It’s hard to think of the carpet industry as a data-intensive business. But at some level it is. Like most manufacturing industries, harnessing data means reducing waste, improving quality and lowering emissions of waste, water and air pollutants.
To learn more, I recently spoke with David Morgan, Vice President of Manufacturing at Shaw Industries Group, the world's largest carpet manufacturer, a subsidiary of Berkshire Hathaway, Inc. Morgan will be a featured speaker at the 2013 VERGE San Francisco conference.
The interview has been edited for clarity and length.
Joel Makower: Give me an overview of how you use data to improve sustainability in the carpet business.
David Morgan: Let me start with the customer. The customer feels pain through service interruptions, when they don't receive the product they expect on time. They receive pain when the product they receive is not as they expected it.
We have a complete set of data and we continue to monitor our products as they move through our operations. So when a customer is not satisfied with a product, we can identify when and where that product was made and we can go back to our processes, which are multiple steps in different facilities over long periods of time. We call that Product Idea and Traceability. We're able to take that information of that dissatisfied customer back to the point in which we think an issue may have happened and study the process.
Makower: What are some of the technologies that have enabled this? Is it sort of basic, bar-coding kinds of things, or is there a sophisticated amount of analytics that really goes into this?
Morgan: Both. First of all, the bar coding and product idea and traceability is very important because you have to have an understanding of when a certain product was processed in a piece of equipment. If you lose that in a reservoir of inventory — where you don't have that granular data about its past and the time that it spent in the machine and when it was in the machines — you can't go back and correlate the process information with the product. So, absolutely, bar coding and product ID are very important. We've gone from hand-written tickets three decades ago to tickets being printed and bar-coded at the product delivery, at the back of the machines. It has helped us define finer and finer points of variation in our process.
That's allowed us to do that cause-and-effect analysis. I'll give you an example. We went into some color technology sensing devices some years ago; you might refer to them as a camera but I'd say it's more of a color detection device. We have developed some fairly intense algorithms around how to look at color off the back end of a production operation, and we have correlated that with variation that is unacceptable in our quality measures downstream and certainly unacceptable to our customers.
So, rather than a human being having to stand and make judgment calls on color, which is fairly subjective, we've turned it into an objective analysis and we do that online, real time. It's done two things for us. The product that moves on downstream has a higher quality. It meets the standard much more often. Number two, it's allowed us to reduce our waste. And anything that reduces wastes improves the experience the customer will have.
Makower: Let’s talk more about the environmental benefits. You just mentioned waste. Are there other benefits in terms of product takeback because you are able to have this traceability?
Morgan: Absolutely. We use some of our identification we place on the back of carpet to identify that as we take back and collect post-consumer carpet across the marketplace. And that collection is not our product necessarily — it's any product. We have to sort that. I'm sure you're familiar with our Evergreen process, recovering post-consumer carpet and moving it back to production. That process is aided by the identification that we place on the back of a lot of our products.
Now, if it's not our product that went on the floor that we are recovering, then it's not as effective at that point. Then we have to use testing mechanisms to test the face of the product and identify the fiber that's in the face of the carpet. But when it comes to sustainability, rather than move back to inside our processes to come to sustainability, our ability to gather data and look at our processes, whether it be the machine or its consumption of materials – natural resources, energy or water – is greatly aided by watching the process and removing variation from unexpected interruptions in the capability of the machine. A breakdown, in other words. Because every time that happens, we lose capacity.
We use additional energy and we have waste material that causes pain to the supply chain all the way to the customer. And so anything that helps us to do that — and the systems that we deployed, both around monitoring and controlling our process — product ID and traceability equipment, knowledge and history of repair and interruptions — several sets of data that we use to analyze our health of our business and to drive improvement — all of those things are translated through the supply chain to reduce waste, reduce energy, reduce water usage and more reliable delivery. Again, the customer's concerned with “I need the product at a certain time. It has to meet the standard and I need it at the most competitive cost I can get it.” And so those things drive those outcomes for us.
Makower: Isn’t this just good manufacturing? I’m not sure how some of this can be seen as part of the company’s sustainability efforts.
Morgan: I defined sustainability in a broad enough scope that says, “When I can produce the same product and use less energy, that's moving in a favorable direction along sustainability road.” When I can use less water, when I use less raw material, when I send a product to a customer and the customer's satisfied and it fulfills their need — that is much more sustainable than when they're not satisfied with the performance of the product and the product has to be recovered and replaced. Now we've used twice as much raw material, twice as much of everything to fulfill the need of the customer. And that is very unsustainable.
Makower: Where is all this going? What would you like to be able to do over the next few years,?
Morgan: One of the challenges we have is the utilization of all this information. It really takes highly skilled individuals. We've named those data scientists. It's kind of a new term we're talking about and it's an individual that really has three capabilities. One is the ability to mine data out of our data warehouses and recover what is needed and sort through that.
The second is an understanding of statistics. They've gotta be able to determine what is meaningful.
And third, they have to answer the question “So what?” If they find meaningful data and they look at it, what would you do about it? So, they have to have a very intimate knowledge of the process. So, if you have a knowledge of the process, you can sort through the data statistically and understand cause and effect, then you can take this information and really act upon the process and make it better.
I would say our challenge and our mission over the next several years will be to develop those skill sets in our individuals and develop more people leveraging what we now have ability to go see within the process.
Makower: How much of the growth of this for you is based on existing technology that we just need to learn to use better, or are you waiting for the technology to improve?
Morgan: Of the two, I absolutely believe that our biggest leverage point to move forward, to continuing gaining benefit from this space, would be individuals with the skill set to exploit what we now have ability to see. To look at what's inside the process and to look at how that affects our products is going downstream. Are the resources we may be using not necessary? So, that data scientist, if we'll just use that term, is the bigger need I see in the next two to three years.
Makower: Are there any other benefits that come out of this?
Morgan: I think our work in this area has had a positive effect on our quality, our customer complaints, our reduction of energy and our reduction of water usage over the past several years. A lot of those can be seen in our sustainability report. I think as time goes forward, I believe technology will help us drive more use of existing materials in a recycled content manner.
That doesn't move quite as fast as developing our ability to analyze data. And rather than a one- or two-year window, those innovations typically could take three to five years and maybe even longer. Though we're continuing to work in those areas, the breakthroughs and the movements just never are as fast as we'd like.