Strategy execution has one of the worst track records in management. Harvard Business Review has estimated that 67% of well-formulated strategies fail because of poor execution, not poor thinking. Kaplan and Norton, in The Balanced Scorecard, put the figure even higher: up to 90% of organizations never manage to execute their strategy as intended. A 2015 academic review by Candido and Santos, screening decades of studies on the topic, found failure rates ranging from 7% to 90%, averaging out at roughly one strategy in two.

Numbers this wide and this consistently bad point to something structural, not anecdotal. And when you look closely at where the failure actually originates, not at the 30,000-foot strategic view, but inside the operating rhythm of the organization, a pattern emerges that has little to do with the strategy itself.

It rarely starts with the plan. It starts with who owns the next decision

I spent fifteen years running operations inside pharmaceutical and biotech organizations, from lean SMEs to multinational structures. Across very different portfolios, the failure pattern was almost identical: the strategy was sound, the team was competent, and yet execution stalled at the same three chokepoints every time.

One case makes it concrete. I took over a multi-project pharmaceutical portfolio with real commercial value at stake — solid science, reasonable roadmap, capable people. And still: inboxes with hundreds of unread threads, decisions reopened two or three times because no one remembered who had actually closed them, and board meetings that surfaced problems the operational team had already known about for weeks. None of this shows up on a strategy slide. All of it eventually shows up on a P&L.

The intervention wasn't a new strategy. It was three unglamorous mechanisms:

  • A decision protocol — a short, explicit map of who decides what, so the portfolio lead stops being the default answer to every question.
  • A gateway process with defined outputs at each stage, replacing status meetings that existed only to generate more meetings.
  • A live dashboard that made progress visible without anyone needing to ask for it, because if it isn't on a dashboard, in practice, it doesn't exist.

Eighteen months later: marginality up 10%, time-to-market down four months, internal email volume down 70%. No reorganization, no new headcount, no new software platform. Just structure applied at the three points where value was actually leaking.

The three foundations, often built in the wrong order

Underneath every operating model that works, there are exactly three layers, and organizations consistently build them in the wrong sequence:

1.  Governance: who decides, and how fast. Most leadership teams build this last, after the cost of ambiguity has already compounded.

2.  Execution: the roadmap, the milestones, the metrics that make progress visible before it becomes a surprise.

3.  Communication: deliberate flows that replace "everyone is copied on everything" with the right information reaching the right five people at the right time.

Governance is the layer almost everyone skips. At ten people it feels like bureaucracy. By fifty, its absence is the leading cause of founder burnout and board-level surprises, and by then it is far more expensive to retrofit.

Borrow the mechanism, not the machinery

There's a real temptation to reach straight for the frameworks that built giants: IBM's process governance, the Toyota Production System, Spotify's squad model. Each contains a genuinely transferable mechanism. None should be adopted wholesale by a fifteen-person team.

What actually transfers is narrower than the framework itself:

  • From process-excellence thinking: you don't need enterprise tooling to expose a bottleneck. Three to five tracked processes and a simple dashboard will surface most of what is actually slowing an organization down.
  • From lean manufacturing: a "stop-and-fix” culture, where a flagged problem halts the line instead of being escalated into next week's meeting, is the real quality lever, not the visual boards people copy for their aesthetics.
  • From squad-based tech models: autonomy without a shared alignment ritual isn't agility, it's fragmentation with better branding. Two or three teams with a clear mission and a genuine weekly sync will consistently outperform ten "empowered" teams working in silence.

The mistake companies make isn't underestimating these methods, it's importing their scale before they've earned the need for it. A fifty-page stage-gate process belongs to an organization running fifty projects in parallel, while is not suitable for a company running five.

The uncomfortable conclusion

Most organizations treat execution failure as a discipline problem, to be solved with more accountability, more urgency, more hustle. Across fifteen years and very different portfolios, that has rarely been the actual diagnosis. What sits underneath, almost every time, is an architecture problem: decision rights left ambiguous, progress left invisible, communication left to good intentions instead of design.

By fixing the architecture, the discipline tends to follow on its own, because for the first time, people can actually see what they are accountable for.