How to guessetimate

the question and details
The process begins (as always) with making sure you understand the question. What are you
ultimately trying to figure out and in what metric? For example, are you seeking total revenue
per store, market size, or total profits for a company? Make sure you have key details such as
the unit of measurement (i.e., should the answer be in dollars or some other unit?). Ask
questions to clarify key details like the period of time the estimate should cover, i.e., per month,
per day, per year, and if it’s an overall value that is desired or a particular segment, i.e., are we
interested in all of Canada, Ontario only, the entire world? Are we interested in the luxury market
or the value-oriented market or the entire market?
Step 2—Break down the problem, and obtain
needed information
You then take the problem, break it apart to the components you need to (and can) figure out,
and then bring the pieces together. How you break it down will depend on the question. It’s
easiest to think where am I starting in my numbers, where do I have to arrive, and then draw a
path for yourself that breaks the question down. You will see in the examples provided later that
a tree diagram can come in handy again by allowing you to identify what information you need
and work your way through it.
You then identify a driver or proxy you can use, i.e., the number of individuals, households, and
stores, and then either build up the information from there or refine it to shrink down to the
ultimate answer.
Step 3—Make assumptions that you can justify
Your assumptions will be very broad at first—refine them to get closer to the actual value. For
example, often the population, household, or individuals you are counting are not all alike—they
can be subdivided into sub-groups that have different habits; if you’re looking at stores or traffic,
the amount of activity they have will vary by time of day.