The Six Things a CEO Can't Delegate

I advise companies for a living, so take this from someone who would profit from the opposite answer: you can delegate almost everything in a company, and you should. Research, analysis, planning and execution all improve when capable people own them. A short list of decisions is different. Hand those off and you haven't freed up your calendar. You've given away the job.

On 9 October, McKinsey Quarterly published a piece that names those decisions with unusual precision, The six strategy roles a CEO must not delegate.¹ My reaction, which you can hear in the video, was simple: McKinsey just put a name to it. The authors are careful to call strategy a team sport. Foresight, research, ideation and analysis can be spread across an organization. The CEO's stature and authority can't, and six roles depend on exactly that.¹

Below are the six, in the order I give them, with what each looks like when it goes wrong. The test at the end takes about an hour. It is uncomfortable, which is usually a sign it's useful.

First, decide how close you stand to the strategy

When the CEO isn't properly engaged, McKinsey writes, accountability for how strategy creates value fragments.¹ Finance shapes the portfolio through hundreds of budget decisions. HR runs the talent agenda. Each business leader runs their own operation. Nobody is wrong, and nobody is responsible for the whole.

The fix isn't more meetings. It's choosing the few places where your judgment changes the outcome and being there in person. McKinsey's example is A. G. Lafley at Procter & Gamble, who turned strategy reviews into focused debates with no presentations and only a few people from the business in the room. In his second tenure he divested or discontinued more than 90 brands to concentrate on the 65 to 70 that created the most value.¹ Nobody makes a call like that from a summary memo.

Then own the thesis and move the money

The second role is owning the value creation thesis, the plain answer to a question every CEO should handle without slides: how does this business make money, and why will it keep doing so? McKinsey's point is that ownership is more than authorship.¹ If the thesis exists only in a board presentation, it isn't owned yet. When your managers can explain it in their own words, it shows up in a thousand small decisions you never see.

The third role is where most CEOs flinch: reallocating resources. The unit that would lose budget always has the longer track record and the better-rehearsed defense. The unit that needs it has a promising idea and a thin spreadsheet. The analysis may lean slightly toward the new; the politics lean heavily toward the old. Only the CEO has the authority to overrule that.¹

The research on this is old but blunt. When McKinsey studied more than 1,600 U.S. companies between 1990 and 2005, a third of them gave each business almost exactly the capital it had received the year before. The third that moved the most, shifting an average of 56 percent of capital across units over the fifteen years, earned on average 30 percent higher total returns to shareholders each year than the third that moved the least.²

Satya Nadella's early years at Microsoft show both the cost and the payoff. In July 2015 the company said it was moving away from a strategy to grow a standalone phone business, recorded an impairment of about $7.6 billion on the Nokia devices business it had bought, and cut up to 7,800 positions, mostly in phones.³ A division head can't make that decision about their own division. It took the CEO.

Fourth, set the tempo in both directions

Every company has a default speed, usually the speed of its budget cycle and of whatever has been working lately. The strategy may need something else. Sometimes it needs patience the organization doesn't have. Sometimes it needs a pace that scares people. Only the CEO can impose either one.¹

Reed Hastings at Netflix is the cleanest example of both directions. He ran streaming alongside the DVD business for four years while broadband and the content library caught up, then accelerated hard.¹ He also got the pace wrong once. In 2011 a price increase and the plan to spin DVDs off as a separate service called Qwikster cost Netflix 810,000 subscribers in a single quarter, and the Qwikster plan was dropped within weeks.⁴ He didn't change the direction. In January 2016 Netflix switched on its service in more than 130 new countries at once.⁵

That distinction matters more than the example. A mistake in pace and a mistake in direction need different responses. Confusing the two is how good strategies get abandoned after their first bad quarter.

Fifth and sixth, make people see it and hold the change

Uncertainty makes organizations wait for the fog to clear. McKinsey's argument is that analysis can sharpen a point of view but can't supply the conviction to act on it, and that the conviction has to come from the CEO.¹ I put it more simply in the video. Inspiring means getting people to see reality clearly and to share one picture of where you are going. You can hire someone to write the speech. You can't hire someone to believe it for you.

The last role is being the catalyst for change, and it's the one that takes stamina. McKinsey describes how David Cote, at Honeywell, fought what he himself called a form of entropy, the drift toward disorder that a leader has to keep pushing against.¹ I've built much of my work around the same idea, which I call Strategic Entropy™. Left alone, a company slides back toward the way it used to work, and my essay on why Thomas Cook collapsed shows how long that slide can look like success.

An hour with last quarter's calendar

This is the exercise I suggest to leadership teams. It works best done alone, with last quarter's calendar open and nobody in the room to impress.

It's also how I work when I go inside a company through operator-led execution. I bring the framework, the discipline and the uncomfortable questions. I don't bring the conviction, because I can't. That part has to be yours, which is the argument of my book on leading in the age of AI, El Último Turno.