I recently stumbled upon the Wikipedia entry for French entertainer and artist Michael Lotito. Listed among Lotito’s many accomplishments (?) are having eaten 18 bicycles, 7 televisions, and a Cessna 150 airplane over his lifetime – the Cessna took about 2 years. I’m not looking to encourage you to gnaw away at your laptop, but to illustrate that unattainable tasks, or unanswerable questions, can be tackled when broken into small pieces. Lotito broke these strange objects down into micro-sized bites or even a powder in order to consume over them over time as part of his “act”.
In consulting, our clients want us to solve big problems or provide quick answers to complex questions – questions where there may not be an immediately accessible answer for. Sometimes the best we can do is to make an educated guess, but we still want to be sure we are taking the proper steps to be as accurate as possible. Answering questions of such a large scale requires us to make some assumptions – if we are able to walk our clients through these assumptions, the sequence of steps, and the overall thought process, we can give them an answer they’re looking for built on a structure and not just random guess making.
Let’s try to tackle the following question – how many road miles were driven in the US last year? Looking at this question, my initial response is “I have no idea, a LOT? Maybe millions or billions.” Let’s walk through the steps needed to try to give us a much more accurate response in just five minutes.
Frame the Question
1. Think about how accurate you need to be and what time frame you’re dealing with.
Do we need to come up with an answer on the fly? Do we have access to a calculator?
Do we have time to pull in more detailed data points?
What are the components that will make up a correct answer?
2. Let’s assume we need to come up with an immediate answer and we have a basic calculator at hand.
3. We’ll need to break it down using either a population or car count and estimate how many miles that person or car is driving in a year.
Pull in the Pieces
4. Make estimates to set yourself with numbers that are round and easy to work with, if possible
.Don’t be afraid to make assumptions, such as even distributions in some of your demographic estimates.
Think about the demographics of your groups. For a question on driving, we might want to look at only people in their prime driving years (ages 16 to 75). But for a question on the number of pairs of boxer shorts purchased last year, for example, we may only want to look at a male sample.
How far can we break this down? It’s much easier to predict the average annual mileage for one person and extrapolate that out than it is to take a stab at the all-up number right off the bat.
For market sizing in the US, you can feel pretty comfortable with somewhere between 300 million and 350 million people, or 100 million households. If scaling for worldwide, you can use 7 billion as a population number, but be careful to account for wildly different worldwide demographics outside of the US.
5. Let’s use a US population of 300 million. Looking ahead, because the question will require us to breakout by driving age, let’s split this population into age demographics. We’ll use an average life expectancy of 80 years, assuming evenly-distributed age demographics. Further, we’ll use this to estimate that about 25% of the US population is not driving on a regular basis, those under 16 and many over 75.
6. We now have an estimated driving population of about 225 million drivers in the US last year (300 million x 75%).
7. Next, let’s estimate that the average person drives about 12,000 miles/year, which seems to be conventional wisdom. If you don’t have a calculator at hand, think about using a round number like 10,000 miles/year and adding 10% or 20% to your final answer.
Answer the Question
8. Once you’ve made your estimates, lay them out in a basic equation that makes sense. We now have a basic equation of:
225,000,000 drivers in the US x 12,000 miles driven/year, or 2.7 trillion miles driven in the US last year.
At first glance, 2.7 trillion seems like huge number, and honestly, I’d have guessed a much smaller number if I was just taking an immediate stab at the question. A quick internet search reveals the general consensus to be around 3 trillion miles driven in the US in 2013. Not bad!
By setting up the problem with a thoughtful and organized approach, we’ve managed to get within 10% of the correct answer in about 5 minutes and avoided an orders-of-magnitude error that a spontaneous guess may have produced. Have fun estimating!
By Grant Klein
Posted by admin