adapting is something a group intends to do rather than a practice it runs. The nine steps are the structured response to watching decisions calcify with no shared place to keep the questions open.