• It's hard to know when they lose their touch. There are two kinds of innovators -- product and business strategy. Steve Jobs has pretty consistently developed market-beating new products at Apple. But he stepped out of management there between 1985 and 1997 -- during which time the company did not do so well. And Google's co-founders -- Sergei Brin and Larry Page -- innovated in search engine marketing -- but it is unclear whether they'll generate a follow up. Meanwhile, Microsoft's Bill Gates was a true business strategy innovator -- very good at capitalizing on competitor's product innovation through a fast-follower business strategy. But he lost his touch after the dot-com crash and his replacement -- Steve Ballmer -- is a janitor who is struggling to catch up with Google.
• They may be more flash than substance. Murray cites HP's Carly Fiorina and AIG's Hank Greenberg as ego maniacal visionaries. Fiorina, though, was more surface than substance. She did not come up with any new product ideas for HP -- just new advertisements. Greenberg, by contrast, was a true innovator who expanded the business by pushing into new markets -- such as China and Japan -- and making bold acquisitions.

Murray misses the point that there are times when janitors are good for shareholders and times when innovators are better. This dents the usefulness of his argument because it's only a matter of time before companies led by janitors will have cleaned up the mess -- driving boards to find innovators who can spur growth. The challenge for investors is to identify the CEO type -- the true innovators and the meticulous janitors -- and to know which company will benefit from each at the right time.

Here's my view on three buys and how they fit:

• Boeing. McNerney is a janitor CEO, which Boeing now needs due to huge demand for its products. Boeing is relatively inexpensive -- its PEG of 1.27 is based on a PE of 30.5 and 23.9% earnings growth forecast for 2008.
• Google. Page and Brin are innovators at a time when new sources of growth are critical -- but will any of Google's experiments pay off? Google is relatively cheap -- its PEG of 1.32 based on a PE of 41.7 and 31.6% earnings growth forecast for 2008.
• AIG. Sullivan is a janitor CEO, and it remains to be seen whether he can spur rapid revenue and profit growth. AIG is not overly expensive -- its PEG of 1.43 is based on a PE of 19.1 and 12.2% earnings growth forecast for 2008.

And here are the three holds:

• Microsoft. Ballmer is a janitor CEO at a time when an innovator would help it compete with Google. Microsoft is fairly expensive -- its PEG of 1.48 is based on a PE of 22.1 and 15% earnings growth forecast for 2008.
• HP. Hurd is a janitor CEO who has squeezed out costs from HP, but it remains to be seen whether he can generate substantial revenue growth. The expensive stock and tepid growth are concerns as reflected in HP's PEG of 1.57, based on a PE of 19.1 and 12% earnings growth forecast for 2008.
• Apple. Jobs is an innovator CEO at a time when innovation is needed. But the stock is very expensive -- Apple's PEG of 2.29 is based on a PE of 31.8 and 14% earnings growth forecast for 2008. But given the potential for upside surprise, I would not count out Apple.

I think this is a useful way to think about investing in CEOs. What do you think?

Peter Cohan is President of Peter S. Cohan & Associates, a management consulting and venture capital firm. He also teaches management at Babson College and edits The Cohan Letter. He owns AIG stock and has no financial interest in the other securities mentioned in this post.