Page 112 - The McKinsey Mind

04 (083-102B) chapter 4 1/29/02 4:50 PM Page 87

Interpreting the Results 87

half an hour before you leave your desk to put it down on paper—
nothing fancy, just a hastily sketched chart or a few bullet points
will do. This exercise will help you push your thinking. Whether
you use that chart or not, once you’ve drawn it, you won’t forget
it. Otherwise, the brilliant insight you had this morning might get
lost by the time you lock up your desk tonight.
Don’t make the facts fit your solution. You and your team may
have formulated a brilliant hypothesis, but when it comes time to
prove or disprove it, be prepared for the facts and analyses to
prove you wrong. If the facts don’t fit your hypothesis, then it is
your hypothesis that must change, not the facts.

LESSONS LEARNED AND IMPLEMENTATION
ILLUSTRATIONS
When interpreting your analyses, you have two parallel goals: you
want to be quick, and you want to be right. Obviously, these two
goals are sometimes in conflict. It’s usually worth taking an extra
day if that will make the difference between getting the right
answer and the wrong one. However, as we discussed in Chapter 2,
there’s probably little point in spending an extra week to go from
three decimal places of accuracy to four.
The results of our survey of McKinsey alumni led us to draw
the following conclusions about data interpretation:

• Always ask, “What’s the so what?”
• Perform sanity checks.
• Remember that there are limits to analysis.

Always ask, “What’s the so what?” When you put together
your analysis plan (as we discussed in Chapter 2), you were sup-
posed to eliminate any analyses, no matter how clever or interest-
ing, that didn’t get you a step closer to proving or disproving your