Bag of Chips (part 1 of 3)

This article was written in 2006 as an essay for product managers at Genie.  Although originally titled Conducting Customer Surveys, due to a formatting error, it took on the title of the first section, A Bag of Chips.  It continues to be timeless in its application, philosophy and approach to understanding customer wants, needs, and market opportunity.

A Bag of Chips (Introduction)

I decided I needed to, as the new Engineering Team Leader, to get a feel for our customers.  I was new to the company, the market, the team, and our dealers.  For reasons which completely escape me now, within a couple weeks of starting my new role I was driving across Texas stopping at 10 different customers trying to see our products through their eyes.  I wanted to understand the “Voice of the Customer” so I could remake our line to be exactly what they wanted.  Halfway between Abilene and Waco I realized there was no voice of the customer, there were voices of customers and they said many different and contradictory things.

I was standing in the hot sun on a dusty parking lot somewhere in Central Texas next to a small scissor lift.  The rental branch manager was not happy.  “Take this back!  Your paint doesn’t stay on!”  Being the good engineer, I dutifully wrote his words on my yellow pad so I could “take it back”.  I didn’t get it.  The increasingly agitated manager turned to his sales person and asked him to get a large Ziploc bag.  What could we possibly need that for?

“Hold this!” says the branch manager to me with the gallon-size bag pushed forward.  Somewhat stunned, I did as he started to pull blue paint chips off the machine.  On and on it went in the hot Texas sun until the bag was completely full. “Now, take this back with you!”  I had just received my first clear message.

Over the next six years, as we put together an all-new Product Management Team, we would learn that there is no single customer voice and clear messages are all too rare.  Every customer has different approaches, wants, and needs.  They tell us what they are feeling or wanting right now, not some scientific study of their business needs over the past year.  Some are excellent sources, analytical people who have been in the industry for decades carefully putting together and weighing their comments. Others are narcissists who are simply in our business to make money with little interest in our products or their customers.  Every one is different and unique.

How do we pull all of these desperate inputs together into a single specification for a new product or product modification?  The answer is not a simple algorithm.  If it were, you would be buying a piece of software called “Customer Input Version 5.3”. Instead, it requires smart, dedicated, and open minded people to go out, spend time with their customers and listen. In the end, a series of judgment calls and compromises produce a specification that will maximize the business performance of your new product.   All customers are special, but their products don’t usually need to be.

This paper is intended to convey a set of principles we put together over a four year period.  It looks at customers as individuals with a variety of traits and needs. Customer input is treated like groups of colored cards.  Our customers metaphorically give those cards to us and with those colors we can create a variety of pictures (new products).  There is no one right answer.  Instead, it is our role to determine which picture (product permutation) is best for our business.  The critical elements in this process are:

  • Brief, well defined surveys
  • Well planned customer visits
  • A team of open-minded, listeners who conduct the survey while evaluating each piece of evidence as it is presented.

One caveat, before we proceed, we use this process for medium volume industrial products with a relative few number of customers.  We do not have the luxury of giving up some of these customers, they are just too big, nor is there ever enough input to be statistically significant.  In fact, it turns out that if you have a customer who provides 40% of your business, they want specification A, and you can’t afford to shrink 40%, a sample of one just became statistically significant. Welcome to the world of industrial products (if you manufacturer toilet paper, this might not be the process for you, for a whole number of unpleasant reasons).


15 Minutes of Fame (Types of Surveys)

If you talk to enough people about understanding customers, you will learn that you need to ask about 150 detailed, open-ended questions at each interview.  You have time for about five.  Our customers are busy, they make time to see us but don’t have time to spend the day with us.  If you do take up 3 hours of their time, they won’t welcome you for the next visit.  We need to give our customers time to talk about what they want to talk about, not just what we do.  We plan on 45 minutes with each customer, when you include introductions, discussions, and seeing what they want to show us, we have about 15 minutes to get through the survey.  Make that five questions, no more.

There are some key criteria to look at when creating the questions for a survey:

  • Know what you need to know: It seems obvious, but the most important first step is to discuss within your organization what you need to know and prioritize those needs. Then ask the top 5 questions.
  • Don’t ask if the answer doesn’t matter: If you can’t provide magnesium wheels, don’t ask your customer if they want magnesium wheels.
  • Make the questions objective: Don’t lead the customer in the question (“Would you like the machine to be an industry-leading 33” wide or a specification dog at 34” wide?”).   Make the question simple then ask why (Would you prefer a 33” width or 34” width?  Why?)
  • Be specific. Don’t ask a general question (“How tall do you want the machine?”) if you only have a couple of choices (“Would you prefer a 45’ machine or a 48’ machine?”)
  • Always end the survey with a general, open-ended question: “What can we do better for you?”; “What product could we make for you that we currently do not?”; “What is a need your customers have that you currently can’t meet? How can we help?”


With only 15 minutes to conduct your survey and 5 carefully chosen questions, you really need to understand and narrow down your purpose.  To think that all of the questions for a major new product development program can be asked and answered in one survey is pure poppy-cock fantasy. It is more than just the number of questions.  As design programs progress the questions and potential trade-offs are better understood. Later surveys let you use your customers to determine your design compromises.  We typically see three types of surveys for the three phases of a design:

Survey #1:  Voice of the Customer:  This survey consists of mostly open-ended questions and is looking for the voice of the customer (“easy to set up”, “intuitive to operate”, “I can understand how to operate it”, “is smooth to drive”).  You will use this data to create the product features but not necessarily the actual specifications.

Survey #2:  Finding Specifications:  In this survey you will attempt to determine what tradeoffs will your customers make and what individual specifications should be (“would you pay $100 more for a cup holder?”).

Survey #3:  Choosing Alternatives:  In this survey your design team has come up with two-three very specific alternative configurations, models, or specifications sets.  You will ask your customer to choose between them.