Speed of Internet-enabled business transactions can fundamentally change the way paper companies function across the supply chain

Paper Industry’s “E” Challenge Is Leveraging Its Information

To understand the challenge, it is necessary to try to predict the changes that are coming from the higher speed of transactions. These changes are fundamental and epoch. We cannot control what we cannot measure, so we must look to see how to acquire the needed information—both in quantity and in form.

We must also understand how to validate and check data before it can be used. And finally, we must understand the analysis methods needed to convert the data to a form that can be used by those who make the decisions.

CUSTOMER-DRIVEN EVOLUTION. How will this Web evolution take place? Customers are in the driver’s seat. Already in the retail world, customers take e-business for granted when they place an order. In real-time, they verify from an Internet order entry screen that there is sufficient inventory to supply the goods, the final cost of the goods and exactly when it will arrive.

What will this look like to the manufacturer?

• Customers access order quality
• Customers access delivery schedule
• Customers access pricing
• Customers place orders for a given quantity
  or
  
• Customers solicit pricing for a desired product and quantity
• Corporations bid on filling an outstanding order from its optimum mill site.

For the manufacturer, this is not always a simple task. First, several companies may own one facility or one company may own several facilities capable of manufacturing the same goods. The answer to “how much product” could be the sum of the amount in the warehouse plus the amount that the mill(s) can manufacture from available raw materials in the requisite time. More and more it means this plus what the traders can get under contract from other mills, and all of this is designed to happen rapidly as deals are made.

FIGURE 1: Sample screen of an e-business site targeted at the chemicals industry.

Every business segment of our industry will be forced into e-business to survive. We must realize that in most ways we will benefit from this rapid exchange of information, but we must also understand the data requirements for all mills in order to make correct business decisions. The data that feeds the cost model at the mills have to be accurate, internally consistent, and available reliably, in real time, and of sufficient detail to support the business practices.

We only need to look towards some of the higher technology industry segments to see the extent of changes yet to be done in our industry. At a minimum, we will see the consumers of paper products drive the market to a new level of real-time pricing. Paper companies will need to be ready to commit to manufacture orders on a daily basis and be able to schedule orders at mills where it is most profitable to manufacture and ship to the end user. Raw materials and power will be brokered in a similar fashion, with individual paper mills as the end user. Product pricing will be based on the real-time balance of raw materials, fixed assets, production, and shipping costs. Gone are the days of large product inventories or price fixing on grades from the largest paper companies.

As corporate paper companies, their mills, their raw material suppliers, the converters and the end customers interact, the “distributor” disappears as the customer becomes more educated. This is a phenomenon knows as asymmetry in marketing, which will further decrease the time it takes goods to get to the consumer.

INDUSTRIAL REVOLUTION, PART II. Most analysts view this vast change in business technology as a second Industrial Revolution. With a PC equipped with the Internet in virtually every home, users have a business enterprise at their disposal. As Microsoft Windows 2000 moves into the enterprise, we are seeing a basic convergence of systems that control the flow of the information—computers, LANs, PDAs, Internet, telephones, wireless communication, and more.

We have seen many improvements in our personal lives from the cars we drive to household electronics to our communications with other people to our personal desktop. As these forces move into the enterprise, there will be a basic change in the way we produce, sell, and ship product. We saw it happen with the banking industry, but these were largely expensive, proprietary systems and it took 20 years because so few people actually understood the technology. Now, with literally millions of developers, installers, integrators, and clients using computers at home and the office, the move to business will be much more rapid and dynamic.

First we have to make sure we are on the same page. What exactly is e-business? Its complexity is most easily shown if we look at the progression of the commercial Internet from a simple web page, to a web application; next to provide a mechanism for e-commerce and then to actually transact e-business from the desktop. The first thing to do is look at the growing layers of complexity as we progress into e-business.

Web publication:

• Simple to use or produce
• Get a request
• Formulate and send a page
• Stateless/asynchronous operation
• Mostly HTML
• Mostly static content.

FIGURE 2: Use of real-time data will allow savvy companies to profit from fluctuations in the marketplace. This chart shows energy prices at an Entergy Exchange site.

Web publication is simple. Most computer users—household and otherwise—have purchased Front Page and built their own web site. The Web is full of numerous places that will host Web sites for free. A web server is a simple device. When the web browser points to a server, it sends a message requesting a page to be sent. Sometimes the message gets there, and sometimes it doesn’t. But when it does, the server formulates a page and sends it back out into the ether not knowing if the requestor is still waiting for it or has moved on or, indeed, if that computer is still on the network.

If the page gets to the requestor, the browser takes the ASCII file (cleverly renamed to HTML so we think it is something new) and renders it onto the screen. It is a loose, stateless, asynchronous operation. Pages are mostly simple HTML and content is mostly static. There are, of course, variations on this theme, but overall, web publication is a very simple request-response situation.

