What McKinsey’s Athlete Mindset Model Gets Right — And Wrong — About Digital Transformation

McKinsey’s athlete mindset model for digital transformation has been circulating long enough now that it’s worth taking seriously — both what it gets right and where it runs into trouble.

The core claim: organizations that successfully navigate digital transformation behave more like elite athletes than like traditional enterprises. They train systematically, they iterate on performance, they have clear metrics, they have coaches, and they treat setbacks as data rather than failures.

There’s something genuinely useful in this framing. There’s also something it misses that becomes more visible when you push on it.

What the Athlete Metaphor Gets Right

The athlete model resonates for a reason. The organizations that navigate digital transformation best do share traits with high-performing athletes:

Deliberate practice over experience accumulation. Elite athletes don’t just play games — they specifically practice the things they’re not good at. The organizations that are winning at digital transformation are the ones with explicit, structured AI experimentation programs. Not just “we encourage innovation” but specific time, specific objectives, specific retrospectives.

Performance feedback at short intervals. Athletes know how they did in the last session. Most organizations do annual or quarterly reviews of digital initiatives. The feedback loop is too slow to learn from. The better ones are running weekly retrospectives on AI workflows and adjusting.

Physical and technical conditioning. Athletes need both physical capacity and technical skill. Organizations need both the infrastructure (technical capability) and the team (human skill). Most digital transformation programs over-invest in infrastructure and under-invest in team skill development.

Coaching as a structural role, not an occasional intervention. Elite athletes have coaches who observe performance and provide specific, ongoing feedback. Most organizations treat change management as a one-time implementation cost rather than an ongoing operational role. The ones that are succeeding have someone whose job is specifically to watch how AI tools are being used and provide that feedback loop.

Where the Metaphor Breaks Down

Here’s where I’d push back on McKinsey’s framing:

Athletes compete in stable rule sets. Digital transformation doesn’t.

A sprinter training for the 100m knows the distance isn’t going to change. The rules of the race are fixed. Digital transformation — and specifically AI adoption in 2025 — is happening in an environment where the underlying capabilities are changing faster than any organization can build expertise in them.

The athlete model implicitly assumes you’re improving for known performance. But a significant part of what makes AI transformation hard is that you don’t fully know what you’re improving for yet. The model that was best for your use case six months ago may not be the best one today. The workflow that made sense when your AI could produce mediocre first drafts looks different now that it can produce quite good first drafts.

This is more like an athlete training for a sport whose rules keep changing mid-season. The conditioning principles still apply, but the specific technical skills require more continuous re-evaluation than the metaphor suggests.

Elite athletes operate in individual performance contexts. Digital transformation is irreducibly organizational.

You can measure an athlete’s personal performance directly and unambiguously. Organizations can’t do that for AI adoption. The value often shows up in team-level outputs, not individual metrics — which makes the feedback loop messier and the coaching relationship more complicated.

The athlete model tends to drive individual AI certification programs and individual performance metrics. But AI transformation value is largely created at the workflow and team level, not the individual level. Improving for individual metrics can actually fragment team-level workflows.

The Missing Element: Adaptive Strategy

What the athlete model lacks — and what the most successful digital transformation programs actually have — is explicit adaptive strategy.

Adaptive strategy in this context means: we have a plan, we have clear signals we’re watching to know if the plan needs to change, and we have decision rules for when and how to change it.

Most digital transformation programs have a plan. Far fewer have the monitoring systems and decision rules that make the plan genuinely adaptive.

In practice, this looks like:
– Monthly reviews of which AI tools and workflows are in use, at what adoption rates, with what outcomes
– Explicit criteria for when a tool or workflow gets discontinued (not just when it gets added)
– Designated responsibility for scanning the AI landscape for capability changes that affect your existing decisions
– Scenario planning for what happens if a specific AI capability becomes available or unavailable

Elite athletes adapt their training when they get injured, when their event schedule changes, when they get new information about competitors. The adaptation isn’t reactive — it’s built into the training system.

That’s the missing piece most digital transformation programs don’t have.

What to Actually Take From the Athlete Model

Strip away the metaphor and here’s what’s operationally useful:

Build the training habit, not the event habit. Treat AI capability-building as ongoing conditioning, not project-based implementation. This means regular, scheduled time for teams to experiment, share learnings, and update their workflows — regardless of whether there’s a specific initiative driving it.

Invest in coaching infrastructure. Someone on your team (or in your network) needs to be watching how AI is being used and providing specific feedback. This doesn’t have to be a full-time role — but it has to be someone’s explicit responsibility.

Measure performance with short feedback loops. Don’t wait for quarterly business reviews to understand whether your AI workflows are working. Weekly output metrics, even informal ones, give you enough signal to course-correct.

Accept that setbacks are data. The athlete model’s strongest contribution: normalizing that high performance requires many failed attempts. Organizations that penalize AI experimentation failures will have slower adoption curves than ones that treat failures as learning events.

The athlete metaphor is useful scaffolding. Just don’t let it prevent you from seeing where the terrain is fundamentally different.

A More Honest Frame for 2025

If I were updating McKinsey’s model for the current environment, I’d combine the athlete’s discipline with something closer to an expedition mindset.

Expeditions have clear goals, rigorous preparation, structured roles, and feedback loops. But they also operate in genuinely unpredictable environments where the conditions change, the route may need to change, and new information constantly arrives that wasn’t in the original plan.

The best expedition teams maintain their training discipline while staying genuinely adaptive to conditions. They don’t mistake the route for the goal.

That’s closer to what effective digital transformation looks like right now. The AI landscape is genuinely uncertain. The capability changes are genuinely significant. The organizations that are building durable advantage are the ones with athlete-level discipline about developing and maintaining AI capabilities, combined with expedition-level adaptivity about which capabilities to develop and when to change course.

McKinsey’s framework gets you halfway there. This is the other half.

Nếu bạn thích bài viết này, hãy để lại email hoặc bình luận để chúng ta cùng trao đổi thêm.