A strategy describes where a company intends to go. An operating model is the machine that gets it there, how decisions get made, who owns what, which work the organization automates, and where intelligence sits in the flow. Most strategies don’t fail because the thinking was wrong. They fail because the operating model underneath them was built for a different, smaller company, and quietly strangled the plan.
Operating model design is the discipline of building that machine deliberately instead of letting it accrete by accident.
TL;DR
- An operating model is the machine that delivers a strategy: how decisions get made, who owns what, what gets automated, and where intelligence sits.
- Most strategies fail not because the thinking was wrong but because the operating model underneath was built for a smaller company.
- Designing it deliberately answers four questions: how decisions are made, how work is organized, what the company automates, and where intelligence belongs.
- The modern operating model is a learning system, so operating model design and AI strategy are now one conversation, not two.
- It is the right work when the constraint is structural: the strategy is clear but will not execute, and growth is exposing the seams.
What an operating model actually is
Strip away the consulting vocabulary and an operating model is the answer to a few concrete questions:
- How are decisions made, and by whom? What gets decided at the front line, what escalates, and how fast.
- How is the work organized? The structure of teams, the handoffs between them, and where accountability actually lives, not the org chart, but how value really moves.
- What does the company automate? The line between work that needs a human and work that should run on rails.
- Where does intelligence belong? Which decisions and workflows are improved by data and AI, and which are deliberately left to judgment.
A target operating model is simply that picture drawn for where the business is going, not where it has been.
Why strategy fails without it
Here is the failure I see most often. A company commissions a sharp strategy, new segments, a pricing change, a faster motion, everyone agrees, and twelve months later almost none of it has happened. Not because anyone resisted. Because the operating model couldn’t carry it.
The new motion needed decisions made two levels lower, and the model still routed everything through the founder. The pricing change needed the finance and sales workflows to talk, and they didn’t. The expansion needed a team that owned it, and ownership was diffused across three people who each assumed another was responsible. The strategy was sound. The machine was built for a different job.
This is why I treat operating model design as part of strategy, not a separate phase. A growth thesis that the organization structurally cannot execute is a wish, and I’d rather find that out while we’re still drawing the map. The growth strategy consulting guide covers the thesis; this is the part that makes the thesis operable.
The operating model of the next decade is a learning system
There’s a structural shift worth naming. For most of business history an operating model was a static thing, a structure you designed, installed, and revisited every few years. That assumption is breaking.
The operating models that win now are learning systems: they instrument their own decisions, feed the results back, and improve without a reorganization. Intelligence isn’t bolted onto the edge of the model as a tool the marketing team uses; it’s woven into how the model decides and operates. Which means operating model design and AI strategy are no longer separate conversations, the question “where does intelligence belong in this workflow” is now part of designing the workflow at all.
The practical version of that question is usually narrower than it sounds. It’s not “how do we become an AI company.” It’s “which specific decisions and workflows in this model should run on data and automation”, the discipline of choosing the right workflows rather than rebuilding everything.
How an operating model redesign unfolds
A serious engagement runs roughly in this order:
- Map the model you actually have. Not the official org chart, the real one. Where decisions truly get made, where work piles up, where accountability is fuzzy, and where the current model is already straining against the strategy.
- Define the model the strategy requires. Working back from where the business is going: what has to be decided faster, owned more clearly, automated, or rebuilt for the plan to be executable.
- Locate the intelligence. Which workflows and decisions should be data- and AI-driven, and which stay human, designed in, not added later.
- Sequence the change. No company can rewire itself all at once. The redesign becomes an ordered set of changes the organization can absorb without breaking, each one making the next possible.
The deliverable is not a chart. It’s a sequenced plan for how the company will decide, operate, and learn, costed, owned, and dated like any other part of the strategy.
When you need this
Operating model design is the right work when the constraint is structural rather than directional: the strategy is clear but won’t execute, growth is exposing the seams of a model built for a smaller company, or decisions that used to be fast now crawl through a structure that’s outgrown itself. Those are also several of the signals that it’s time to bring in a strategy consultant at all.
If, instead, you don’t yet know where the business is going, the operating model is premature, you can’t design the machine before you’ve decided the destination. Strategy first, then the system that delivers it. But once the destination is set, the operating model is the difference between a strategy that ships and a strategy that frames nicely on the wall.