Your data is telling you where to stand, not what to fix

By Patrick Smith

In today’s manufacturing world we’ve become focused on collecting and analyzing data - creating pivot charts and paretos, moving data into buckets, rearranging so that it can tell us how our process is performing. Although good for initial prioritization, taking this analysis too far can get us into trouble because it distracts us from getting on the floor and starting the actual problem solving needed to improve. As a leader, it is up to you to help your people use data to know where to focus rather than trying to solve a problem sitting in front of their computer.

One example from my past experiences that demonstrated this trap very clearly was during a scoping session for an improvement team. We had financial data that showed some key areas that we’d need to focus on - specific raw material losses and specific losses across individual manufacturing lines. This data was presented to the broad group showing some of the general areas that we’d need to observe and improve. I thought this was great and was ready to huddle up and go out to the floor to start problem solving, but then we starting going further into the data- into individual SKU’s, looking at specific days, looking at various paretos of different ingredients. For 15 more minutes the group continued to jump to conclusions delving deeper and deeper into the data analysis without getting us any closer to understanding a root cause and therefore a solution. The broad data showed us where the problem was, and the next step we should have taken was to get on the line and observe the process and problem in action. The issue for you as a leader is that in this case everyone was ok with continued digging even though there was no value and significant distractions were being created.

Data can be powerful in focusing in on opportunities, but often it will not tell you what to do to improve. We need to add in the element of getting out into the operations, observing the problem, and fully understanding the patterns that are occurring. These often overlooked steps in the process are vital to actually realizing improvements.

So then next time you find your team going down the rabbit hole how can you bring them back on track?

My favorite first step is to observe the problem in real time by completing a line study. Let the data guide you to the biggest problem areas, and then you can observe the process when things are going wrong as well as when things are going right. Looking at both of these cases is useful in highlighting differences between good and bad running. A key tool I use here is to complete a line study looking at run speeds, downtimes, waste, and idle times. This can be done with the classic pen and paper, or with a line study app that we’ve developed to make it easier (available on iOS in the app store today, called LineStudies by Stroud International). I want to focus on pinpointing areas and causes that are visible during the study and validate the data that has lead us to this point.  With the simple exercise of getting out on the floor, you can learn a lot and even solve some problems with the initial observations.

A second step is to figure out the pattern of failure. Here we are looking for patterns in timing, location, personnel, or any other criteria that could cause a problem. This is normally a combination of observing and data analysis that needs to start with solid observations of the problem. Here we can analyze our line study data, or other observational data, to help pinpoint patterns that aren’t visible at a computer. Finding patterns in this sense provides a powerful second step that will solve a significant amount of your daily operations opportunities. In many cases you’ve got to be a bit more sophisticated with how you “watch” to really find the patterns occurring. This can come in the form of video recording at high speed, and then looking at the footage slowed down to understand something at high speeds. Or it could be using multiple synchronized videos to try and recognize patterns in isolated areas.

It can be tempting to continue digging into the data with the hope that a solution presents itself. Instead I’ve always had more success shifting to observing the process and problem while it occurs. When trying to lead your team to success in problem solving remember to ask these two questions:

  • What did our line studies and observations tell us about the problem?
  • What did the pattern of failure tell us about the problem?


Like this article? You might like: 

The True Cost of Guessing in Problem Solving

Big-Project Engineers Have to Deal with Too Much Red Tape

Truck Engines Replaced 80% Faster