Executive Briefings

Better Forecasting, S&OP Support Transformation At Campbell's Soup

A conversation with Mike Mastroianni, vice president of North American planning and operations at Campbell's Soup Co.

Before joining Campbell's three years ago, Mastroianni spent 14 years at Nabisco, another big-brand company in the food industry. There, he worked in financial, manufacturing and business development positions before getting into logistics. At Campbell's, Mastroianni is responsible for overall operations support for the North American manufacturing group. In that capacity, he led an initiative to establish a Sales & Operation Planning process that has been instrumental in Campbell's enterprise-wide transformation project. Mastroianni also has responsibilities for demand- and supply-side planning, distribution and continuous improvement within the plants. He acknowledges that this last function makes his job "a bit of a hybrid" but he likes the opportunity to stay close to the plants.

Q: What does Campbell's supply-chain look like?
Mastroianni: The North American unit has soups, sauces and beverages and generates around $3b annually in sales. We have a handful of major plants in the U.S. that we call jumbo plants and we also have a series of vertically integrated operations. So it is a combination of traditional manufacturing plants with a sizeable warehouse bolted on and a series of vertically integrated facilities.

We manage around 4,500 SKUs. There are about 850 specific products, but by the time you take into account all the different sizes and so on, you are talking about 4,500 SKUs. Our brands include soups in the familiar red and white cans, but also Swanson broths, Franco-American products, Pepperidge Farm bakery items, Prego and Pace sauces, and the V-8 line of beverages. It's a diverse product line, which is one of our challenges, and also an opportunity.

Q: What were the primary "pain points" in the supply chain when you came to the company?
Mastroianni: The majority of issues that we were dealing with as a company actually were upstream and not within the four walls. By upstream I mean forecasting and supply planning. This was critical because about two and a half years ago the company launched a transformation plan that meant we were going to be changing at a fairly rapid pace relative to innovation and focusing on the customer. So we absolutely had to advance our sales and operations planning capability. I was asked to lead the initiative to put an S&OP process in place.

Q: What was driving this transformation strategy?
Mastroianni: The gist of the plan was to essentially step up the innovation - introduce more new products and get them to market faster. Up until the transformation plan was announced, Campbell's had been experiencing a sales line that was fairly flat to declining. We knew that to grow the business, we absolutely had to step up innovation. At the time we were basically a condensed soup company, but there was a lot of pressure on the consumer side of the house for more convenience. So that was an area where we had to step our game up considerably.

What that meant to us in the planning and execution side of the house was that there certainly were more SKUs on the horizon. Our overall planning capabilities were going to be significantly challenged because we were, as I said, in a fairly stagnant environment. Competencies within our planning organizations were OK for a stagnant environment, but they were not adequate for a dynamic environment - that went for systems as well as processes and people. We had to address all three of these areas and get them all in line. So what we were facing was a significant increase in new products and more SKUs overall, because people never want to let go of the old ones. And that places a tremendous amount of pressure on inventory control as well.

Q: What systems changes did you make?
Mastroianni: Well, we did a lot of work on the systems side just to better utilize the tools we had - that was the first thing. We were and are still using Manugistics, but we were not using it well. The system is capable, but we needed to get into all the planning parameters, to educate people, to optimize the parameters and to maintain them. We needed to upgrade skill sets and, in some cases, we had to put some new bodies in place.

We also have been in pilot for more than a year with the real-time forecasting tool from Terra Technology. Terra was a partner that we had brought in to work on some other planning tools for us, and they recognized very quickly what a great opportunity it would be for us if we could get our arms around the demand picture earlier in the process. We have been running this application every week - every day, really - for over a year and we have consistently seen improvements of about 50 percent in short-term forecast accuracy.

That is huge because there really are two components here. Sales and Operating Planning looks a little bit further out, by design. Typically S&OP looks two to three months out as a starting point and then goes on out to 18 or 24 months.

But here is where we wanted to go a step further. We wanted to get more precise in that one-to-four-week horizon. I think most people in this business will tell you that is where they struggle the most. That is where the pain is. So we wanted the ability to systematically adjust to the difference between actual and forecast.

What the Terra technology tool has done beautifully for us is to bring some very robust and agile algorithms to solve this short-term problem. Specifically, when these guys really got under the hood, they saw that we had a lot of data we were not leveraging. We have tremendous data history relative to how our customers perform. We had the Manugistics forecast. We had Vendor-Managed Inventory data. We had all kinds of other causal factors, all in house. And our customers are placing orders all during the week. So at any given point, we can calculate what percentage of orders are currently in hand. Then we know what is left out there, still to be determined.

So the Terra guys said, "Let's not start from scratch here. You have a lot of information already blocked and loaded." For next week's shipments, for example, we probably already have in house 50 percent to 75 percent of our orders. So let's use it. Use it but be real smart about it, because there still is that other 25 percent that we have to put some math to.

