PPG Industries is a leading diversified manufacturer that supplies paints, coatings, chemicals, optical products, specialty materials, glass and fiberglass around the world. PPG operates in more than 60 countries and has more than 150 manufacturing facilities and equity affiliates. Last year the company generated sales of $11.2bn and spent more than $7bn on direct and indirect materials to support its operations. To help the company control and leverage its indirect spend, Jim Polak developed and led a corporate spend analysis project. Using tools from Ariba, the resulting sourcing strategy is having a significant impact on PPG's bottom line.

Q: How did you come to head up this spend management project at PPG?

Polak: I've been with PPG for more than 34 years now. During that time I have worked all over the company-as an engineer, in management science, in market research, in customer service, as a plant manager, as a sales manager, as director of marketing and as director of production. So I had all that experience before I became a director in our purchasing department. When I first came to this job in 2000, the things that amazed me were how many processes we had that were still manual and how we had people in all these different pockets across North America acting as independent groups with very little visibility into our spend. This was true globally as well. As a result, we were not leveraging our spend. We would optimize it for businesses or regions, but we were not optimizing it across North America or across the enterprise.

Q: How did you approach the problem?

Polak: The vice president I reported to at the time really gave me a free swing at deciding what we should be doing in terms of automation. That was the height of the internet boom and "e-procurement" was the biggest buzzword in purchasing. Every company wanted to do "e" something. I had the luxury of having several months to really take a look at this and I talked to a lot of people about it. It became clear that if we were going to do this the right way, we needed to start with spend analysis. We were never going to be able to automate transactions until we knew where the money was going and what all of our associates were buying. So we started with spend analysis and we realized some pretty good savings-in the millions of dollars. Then in 2003 we decided we needed a world-class interface in spend analysis, and we opted to go with Ariba. Taking a best-in-class approach, we also decided at that time to go with Freemarkets as our e-sourcing vendor and with a small company in California called Softface for data enrichment. Within six months, though, Ariba acquired both those companies. Basically, what we wanted to do was to improve our ability to automatically categorize data and to generally step up our game in the e-sourcing arena.

Q: In this context, what does it mean to categorize data?

Polak: Well, we have a PPG version of the United Nations Standard Products and Services Code (UNSPSC) that we use to classify both products and services that we purchase. For example, take a commodity like electrical MRO [maintenance, repair and operations]. Instead of having one big categorization with that title, we give different buckets under that category an identifying PPG UNSPSC. This allows us to create a hierarchy of classifications within a big category. Within electrical MRO, for example, you have things like lead ballast, motors, motor repair, switches, wiring, cable and a lot of other things. Some vendors might be able to offer really good pricing in one area but not another. So we identify unique buckets of spend that we can aggregate, optimize and leverage. Typically, as time goes on you get smarter so you can create more finite and more uniquely identified categories of spend. Then you have more opportunities to leverage each one. Electrical MRO is one example, but this applies to office supplies, pipe valves and fittings, contingent labor-all the other categories you can think of.

The key thing about categorization is that it has to make sense to you, because you are using it to dig into your data to tell you what you are buying and where your spend is going. You should not try to adhere to somebody else's idea of categorization. It's fine to use somebody else's structure-at PPG we use the UNSPNC as a basic structure, but we add our own unique features to it.

And we give our buyers, who are experts in a given category, full accountability and full responsibility for creating the levels of categorization within a given commodity. If they want to go two levels down, fine. If they want to go 10 levels down, fine. Because at the end of the day they are responsible for driving spend to the bottom line, so we make them responsible for categorization. They are the ones that are going to use that information to drive savings, so why not let them control it? If they make a mistake either way, as time goes on they will understand that and will create more optimal, better categorizations.

Q: When you start digging into the data, do you often find surprises?

Polak: Yes. Electrical MRO was one of the first areas we attacked at PPG. The experts here thought that we probably had, at most, 40 suppliers in this category and that it wasn't that big of a spend. When we did even a crude categorization and got data from all of the different sources within PPG, we found out that we have more than 300 suppliers and more than $10m of spend in the category. So when we went out and leveraged that, we saved a lot of money in our very first event, even with data that wasn't very good.

I have talked to a lot of different companies about this issue and often the mistake people make is in thinking that the data they get from their archives and ERP systems has to be perfect and, unless it is perfect, they can't do any analysis and can't hold a sourcing event. Nothing could be further from the truth. It's true that data enrichment and data cleansing are never ending tasks, but one of the things I stress to people is that they are not doing spend analysis to create pretty charts or to impress people. They are doing spend analysis because they want to see where the money is going and they want to run sourcing events and leverage their spend among fewer suppliers and get savings for their company. And they can do that even with relatively poor data.

When we started this project, we believed that the savings would be there, but we had a number of people within PPG who had been with the company for a long time and their natural reaction was that we would never realize the savings we were talking about. We would go into a group and tell them that the savings target was for a minimum of 10 percent and they would look at us like we were crazy. But you just have to persevere.

One area where no one thought we could achieve savings was contingent labor because the base wage rate is controlled by our HR department. But then our buyers became experts in overhead and found out that the overhead markups from vendors varied from a low of 30 percent to more than 200 percent in some cases. They were able to use that knowledge to save substantial money for PPG.

Q: Have most of your savings come from leveraging the volume of spend?

Polak: Sure because you have suppliers that are trying to optimize their delivery, their relationships and their logistics and the greater the volume the easier it is for them to save costs on their side.

Again, using electrical MRO as an example, we went from 307 suppliers down to six. So 301 suppliers lost their business with PPG, but six increased their volume and were able to improve their cost structure and give us greatly improved pricing.

Here is my rule of thumb on the savings a typical company can achieve: take 20 percent of your total indirect spend and you can easily save 15 percent of that. So if you spend $1bn on indirect goods and services, 20 percent is $200m and 15 percent of that is $30m. I have used this for five or six years in many presentations and I have never had anyone tell me that's too high a bar. I have had a lot of people tell me that they exceeded that benchmark.

Q: Are there other technologies that have contributed to this project?

Polak: In 2005 we added several additional solutions from Ariba. We wanted to automate our contracts so we implemented Ariba Contract Workbench. We implemented Ariba Buyer to further automate transactions and Ariba Invoice, which enables our suppliers to not only receive purchase orders electronically but also to invoice us electronically. I've always been big on determining what your strategy is for a commodity before you go out and try to source it, so we acquired Ariba Category Management to help with that.

Q: What have you learned from this project that you would like to share with our readers?

Polak: Don't get discouraged if early on the experience within your organization says that what you are trying to do is not possible. You have to believe the savings are there and not let anybody sidetrack you. If you can get a high level of sponsorship and buy in, your job will be a lot easier. Lastly, make sure you have an organization and reporting structure that will make it possible to get things done. At PPG, all the purchasing groups across the country either directly or indirectly reported to me and that made my life easier.

RESOURCE LINK:
Ariba, www.ariba.com