CNET también está disponible en español.

Ir a español

Don't show this again

Tech Industry

Commentary: Making manufacturing flexible

By 2008, companies will connect physical shop-floor assets to enterprise applications using "X Internet" technologies that sense, analyze and control events in the physical world.

    Commentary: Making manufacturing flexible
    By Forrester Research
    Special to CNET News.com
    April 9, 2003, 12:15PM PT

    By Jennifer Chew, Senior Analyst

    Today, the manufacturing shop floor is a black box. But by 2008, companies will connect physical shop-floor assets to enterprise applications using "X Internet" technologies--devices and applications that sense, analyze and control events in the physical world.

    The X Internet will enable companies to begin to manage the manufacturing process by exception. When Forrester interviewed 28 manufacturers last year to learn about the flow of information to and from the shop floor, we found that manufacturers:

    • Perceive tangible benefits from shop-floor integration. Companies expect that real-time integration will let them respond promptly to demand variations and make manufacturing adjustments without blowing the cost structure.

    • Are unclear who owns integration. Our interviewees believe shop-floor integration is someone else's job--such as facilities managers or a nameless steering committee.

    • Worry about technical impediments. Companies were uncomfortable with their ability to support shop-floor integration due to a confused enterprise application landscape and technologically unsophisticated shop floor.

    Supply chain dynamics
    Despite our interviewees' prudent optimism about shop-floor integration and associated benefits, most manufacturers still lack top-management support for full-blown integration initiatives. But the top floor must realize that without shop-floor visibility, their organizations will never be able to adapt to new supply chain realities such as these:

    • Outsourcing. As manufacturers like Motorola increasingly turn to outsourcing, the ability to manage multitier product quality, scheduling and costing will decide the winners.

    • Mass customization. Manufacturers like Ford Motor will increasingly face pressure to customize products for demanding consumers. For example, 21 percent of all new cars purchased by 2010 are expected to be built-to-order, up from 5 percent today.

    • Shrinking product life cycles. As the major drug companies face increasing competition from generics and me-too drugs, companies like GlaxoSmithKline will be forced to shrink their development cycles by 25 percent to remain competitive.

    Responsive processes needed
    Forrester believes that over the next decade, existing static supply chains will be replaced by adaptive supply networks. But these sense-and-respond networks won't materialize overnight, and manufacturers will have to crawl toward improved shop-floor transparency and flexibility before they can run in adaptive supply networks.

    Companies must take their first steps toward this goal by connecting their data-rich shop-floor processes to their enterprise resource planning (ERP) and supply chain management (SCM) applications so they can begin to do the following:


    Special Report

    Pervasive computing could
    revolutionize industries.


    • Redirect orders to meet service levels. When capacity at one plant drastically changes because of a three-week unplanned shutdown from a chemical spill, production planners must be able to rapidly decide which other sites they can shift demand to. The more insidious problem is manufacturers' inability to detect quality issues before they become a crisis. Companies like BASF must be able to redirect demand as soon as they spot a budding production quality issue.

    • Allocate capacity to maximize revenues. ERP applications like SAP's R/3 operate with the presumption of infinite capacity--enabling unlimited overscheduling of orders. Worse, lack of integration between plant scheduling tools means downtime for repairs is ignored. But if Wyeth, working at full capacity to produce the smallpox vaccine for the U.S. government, can prioritize shop-floor orders across its limited manufacturing capacity, it could boost profitability.

    • Build to actual demand to eliminate inventory buffers. Companies today rerun their manufacturing resource planning (MRP) or ERP plans nightly or weekly to make sure that the corresponding manufacturing orders meet the demand. Meanwhile shop-floor operators are producing to outdated plans or backward scheduling based on incorrect customer requirement dates until a new ERP plan is generated. To compete with lean manufacturers like Honda and Toyota, companies like General Motors should be able to minimize their inventory buffers by linking customer order management to shop-floor scheduling.

    Flexible manufacturing
    Today's application islands won't help manufacturers react quickly to operational changes. Instead, to expose shop-floor data like individual machine capacity and actual work-in-progress inventory locations, companies will exploit the extended attributes of the X Internet--embedded chips and silicon eyes that can sense, analyze and control the real world. The X Internet will allow companies to curb info lag times, reduce data interpretation errors, and react quickly to changes in supply and demand. The X Internet will enable flexible manufacturing through the use of technologies like:

    • Adaptive agents--to trigger rescheduling. In today's shop floor, Machine A's malfunction would not be caught for several days--not until the finished goods failed quality inspection. Instead, by linking to manufacturing execution system (MES) and ERP applications, as soon as Machine A starts malfunctioning, a resident agent will alert the MES so it can reschedule around the looming failure. In turn, the MES can signal the ERP system that the replenishment order should be shifted to another plant because capacity is maxed out. The result? When the Lipitor granulation machine at a Pfizer plant acts up, the drug manufacturer can catch the problem well before the final inspection.

    • Wireless sensors and tags--to reduce inventory costs. Because companies generally use a disconnected ERP as the inventory system of record, they end up performing costly physical counts or audits. Instead, by tagging their inventory with ever cheaper radio frequency ID chips, companies will be able to track inventory status in real time. So manufacturers like Johnson Controls can minimize recounts, improve planning accuracy and give integrated supply chain partners a view into actual inventory.

