Eryc Eyl: Chapter 15: The Old 80-20 Rule or Making Pareto Practical

Add Value and Sell More with a Simple Data Analysis Tool

Last time we talked – instead of my usual practical advice – I gave you a history lesson, some theory and not one bit of homework. By the end of this week’s column, I hope you’ll see that the groundwork of history and theory was necessary to get us here. Strap yourself in – this is going to get hairy and fun.

If you haven’t read my last column about Vilfredo Pareto and his powerful contribution to solving business problems, I strongly encourage you to do so before reading on to find out how to actually apply Pareto’s concepts by making a Pareto chart. For now, I’ll completely oversimplify Pareto’s vast accomplishments in economics with the Pareto principle: 80 percent of problems are caused by only 20 percent of causes.

If you remember our previous conversations about selling stuff, we agreed that it all comes down to asking questions and listening actively to figure out what your customer’s biggest problem is and then solving it.

Let’s say, for example, that your customer is Wikimedia, the company responsible for the hugely popular Wikipedia. The company got a lot of unwelcome attention earlier this year when its data center in Amsterdam overheated, effectively shutting the website down for European users. Your job, then, would be to figure out why the data center overheats, and a Pareto chart could help you.

But since not everyone can relate to overheated data centers, let’s focus on something we can all understand.

Over the past two years, let’s say I’ve been late to work 143 times [Note to my boss: This hypothetical scenario is for learning purposes only and should not be included in my annual review]. Fortunately, I’ve had the immense foresight and perspicacity to keep track of my lateness and the reasons for it. My tally sheet looks something like this:

As you can see, I’ve sorted my tally of reasons for lateness in descending order by frequency. This is a crucial step in creating a Pareto chart in a simple office tool like Microsoft Excel. I’ve also added a column that figures out each reason’s contribution to the total, e.g. not hearing my alarm accounts for 22% of the 143 times I was late to work.

I’ve also added another column to track the cumulative contribution. This will be very important as we apply the Pareto principle to solving this problem.

As you can see, there are 10 separate reasons why I was late to work those 143 times. However, only three to four of them account for 80% of the times I was late. This might be made clearer with another picture. Here’s that same data, represented in graphic form:

Like Facebook junkies, server logs freely spit out endless amounts of data about themselves that will help you get to the bottom of this, so you won’t be fumbling around in the dark on this one. If the data center overheated 40 times in the past five years, you’ll probably find 10 different causes for the overheating, but only two of those causes were behind 80 percent of those situations. This is the Pareto principle in action.

Once you’ve figured out, hypothetically, that unnecessary read cycles on a tape drive and faulty power management on a web server have caused 32 of the 40 overheating incidents – allowing you to ignore things like that one really hot day in Amsterdam or the time a trash can caught on fire [Please note: I am totally making these things up. I have no reason to believe there’s anything wrong with Wikimedia’s tape drives or trash cans.] – you can start focusing the conversation on solutions that will truly make a difference for the customer.

Your challenge, should you choose to accept it, is to go out and start gathering some data about a problem that you or your customer is currently facing, and plug that data into a Pareto chart. Maybe it’s the number of RMAs you’re getting or the number of deals you’re losing to the competition. No matter what it is, you can easily start zeroing in on the key drivers by tallying up the causes and listening to Vilfredo Pareto.

By the way, if you’d like assistance creating a simple Pareto chart in Excel like the one shown above, don’t hesitate to contact me. I’m always happy to help. You can find me at SuddenSixSigma@gmail.com.