Page 115 - The McKinsey Mind

04 (083-102B) chapter 4 1/29/02 4:50 PM Page 90

90 The McKinsey Mind

I actually had an analyst run lots of numbers from many dif-
ferent sources and then come to me and say, “Well, here’s the
answer.” I took one look at the numbers and said that can’t
possibly be right, because if it were, the world would look a
whole lot different. So, when you’re analyzing the data, just
be sure that you’re stepping back from it and doing a high-
level sanity check.

—Bill Ross, General Electric

I always ask, “How far off would our current answer need
to be before we change our conclusion?” I push very hard on
testing assumptions by making sure the drivers of those
assumptions are very clearly identified. I then focus the
analysis on these drivers. This has fundamentally improved
our acquisition strategy; the results of our recent acquisitions
speak for themselves.

—Ron O’Hanley, President, Mellon Institutional Asset
Management

Although there’s no one best way to do a sanity check, asking
a few pointed questions about your analyses before you put
together your big presentation can save you a lot of trouble.
Remember that there are limits to analysis. Analysis plays a
vital role in the McKinsey problem-solving process, but when all
is said and done, it can take you only so far. You have to draw
inferences from the analyses; they won’t speak for themselves.
You’ve reached the point in our consulting model where intuition
takes the lead from data. You’ve come to Mr. Berra’s fork in the
road, and you have to take it.