• Tweet

  • Mail

  • Share

  • Save

  • Get PDF

  • Buy Copies

  • Print

In the early 1980s, when I was heading up the automotive seating operations at Hoover Universal (afterward caused by Johnson Controls), managers from a Japanese competitor that supplied parts to Toyota asked permission to visit our plant. We agreed, on the condition that they reciprocate and because we believed they would larn little from a brief tour. The visitors spent less than an hour in one of our all-time plants, taking no notes. Eventually, nosotros got to read their tour report, and we were shocked at the detail with which they had described our plant and our technology, right downwards to an accurate estimate of our cost of sales. Meanwhile, our senior managers had visited their Japanese plants and learned next to nothing.

Afterwards that experience, I resolved to train our managers—and myself—to approach a plant tour with an educated eye, one that could discern a plant's strengths and weaknesses as accurately every bit the Japanese managers had read ours. The Rapid Plant Assessment (RPA) process, the tool I've developed over the years to achieve this task, has been used in more than 400 tours of over 150 operations since 1998. The information this tool has given us has influenced activities and decisions ranging from benchmarking to competitor assay to strategic acquisitions. And the results of a bout are available in a day or less, whereas nigh rating systems typically take weeks to complete.

Let me give you an example of how powerful the tool tin be. When I was the CEO of Oshkosh Truck, nosotros were engaged in a highly competitive sale for Pierce Manufacturing, a leading U.S. fire truck maker. Pierce's executives didn't favor Oshkosh as the acquirer, so they simply permitted u.s. one 30-minute tour, afterward hours, of each of their three plants. Only we learned so much nigh their operations in those short tours that we were certain we could cut costs by a few million dollars per yr—for instance, past eliminating materials-handling bottlenecks, consolidating plants, reducing inventories, and running the pigment shop on one shift instead of three. As a effect, we offered a higher toll than the company's financials would take indicated and won the sale.

To be sure, the RPA process is non a substitute for due diligence when you're making an conquering, and likewise, y'all'll consider a wide range of factors when choosing a supplier. Merely managers all too oftentimes ignore visual information in favor of the numbers, and as a result, you might miss vital cues almost an operation's strengths and weaknesses. This may cause you to miss out on a highly desirable opportunity or to enter into a relationship that sounds promising only to discover a performance problem later. You lot tin also apply the tool to your own operations to learn what your plant is telling visitors and where you lot might find opportunities for comeback. Donnelly Electronics, Eaton Corporation'due south Aeroquip Group, Haworth's office piece of furniture plants, a Lockheed Martin division, and Seagate Technology are just a few of the organizations I know of currently using the RPA process in their own lean transformation journeys. I'll lay out the tool beneath, forth with two work sheets that will help you codify and analyze the results of the tour.

A Tool for the Tour

At the heart of the RPA process are two assessment tools for teams performing plant tours. The RPA rating sheet presents 11 categories for assessing the leanness of a institute, and the RPA questionnaire provides 20 associated yes-or-no questions to determine if the institute uses best practices in these categories. Following a bout, team members will capture their observations in work sheets like the two shown on the next pages. There are many quantifiable factors by which to assess performance in the rating canvass's xi categories; an exhaustive list of elements to consider in conjunction with them is located on the Spider web at www.bus.umich.edu/rpa along with other avant-garde assessment tools mentioned in the concluding sidebar in this article.

During a tour, team members will be observing all aspects of a establish's surroundings, talking with the workforce and managers, and looking for prove that the institute adheres to best practices. It's important that squad members not take notes during a tour, considering note taking detracts from picking up visual cues and impedes advice with employees on the plant floor. Instead, each member of the squad is assigned master responsibility for a few categories, and the team should meet immediately after the bout to share impressions and fill up out the work sheets. We likewise recommend that everyone on the team answer the last question, "Would yous purchase the products this operation produces?" (Meet the sidebar "Team Limerick and Preparation" for a discussion of who should be on the tour team and how to set up for it.)

An RPA team is small, usually four or five people, with ane person designated as the leader. Someone with equipment knowledge for the industry being examined is a practiced choice for the team, and people with production experience are specially valuable members, but squad members should possess a multifariousness of types and levels of experience. The RPA reports developed past different teams I've sent to bout the same plants take come in with remarkably consequent ratings and recommendations for comeback despite the fact that their members had widely varying levels of experience.

Depending on how experienced team members are with lean operations, it takes ane to iii days to railroad train a team to read a found using this tool. After initial training classes most leanness, tours of employees' own plants provide fantabulous opportunities to piece of work with the RPA process. What's more, reports completed during these and future tours tin get vital additions to the company's records, providing benchmarks for functioning improvement and example materials for time to come trainings.

Raters familiarize themselves with full general background about a plant by examining annual reports; analysts' reports; and the Web site of ten-Chiliad Wizard Engineering science, an industry association. Easily located benchmarking Web sites that focus on specific manufacturing and service industries also provide valuable initial insight.

Information technology's vital that the squad review industry-specific characteristics prior to conducting a bout since inventory or practices that would be considered excessive in some industries may be unavoidable in others. The pharmaceutical manufacture, for instance, must comply with purity requirements that call for a level of oversight that would be redundant in a commodity concern. And restaurants or electrical power plants volition typically accept either loftier inventory or under-utilized capacity, traits that wouldn't be desirable in other industries. By taking such practical and regulatory requirements into account, a squad volition ensure greater accuracy in its assessment.

Permit'south have a expect at each of the categories in plow.