Web application:

• Microsoft document equals “Strategy”
• Uses web protocol as client/server
• Presentation tier (IE)
• Middle tier (IIS/MTS)
• Data tier (SQL Server)
• Backbone of DNA.

FIGURE 3: Numerous customer pressures are driving paper mills and their parent companies to change the ways they do business.

A web application is a significant step increase in functionality. In a paper titled simply “Strategy” a few years ago at the PDC, Microsoft pointed out that a web application is the quintessential three-tier application, and the Internet server was really one of several middle-layer servers in a complex client/server network. Since a web application must support interaction between multiple users and the server, there are issues like validation, security, personalization, and a whole raft of others that were not present with web publication. As the systems become more complex, we see additional issues of scalability, redundancy, and more. A web application puts significantly more stress on the client/server methods of the Internet—significantly more complex than ODBC or even Socket calls over a LAN.

E-commerce:

• Replaces existing business processes

   

• Web-centric

   

• Requires domain knowledge of the business practices and procedures

   

• Allows a free flow of information.

   

E-commerce begins when a web application is combined with the domain knowledge to create the front end for a business process. These software packages are complex for the developers but friendly to the users, providing familiar ways to search, select, browse, and enter information only when needed. A good web application seems to know where a user might want to go next (even when the user does not) and presents these as choices for clicking instead of making the user decipher obscure codes typed into a form.

E-commerce applications wrap or replace the essential processes of commerce. Some common ones are procurement, sales, human resources, timesheets, accounting, inventories, expense accounts, and more. The big difference between e-commerce and a generic web application is the software support complexity factor. It is critical that the developers and integrators of information have the domain knowledge of the business process.

THE CASE FOR E-BUSINESS. Finally, there is e-business. Whereas e-commerce is the “webification” of a business process, e-business is the effort of changing the business processes to take advantage of the web. Industry simply will not conduct business in the style or the speed of the past. The decreased cost of transactions greatly increases the volatility of the market and lowers barriers to entry for new competitors.

What are the distinct advantages of e-business?

• Change the business
• Faster turnaround of product (order to delivery)
• Decrease in response time
• Increase use of mill information
• Opportunity for making the best use of the available equipment
• “No web” equals “No business.”

Compared with e-commerce, there are greater issues of reliability, redundancy, and security. In today’s market, if you were to lose production at a mill due to a catastrophe, traders could sell from the warehouse, change or downgrade other paper grades, or even procure paper from a competitor to fill the order. If the e-commerce web servers go down, you are out of business until they get back up, and you lose many of the customers that try to access your business in the meantime. As businesses start to depend upon these large computer servers, the enterprise will place more value on hardened, secure facilities with redundant power, redundant telephone companies, and redundant HVAC.

So what does e-business look like? Figure 1 shows an example from the chemical industry—a screen from a web site called www.chemconnect.com that supports buying and selling of chemicals. There are sites specializing in pulp and paper, crude oil, refined products, and others. Here we see some HDPE Film Grade offered f.o.b. Long Beach for $0.43/lb and there are two bids, one for full price f.o.b. Long Beach for all of the available material, and one for $0.34/lb f.o.b. India—it is not much of a mystery who will get this material if a new bid does not come in.

Suppose, however, that you have an extruder plant in the Los Angeles area and you need HDPE Film Grade to produce products for your customers next week. Your bid had better be more than $0.43/lb. This is a key point of e-business. Many think it will only lower prices, but in this example, it could cause prices to increase by taking politics out of price increases.

In a presentation, the CEO of ChemConnect noted that when requesting bids (a different page) it was not unusual for a user to get offers from their existing suppliers that are lower in price than the long term contract rates.

FIGURE 4: A diagram of a simple facility shows the complexity of deriving and interpreting accurate data.

THE EFFECTS OF E-BUSINESS. So what will e-business do? Increase prices? Decrease prices? Actually, it will allow everyone to sense the market price for goods in real time. This is evident in the stock markets with e-trading, in chemical trading, and in power brokering. And soon we will see this effect in every commodity and service sold—even pulp and paper.

If still in doubt, there is further proof by looking at what is happening to the price of power. With “privatization,” power is purchased from the generating companies using a bid process, and there are many, many players in this field. To understand why, look at one plot of about 9 months of the power prices in the New Orleans area (Figure 2).

The normal price is a bit over $20/mW hr., but in a matter of a few hours in August, the price climbed to 10 times this number. This could have been caused by an outage, unusually high demand, loss of an inter-tie, or any of many, many other causes or a combination of causes. Because the power infrastructure was never designed for market priced power prices, these spikes are inevitable. For this plot, the two-day opportunity for a 1,000 mW turbine is $5 million.

There are companies today building power plants in areas where they know that sooner or later they will pay for the entire plant during a week long transient. They just look for growth and loaded circuits. With the roadblocks to new construction combined with the ban of nuclear power plants and rapid growth in certain areas, these overloaded circuits are not hard to find. Perhaps that is why there is a two-year backlog for turbines.