What these algorithms do is really focus in on the short term. They take the information that we have and work with it. Then, if we integrate it into our execution systems, we will have immediate visibility on the shop floor every day when we run our batch. So that is where we have applied it. That is the 50 percent improvement in forecast accuracy. And that 50 percent includes a lot of new product launches last year, so the tool has performed extremely well.

Q: Was one of the goals of this project to improve stock-outs?
Mastroianni: Stock-outs are an issue for every company. Here is the challenge: It is one thing to achieve a high level of precision and accuracy. The second issue is how do you execute this on the shop floor? How do you meaningfully take advantage of this early visibility to demand? You need to have an agile environment, but you also need to have an environment that puts customer service very high on its priority list and that allows you to expedite or de-expedite in a cost effective manner. You also need to have the right metrics in place and the right objectives so that plants are not penalized for changing over more frequently.

Q: Are there other initiatives you are working on that you can share?
Mastroianni: An issue that has been really big for Campbell's is the work we have done over the last year to de-compose, at a very deep level, the drivers of our inventory. That has given us the visibility to act on some of those drivers in a very specific way. For example, if we can specifically quantify by SKU, by location, by manufacturing line, what our cycle-stock inventory is, we can implement programs to reduce our cycle stock. We can launch a lean initiative, we can possibly spend capital against changeover parts, etc. Before, we never had that kind of granularity. So we have invested in this type of analysis and we actually are using Terra Technology there as well. So de-composition analysis of inventory is very important for us. It is all about agility at the plant and really getting focused on that agility by understanding the drivers.

Q: Did this analysis turn up any surprises?
Mastroianni: Yes there were some ah-hahs. Specifically, we had one quality-driven initiative that resulted in our single-sourcing one group of SKUs. It was the right thing to do and everyone was excited about it. But with this granular analysis, we saw that the inventory from this group jumped off the radar screen. As we got under the hood, we found that it was all about the implications of single sourcing: more lead time, more safety stock in the system, and cycle stock increased significantly because of new requirements on the quality side. So the "ah-hah" there was that on any project going forward, I will guarantee you that we will do a de-composition analysis and impact study to understand those dynamics.

Overall, the value of the information we gained from comprehensive, deep-dive de-composition of inventory drivers has fundamentally changed the way we are managing inventory because it touches everything. It touches how you execute on the plant floor, it impacts your capital investment, it impacts commercialization, capacity planning, new product launches, etc. It impacts everything we do. Visibility at that level has changed the game for us.

Before joining Campbell's three years ago, Mastroianni spent 14 years at Nabisco, another big-brand company in the food industry. There, he worked in financial, manufacturing and business development positions before getting into logistics. At Campbell's, Mastroianni is responsible for overall operations support for the North American manufacturing group. In that capacity, he led an initiative to establish a Sales & Operation Planning process that has been instrumental in Campbell's enterprise-wide transformation project. Mastroianni also has responsibilities for demand- and supply-side planning, distribution and continuous improvement within the plants. He acknowledges that this last function makes his job "a bit of a hybrid" but he likes the opportunity to stay close to the plants.

Q: What does Campbell's supply-chain look like?
Mastroianni: The North American unit has soups, sauces and beverages and generates around $3b annually in sales. We have a handful of major plants in the U.S. that we call jumbo plants and we also have a series of vertically integrated operations. So it is a combination of traditional manufacturing plants with a sizeable warehouse bolted on and a series of vertically integrated facilities.

We manage around 4,500 SKUs. There are about 850 specific products, but by the time you take into account all the different sizes and so on, you are talking about 4,500 SKUs. Our brands include soups in the familiar red and white cans, but also Swanson broths, Franco-American products, Pepperidge Farm bakery items, Prego and Pace sauces, and the V-8 line of beverages. It's a diverse product line, which is one of our challenges, and also an opportunity.

Q: What were the primary "pain points" in the supply chain when you came to the company?
Mastroianni: The majority of issues that we were dealing with as a company actually were upstream and not within the four walls. By upstream I mean forecasting and supply planning. This was critical because about two and a half years ago the company launched a transformation plan that meant we were going to be changing at a fairly rapid pace relative to innovation and focusing on the customer. So we absolutely had to advance our sales and operations planning capability. I was asked to lead the initiative to put an S&OP process in place.

Q: What was driving this transformation strategy?
Mastroianni: The gist of the plan was to essentially step up the innovation - introduce more new products and get them to market faster. Up until the transformation plan was announced, Campbell's had been experiencing a sales line that was fairly flat to declining. We knew that to grow the business, we absolutely had to step up innovation. At the time we were basically a condensed soup company, but there was a lot of pressure on the consumer side of the house for more convenience. So that was an area where we had to step our game up considerably.

