Managing Automation - December 2008 - (Page 24)

[ COVER STORY ] EXPERTS offer FIVE IDEAS that IDEA 1 USE THE RIGHT FORECASTING TECHNIQUE AND TOOLS FOR THE JOB. Too often, manufacturers unwittingly make demand forecasting more complex than it needs to be by taking a one-size-fitsall approach. In fact, Deloitte’s John Simrose says, different forecasting jobs should be approached differently. Most manufacturers, for example, don’t need to bother with data at a detailed level, down to the specific SKU or retail location, when they’re developing longer-term demand forecasts — longer than 18 months in advance of production, adds S&OP consultant and author Tom Wallace. Instead, planning that extends outside the so-called planning time fence — the point in time when materials need to be lined up for production — should be done at the aggregate level and should factor in inputs from sales and marketing organizations. which often have an understanding of longer-term trends. “Beyond the planning time fence, too much detail and too much statistical forecasting just gets in the way,” Wallace says. “There are just too many factors mitigating against the future being like the past.” Best-in-class demand planning organizations routinely use different planning tools and techniques for long-term and short-term forecasts, or for product groups with different characteristics. Cable manufacturer General Cable Technology Corp., for example, has three product groups, each operating in markets with different customer requirements. The $6 billion company’s utilities unit must emphasize customer service. That means no out-of-stocks. But the company’s electrical distribution and aftermarket business must balance acquisition and inventory carrying costs, says Al Rebelo, director of demand planning and customer integration. General Cable manages to meet those different planning needs, in part, by using a demand planning tool from Smart Software that supports a range of analytical algorithms and methods, such as intermittent demand forecasting. Offerings from other vendors, including SAP, Oracle/Demantra, and JDA, also easily support a wide range of forecasting techniques and algorithms. Other manufacturers, such as Procter & Gamble, are beginning to opt for dedicated bestof-breed tools for specific demand planning tasks, such as highly complex short-term forecasting. COLLABORATE TO PRODUCE A SINGLE FORECAST THAT THE ENTIRE ORGANIZATION BUYS INTO. At 70% of manufacturing companies, different functions — such as operations and sales — are still creating their own demand forecasts, according to a recent survey by AMR Research. This undermines demand planning in a number of ways, experts say. First, it tends to erode forecast accuracy since functionally driven forecasts are usually heavily biased, depending on which department is conducting them. Multiple forecasts also lead to a duplication of effort. And, worse, they can easily derail important business processes, such as S&OP. “When different functions come to the table with their own forecasts, what you end up with is a lot of arguing over whose numbers are right,” Wallace says. “Having a single number is an important part of implementing S&OP and, ultimately, running the business.” But how do you gain cross-functional consensus on a single demand forecast? A good start, experts say, is to pull all of the demand planning into a center of excellence that can more easily root out biases and tie demand planning to the overall enterprise strategy, not the interests of a single function. That’s what Welch Foods Inc. did as part of its push to improve forecast accuracy and cut unnecessary inventory. Before implementing Demantra, the company had all the different functions doing their demand plans, profit analysis, and trade promotion analysis independently, usually using spreadsheets. The result was inter-function disagreement over forecasts and “tens of thousands of hours per year of duplicated effort just assembling the data,” Welch’s CIO Ray Gosselin says. Welch’s replaced that with a centralized demand planning organization, reporting into supply chain management, that uses a standard, monthly process to gather demand data from multiple functions and assemble a single forecast. Demantra lets Welch’s slice, dice, and present the results to different functional groups in the ways they need it. manufacturers can implement to improve their demand forecasting. IDEA 2 planning process is not tightly integrated with other processes that can actually do something with the information, such as shape and respond to demand. On the front end, experts say, demand planning should be linked closely to processes such as vendor-managed inventory (VMI), which can provide critical demand trend information. “While many companies are in VMI relationships, they do not use VMI to improve demand sensing,” wrote AMR Research analyst Lora Cecere in a recent report. “VMI usually operates in an island disconnected from other demand systems.” At the same time, it’s important to close the loop between demand forecasting and execution processes, such as sales, production, and inventory management, experts say. “We see a lot of companies not doing a very good job of tying together the demand forecast with other activities, such as the capacity plan,” Deloitte’s Simrose says. At Welch’s, the new demand planning organization, as part of its monthly planning process, gives the company’s sales group specific feedback on what SKUs should be more aggressively sold, given demand trends. “Creating that feedback loop is an important part of making the most of the aggregate demand information we’re pulling together,” Gosselin says. Software tools and related methodologies can support and reinforce that feedback loop. Supply chain management software vendor i2, for example, has developed what it calls Process Playbooks, a set of cross-functional prescriptive steps that manufacturers can implement when specific demand/supply problems arise. Similarly, vendors such as Optiant and Kinaxis offer tools that link demand and supply planning. Optiant, for example, uses probabilistic math from demand forecasts to determine how much inventory should be where. IDEA 4 IDEA 3 LINK DEMAND PLANNING WITH OTHER KEY PROCESSES. Even a relatively accurate demand forecast won’t provide maximum benefit if the demand DON’T FORGET THE PEOPLE. Effective demand planning organizations require people with a wide range of skills, some of which can be hard to come by. Manufacturers need a plan for finding or developing these skills, experts say. “You need a couple of people who are awe- ma December 24 2008

Table of Contents Feed for the Digital Edition of Managing Automation - December 2008

Managing Automation - December 2008 Contents Take 1 Business Objects Chief Says Union with SAP Meets Objectives After One Year Yes, Emerson, Too, Is in the MES Market Infor Chief Puts Off IPO, Restarts Buying Plans Kronos Now Tracks Shop Floor Machines IQMS Rolls Out User Interace, Other Upgrades Notes Five Ideas for Demand Planning Building on the SOA Blueprint Innovation Now A Team Effort Lean %2B Technology = LEAN^2 Finding Flaws Before They Spread Product Scan Advertiser Index Next

Managing Automation - December 2008

For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.