If you think this is an unusual circumstance, we have already seen mills curtailing production and routing electrical generation back to the power company because the price of electricity is more valuable on a given day than paper production. Many other paper companies could take advantage of power exchange when rates and consumption is high, but most paper mills do not have the e-business systems necessary to arbitrage their production costs. People that are not in the game will give away these added profits to others while likely holding on to the risk of their own business.

SOME REAL CONCERNS: So what is the real concern? The main goal should be to first collect credible information from all areas in real-time. To understand the importance of this concern, we first must understand what is driving this issue (Figure 3).

There are incredible pressures on every business. At all times, companies live in fear of someone introducing the better mousetrap. Consider what would have happened to Barnes and Nobel if Amazon.com had hit the market when the Internet was in 70% of homes instead of 5%! There would have been no time to recover.

Most companies today do not have the right information in the right form and at a time that it does them any good. They do not know how much it costs them to make a product (real-time costing), margins on the orders as they go out the door, or incremental cost to produce more product. Yet they will have to bid in real time for the business. The data to support e-business must be available in real time, accurate and reliable, of the proper scope and aggregation, and of sufficient fidelity to make all the decisions needed in minutes.

The best description of e-business requirements is that we want to do business at the speed of thought, but we get data at the speed of paper with accuracy at the level of rumor. Businesses of tomorrow will be centered on a trader function. This function is one of the most data intensive in the world. The data infrastructure in a mill must be able to provide its decision-makers with the appropriate data in a form that they can use and understand so that they can draw simple and accurate conclusions. Most traders will know that they should sell at the highest margin and buy at the lowest cost if they only knew the numbers. Any plan to operate without the appropriate infrastructure in place will be myopic.

Let us look at the basic problem. Figure 4 shows a very simple facility, with material flows from two tanks to a large tank. Now consider how to compute the production for a 24-hr period. The problem is that if you look at a real data set you will see things like blocked flow controllers that have slightly negative flows. Should we subtract these or set them to 0? There are missing data. Do we assume a straight line, average, or more of a stair step? There are manual entries and quality data that come in later after the sample is run. How do you multiply a concentration times a flow? There is much redundant information—levels fall and levels rise. How can this be used?

The list goes on, but with this simple example, there are more than 300 different answers to the query as to how much was produced in a day compounded with the various methods one can use to aggregate this data (hourly, daily, shift, month). All methods to treat the data need to be internally consistent.

So therein lies the problem. If everyone comes to the morning meeting with their own data set and these data sets are different, it paralyzes the decision process. Now consider a more realistic application. In a paper company, there could be 10 mills with 25 machines with the capacity to fulfill an order of a given grade. Each mill has daily issues that affect production pricing ranging from power and transportation costs to downtime on their machines due to mechanical problems. Yet the company has to know the lowest cost opportunity to produce the 100,000-ton order of grade X and deliver it to Kalamazoo, Mich. If they cannot price the product competitively and deliver it on time, the customer will merely buy from someone who can. Clearly the task of accessing the information in real time is a challenge.

SO WHAT IS THE ANSWER? The easiest answer is found in validated data. Reasonable people, provided with the appropriate data, will usually come up with the right answer. The problem is that everyone has his or her own outdated, inconsistent data from which any conclusion can be drawn. Also we have to recognize that the Microsoft infrastructure that has done so well for us at the desktop is moving into the Enterprise level. No one likes an IBM 3270 display if they can use a spreadsheet! This gives rise to a good plan of action.

1. Fix the infrastructure: If you have old PC that does not support current operating system technology, outdated LANS, flake routers, no instrumentation, etc., fix it or give up. No modern computer application will integrate cost-effectively with an antiquated environment. It is not worth the manpower to ‘patch’ because you are kept from the best tool set.

2. Acquire all the data: Every application needs data—not just the measurements, but the shipping transactions, order quality specifications, material used in manufacture, operator comments, targets that were being set, failures, etc.

3. Validate the data: Give the users credible data stemming from process data that has been washed of errors. There are many ways to improve data quality. The main method is reconciliation of the data over a time period and over a boundary. This is usually in the form of daily material-balanced data on a unit-by-unit basis. This is similar to a checkbook. Once your balance matches the bank’s, your confidence in the data is higher.

4. Structure the data so that it can be used: Just giving management access to 100,000 data points and their time history with no structure does not work. Try to find the natural structure of the business. Time synchronization of the data is necessary because all attributing process conditions did not occur at the same time in the manufacture of a roll of paper. In manufacturing, there are three different models: physical, product, and process. It is likely that most people and their applications use all three.

5. Put the analysis tools that people need on their desktops: Workers today have no problems with databases, spreadsheets, statistical methods, and graphics. If they do, get new workers.

CAROL JACKSON is pulp and paper industry manager, OSI Software, Inc., Jacksonville, Fla. J. PATRICK KENNEDY is president, OSI Software, Inc., San Leandro, Calif.