What that meant to us in the planning and execution side of the house was that there certainly were more SKUs on the horizon. Our overall planning capabilities were going to be significantly challenged because we were, as I said, in a fairly stagnant environment. Competencies within our planning organizations were OK for a stagnant environment, but they were not adequate for a dynamic environment - that went for systems as well as processes and people. We had to address all three of these areas and get them all in line. So what we were facing was a significant increase in new products and more SKUs overall, because people never want to let go of the old ones. And that places a tremendous amount of pressure on inventory control as well.

Q: What systems changes did you make?
Mastroianni: Well, we did a lot of work on the systems side just to better utilize the tools we had - that was the first thing. We were and are still using Manugistics, but we were not using it well. The system is capable, but we needed to get into all the planning parameters, to educate people, to optimize the parameters and to maintain them. We needed to upgrade skill sets and, in some cases, we had to put some new bodies in place.

We also have been in pilot for more than a year with the real-time forecasting tool from Terra Technology. Terra was a partner that we had brought in to work on some other planning tools for us, and they recognized very quickly what a great opportunity it would be for us if we could get our arms around the demand picture earlier in the process. We have been running this application every week - every day, really - for over a year and we have consistently seen improvements of about 50 percent in short-term forecast accuracy.

That is huge because there really are two components here. Sales and Operating Planning looks a little bit further out, by design. Typically S&OP looks two to three months out as a starting point and then goes on out to 18 or 24 months.

But here is where we wanted to go a step further. We wanted to get more precise in that one-to-four-week horizon. I think most people in this business will tell you that is where they struggle the most. That is where the pain is. So we wanted the ability to systematically adjust to the difference between actual and forecast.

What the Terra technology tool has done beautifully for us is to bring some very robust and agile algorithms to solve this short-term problem. Specifically, when these guys really got under the hood, they saw that we had a lot of data we were not leveraging. We have tremendous data history relative to how our customers perform. We had the Manugistics forecast. We had Vendor-Managed Inventory data. We had all kinds of other causal factors, all in house. And our customers are placing orders all during the week. So at any given point, we can calculate what percentage of orders are currently in hand. Then we know what is left out there, still to be determined.

So the Terra guys said, "Let's not start from scratch here. You have a lot of information already blocked and loaded." For next week's shipments, for example, we probably already have in house 50 percent to 75 percent of our orders. So let's use it. Use it but be real smart about it, because there still is that other 25 percent that we have to put some math to.

What these algorithms do is really focus in on the short term. They take the information that we have and work with it. Then, if we integrate it into our execution systems, we will have immediate visibility on the shop floor every day when we run our batch. So that is where we have applied it. That is the 50 percent improvement in forecast accuracy. And that 50 percent includes a lot of new product launches last year, so the tool has performed extremely well.

Q: Was one of the goals of this project to improve stock-outs?
Mastroianni: Stock-outs are an issue for every company. Here is the challenge: It is one thing to achieve a high level of precision and accuracy. The second issue is how do you execute this on the shop floor? How do you meaningfully take advantage of this early visibility to demand? You need to have an agile environment, but you also need to have an environment that puts customer service very high on its priority list and that allows you to expedite or de-expedite in a cost effective manner. You also need to have the right metrics in place and the right objectives so that plants are not penalized for changing over more frequently.

Q: Are there other initiatives you are working on that you can share?
Mastroianni: An issue that has been really big for Campbell's is the work we have done over the last year to de-compose, at a very deep level, the drivers of our inventory. That has given us the visibility to act on some of those drivers in a very specific way. For example, if we can specifically quantify by SKU, by location, by manufacturing line, what our cycle-stock inventory is, we can implement programs to reduce our cycle stock. We can launch a lean initiative, we can possibly spend capital against changeover parts, etc. Before, we never had that kind of granularity. So we have invested in this type of analysis and we actually are using Terra Technology there as well. So de-composition analysis of inventory is very important for us. It is all about agility at the plant and really getting focused on that agility by understanding the drivers.

Q: Did this analysis turn up any surprises?
Mastroianni: Yes there were some ah-hahs. Specifically, we had one quality-driven initiative that resulted in our single-sourcing one group of SKUs. It was the right thing to do and everyone was excited about it. But with this granular analysis, we saw that the inventory from this group jumped off the radar screen. As we got under the hood, we found that it was all about the implications of single sourcing: more lead time, more safety stock in the system, and cycle stock increased significantly because of new requirements on the quality side. So the "ah-hah" there was that on any project going forward, I will guarantee you that we will do a de-composition analysis and impact study to understand those dynamics.

Overall, the value of the information we gained from comprehensive, deep-dive de-composition of inventory drivers has fundamentally changed the way we are managing inventory because it touches everything. It touches how you execute on the plant floor, it impacts your capital investment, it impacts commercialization, capacity planning, new product launches, etc. It impacts everything we do. Visibility at that level has changed the game for us.