    • Web services--to enable dynamic capacity reallocation. Companies today don't know the real cost of any widget rolling off their assembly line. Why? Strategies like activity-based costing allocate overhead based on the number of orders or average labor costs. But with Web services lowering shop-floor integration costs, companies will be able to view both works in process and equipment behavior. The result? By precisely calculating its products' total costs--real maintenance, quality inspection, labor, electricity, water, machine and inventory costs--Siemens can quickly decide which ones to shift to lower-cost facilities.

    Managing by exception
    As companies begin connecting their shop-floor assets to their ERP and SCM applications, they will find themselves overwhelmed with data--unless they define precise rule-based information flows for key stakeholders like plant managers, production planners, line workers, machines, planning applications, execution applications and inventory. These new rules will help minimize human intervention by managing exceptions about these things:

    • Supply variability. When inbound materials fail quality inspection, the purchasing manager should receive an alert that triggers follow-up action with the supplier. With quality management systems like the laboratory information management system (LIMS) tied to ERP applications, a key supply issue can be quickly resolved. Companies like Kellogg--whose Pop-Tarts reputation hinges on the quality of raw materials like flour and sugar--could also avoid costly recalls by spotting inbound quality issues even before the materials leave the suppliers' shipping docks.

    • Operational variability. When a machine goes on the blink, several fixes must contribute to the final solution--like temporary rescheduling of manufacturing orders, the creation and execution of new maintenance orders, the activation of more stringent quality inspection plans, or notification to customers of new availability dates. With integration, companies can expedite problem resolution--and minimize inventory buffers--by making informed decisions about which orders to reschedule or what new dates to offer. So companies like Procter & Gamble can respond faster and better to failures on the shop floor--or avoid them altogether.

    • Demand variability. When a customer suddenly cancels a large order or makes changes to an existing one, planning often has to start again from scratch. Instead, with order management systems connected to MES systems, companies can rapidly adjust to demand fluctuations rather than get caught with massive overcapacity or undercapacity on the shop floor. So contract manufacturers like Solectron can begin optimizing outcomes automatically to redirect orders to the best site.

    What to expect
    Several of the components for building a flexible manufacturing environment are emerging--like Web services and enterprise application integration (EAI) for integration, radio frequency ID chips for inventory tagging and Net-native process control systems for controlling individual machines. But users will hit a wall when attempting to fully integrate the shop floor with today's technology. To move beyond these limits, manufacturers must encourage vendors to do the following:

    • Upgrade their batch optimization engines. Today's supply chain optimization tools from vendors like Manugistics, i2 Technologies and SAP rely on static systems with limited data sources for its calculations. But pushing inventory updates or labor changes at these tools every second will shut down the process. Vendors must shift to adaptive planning engines--fed with live data from agents that monitor the performance of networked manufacturing systems--to react quickly to changes in supply and demand as well as shifts in capacity.

    • Agree on shop-floor interoperability standards. Early adopters of shop-floor integration--like semiconductor or pharmaceutical companies--will hit major roadblocks as they try to tie together systems built with proprietary data models from vendors like Rockwell Automation or Emerson. Users must push ERP, MES and shop floor control (SFC) vendors to agree on data interoperability standards--or expect to spend a bundle on chine.

    Users must also encourage integrators to do the following:

    • Extend skill sets from SCM and ERP to true shop-floor management. Most integrators we spoke with for this report have ERP and SCM implementation skills and manufacturing process skills--but virtually no one ties the two together. Early examples of how to build these combination skill sets include KPMG's partnership with Procter & Gamble to comarket PowerFactore. P&G provides the tool set while KPMG provides the consulting project expertise to deliver a tailored solution for streamlining manufacturing operations. P&G claims to have saved more than $1 billion in operational costs since implementing the tools in 1991.

    • Revive technical integration skills until standards catch up. Other integrators, like PwC Consulting or Deloitte Consulting, can't afford to catch another P&G on their hook, so they must develop skills the old-fashioned way--building integration expertise project by project with early adopter companies like Boeing, Unify,and Corning. Integrators will have to dust off their programming documentation guidelines from the late '80s or early '90s until one-size-fits-all integration standards crystallize.

    Adoption rates
    While Forrester believes that by 2008 most large manufacturers must eventually deploy X Internet-enabled flexible manufacturing techniques to remain competitive, certain industries will move faster than others. These include manufacturers that:

    • Produce high volumes... Asset-intensive manufacturers like Gillette and Georgia-Pacific, which produce thousands of razors or reams of paper per day, need real-time visibility into their shop-floor operations to optimize the use of labor, machine, and inventory assets.

    • ...of customized products. As make-to-stock manufacturers--like Pfizer in pharmaceuticals and GM in automotive--tilt toward greater customization of their products over the next decade, they will be pressured to boost their manufacturing flexibility to accommodate the growing number of custom orders.

    • Require complex quality management... Semiconductor players like Taiwan Semiconductor Manufacturing and specialty chemical suppliers like Rohm and Haas that depend on high-yield factors with minimal quality failures must be able to identify and fix manufacturing snafus--because the cost of doing it after the fact is prohibitive.

    •  ...and operate in a regulated environment. Ford probably did not think of itself as operating in a highly regulated industry before the Firestone tire issue emerged. But as industries like auto, chemicals, consumer goods and pharmaceuticals face ever more stringent traceability, visibility, and reporting requirements, they will be under increasing pressure to identify where and when operational glitches occur--to avoid costly recalls, expensive fines or even executive jail time.

    © 2003, Forrester Research, Inc. All rights reserved. Information is based on best available resources. Opinions reflect judgment at the time and are subject to change.