Creating a Data Driven Culture of Thinking: The Y of Y=f(x) + N

 

 

Last month, we discussed the idea of creating a data driven culture of thinking by using the mathematical formula Y=f(x) + N. Let’s dig a little deeper on the left side of the equation, the Y.

 

The Y is the output, also known as the response variable. It must be something specific so we can collect data about its performance on each individual outcome of the process or product. If we (or the customer) value something, we must measure it. If you are not measuring critical outputs, how will you know if things are getting better or worse? When we have no knowledge about something, that’s when things feel unpredictable. Trying to use hope as a strategy is a slippery slope towards chaos, bad decisions and emergencies.

 

Getting the output right must come first if we want to avoid wasting time and resources studying the x. There are at least two things to consider when thinking about the Y.

 

There can be more than one Y and the customer is who gets to dictate what they are. So, we could adjust this first point to say, “Do we have all the correct Ys?” We can easily get excited about a feature or continue to optimize an output from a process well beyond what the customer is willing to pay for.  And although we end up with something great, we can’t run a business if the customer isn’t paying.

 

To be most efficient, always start with the customer. We have both internal and external customers. So, if you don’t interface with the end customer, you probably have some internal customers you talk to - those who use your output as their input downstream in the organization. 

 

Most of the time you’ll find it’s straight forward to define the Y. Sometimes, though, even the customer doesn’t always know what they want or doesn’t articulate their needs. However, it’s still our job to sort this out or there could be big consequences. Seek help if you are needing to venture into the world of market research because it’s certainly a jungle. See how The Center can help here.

 

Now onto the second point. It’s going to be difficult, if not impossible to improve the Y if we don’t have a good way to measure it. How do we know the values we record are good? Just because an instrument is calibrated, does not mean the data you are gathering is useful.

 

A measurement system analysis (MSA) is a useful tool taught in Six Sigma training that allows us to fully validate the measurement system. This can be done quickly and easily (many times within a single day) for both variable data (numbers on a continuous scale) or attribute data (counts, percentages, pass/fail). The noise present in the measurement system, if left unstudied, could lead us to incorrect conclusions like accepting bad outputs, or rejecting good outputs. On top of that, a poor measurement system can make it difficult to find the x to solve your problem!

 

Surely multi-billion-dollar corporations have tackled this, and we can learn from them, right? In practice these concepts are not always easy to accomplish. Have you heard the story of New Coke? A summary is provided in the book Blink by Malcom Gladwell, which highlights the mistake of not getting the output right. I’ll provide a shortened version below, but I recommend you check it out if you haven’t read the book.

 

In the early 1980’s Coca-Cola, the long-standing champion of soft drinks, had been steadily losing market share to Pepsi, the newcomer. Pepsi ran a marketing campaign they called the Pepsi Challenge in which dedicated Coke drinkers were asked to take a sip from two glasses and choose their favorite. Pepsi capitalized on this study because most participants would choose the glass with Pepsi. 

 

Coke of course disputed their findings, but when they privately conducted the blind head-to-head taste test on their own, they had the same results! Coke felt they needed to react or continue to lose market share. So, after applying this sip test to prototypes in further studies, Coke developed a lighter and sweeter “New Coke” that began to take the edge over Pepsi. They then conducted a validation study which included hundreds of thousands of consumers across North America. New Coke beat Pepsi by 6-8 percentage points and the CEO of Coca-Cola said it was the “surest move the company has ever made.”

 

But you don’t see New Coke on the shelves today, in fact Coke had to bring back the old formula after only 77 days due to the outcry from the public.

 

So, what happened? All that data they had wasn’t any good! It was not the correct Y, or at least it wasn’t the only critical Y. The Coke formula was more to the public than just a blind sip test. The brand matters, the packaging matters, the context matters. When was the last time you took a blind sip of a soda in a lab setting? Most of us would say never; we drink the whole can in the comfort of our home knowing that we bought a case of it. 

 

Pepsi is a sweeter drink with a citrus burst. It was born to win in a sip test. But a drink too sweet can become overwhelming and not enjoyable after a whole can. Getting their output wrong is where New Coke failed. This CLT (central location test) wasn’t a good representation of what happens in real life. And although CLTs are an important part of market research to bring standardization, they don’t tell the whole story. Coca-Cola failed to recognize all the relevant Ys they should have been measuring and spent a lot of time and resources optimizing on the wrong Y.

 

Progress and learning can still be made regardless, but it can be a lot less painful if we take the proper time to make sure we get the Y right in the first place. See the summary Coca-Cola provides on their website.

 

Let your questions guide you during the defining steps of a project before you get ahead of yourself and waste a lot of time experimenting. We’re often tempted to jump right into analysis, especially if there is an abundance of historical data available. Don’t fall victim to that. Use critical thinking and make sure you are studying the correct Y. Then validate your measurement system, so you know you’ll be able to collect useful data and answer your project questions.

 

The Center’s courses combine proven problem-solving methods with powerful statistical tools to construct a reliable roadmap to improve business operations. Learn more about the Six Sigma courses we offer here.

 

 

MEET OUR EXPERT: Anthony Welsh, Six Sigma Master Black Belt
Anthony Welsh is a Six Sigma Master Black Belt with 20 years of experience delivering projects to both the automotive and consumer products industries. In his role at The Center, Anthony shares expert tools in critical thinking and data-driven decision making to assist clients with using Six Sigma methods to achieve real results.