How To Do Market Sizing
My engineering colleagues think that market size is an easy thing to find. "Just call Gartner." Some times that's easy: they know how many Linux servers or Oracle databases or Microsoft Vista installations because these numbers are available. While Gartner, IDC, and others are good, their numbers may be soft too--only as good as the numbers reported by the vendors. Do you have any suggestions for sizing a market?
Here's my story: use logic.
You've probably heard about interviewing at Microsoft. At some point in the interview they ask an impossible question to see if you can reason. "How long would it take for a bowling ball dropped from a cruise ship to hit the bottom of the ocean?" Let's use similar logic for market sizing. Of course look at the numbers that are available. Look at the associations if they exist. But how many piano tuners live in America? I dunno. How many live in Raleigh? Look in the phone book, calculate the per capita number and apply that to 300 million in America.
There are two more numbers to worry about. How many will buy anything and how many will buy from you?Now you have to justify those two numbers. What research suggests that 10 percent will buy? 20%? I'm always leery of a number that says more than 20% will buy in a single year.
OK, so 10% will buy something. How many will buy from you? Unless you truly have something novel, it's unlikely that you'll get more wins in future than in the past. Therefore, your market share percentage is likely to be the percent who will buy from you.
Take the addressable market times the percentage who will buy something times the percentage who will buy from you to get a realistic customer potential.
Answered By Steve Johnson