Category 1

Customer Satisfaction

Workers in the all-time plants clearly know who their customers are—both internal and external—and make client satisfaction their main goal. What's more, they empathise that information technology's their chore to make tours infrequent experiences then visitors go out with resoundingly positive feelings about the facility. Such care for customers, or lack thereof, is readily credible in a brief plant tour. You should exist welcomed to the plant and given an overview of its layout, work-force, customers, and products. Quality and customer satisfaction ratings should exist prominently posted. And try asking an employee, "Where does your product go next?" If you hear, "Ford" or "John, over on line 6," you tin rate the establish higher on this measure than if you lot hear, "I put information technology in this bucket and I don't know what happens to it after that." (Questions 1, two, and xx on the RPA questionnaire relate to this measure.)

Category ii

Safety, Environment Cleanliness, and Order

In a clean and orderly plant, parts are easy to find, inventory is easy to count or guess, and products move safely and efficiently. The institute should be well lit, the air quality good, and noise levels depression. A visual labeling system should clearly marker inventory, tools, processes, and menstruation. A short constitute bout volition readily reveal how successfully the company attends to all these factors.

All component parts should be treated with equal intendance. Many companies become to great lengths to go along expensive parts in order while giving brusque shrift to low-cost ones like labels or fasteners. That habit tin can exist costly. Indeed, when we were making seats at Johnson Controls, we never lost a seat dorsum or cushion, only occasionally the bolts that joined the recliner mechanisms to the back and cushions were left off or not available. We couldn't ship a seat that was missing a bolt (or collect payment on information technology, naturally), so a single bolt was substantially equally valuable equally a larger, more than expensive role. (Questions iii–five and 20)

Category 3

Visual Management System

Tools that provide visual cues and directions are readily credible in well-performance plants. Such signage, clearly guiding employees to appropriate locations and tasks, tin can greatly raise productivity. Look for organizational tools such as kanban scheduling and color-coded production lines besides as patently posted piece of work instructions, quality and productivity charts, and maintenance records. Other indicators of skillful visual management include kiosks displaying data like the names of team members, productivity measures, and vacation schedules, as well equally a central location such as a control room or status board from which you can meet the current country of the overall operation. Chemical and other process-manufacture plants typically have potent visual management practices (every bit opposed to multiple, fragmented displays); even the largest plants tend to brandish product line flows, plant layouts, and other key information on a unmarried display. (Questions 2, 4, half-dozen–10, and 20)

The next three categories are intertwined. Rating a establish apace on these three is straightforward from obvious visual clues.

Category 4

Scheduling Organisation

The best plants rely on a single "pacing procedure" for each product line and its suppliers. This process, usually at the stop of the line, controls speed and production for all the upstream activities, much equally a footstep motorcar sets the speed at a racetrack. Demand for product at each work center is triggered by demand at the adjacent. This keeps inventory from building up, improves quality, and reduces reanimation because production lines aren't kept waiting for parts.

Plants that use a central scheduling system virtually always over- or under-produce some parts at some point in the process because instructions come up to each line from a central computer, not from the product line that really uses the part. I saw the downside of central scheduling in the extreme when I visited a tractor manufactory in the Soviet Union. The plant was diligently producing co-ordinate to its centralized schedule, merely the engine constitute wasn't aircraft plenty engines, and then each incomplete tractor was towed out to the g. I counted a full six months' supply of lifeless tractors, each waiting for its final, earth-shaking, office.

Rating Leanness RPA Rating SheetTeam members use the RPA rating canvass to assess a institute in 11 categories on a scale from "poor" (1) to "excellent" (9) to "best in class" (11). The total score for all categories will fall betwixt 11 (poor in all categories) and 121 (the best in the world in all categories), with an average score of 55. Factors to consider to rate a found in each category are described in this article; a more detailed list of evaluative factors appears on the Web at www.bus.umich.edu/rpa. The rating canvas likewise guides team members to questions in the RPA questionnaire (opposite) that relate specifically to each category. When plants are rated every twelvemonth, the ratings for well-nigh tend to amend. Ratings are usually shared with plants, and motivated managers commencement improve their plants in the categories that receive the lowest ratings.

Yous tin notice out how the plant schedules its lines by asking workers merely also past looking at inventory levels. If inventory piles up at one work heart, then the scheduling systems are probable independent of each other or the procedure is inadequately paced. Besides, if processes are scheduled from a key MRP system, you'll encounter computer screens or stacks of work orders at the line.

What's more, you tin can lookout for visual and verbal communication between procedure operators on the same line: People should be close enough to speak to one another and to see another's inventory. Operators tin and do override central scheduling and respond to visual cues; for case, they might slow production if they see inventory piling up down the line. (Questions 11 and 20)

Category five

Employ of Infinite, Move of Materials, and Product Line Catamenia

The all-time plants use space efficiently. Ideally, materials are moved just once, over as brusque a distance as possible, in efficient containers. Production materials should exist stored at line side, not in split up inventory storage areas. Tools and setup equipment should be kept near the machines. And the institute should be laid out in continuous product line flows rather than in "shops" dedicated to particular types of machines. On my first visit in the mid-1990s to Republic of austria-based Rosenbauer, i of Europe's largest burn down truck and equipment makers, I gave the plant excellent ratings except on its pump and truck assemblies, which were washed in traditional cells. Past the time I visited 2 years subsequently, the company had converted pump and truck assemblies to lean product lines, and managers reported that both productivity and quality had improved.

Counting forklifts is an easy style to get a sense of a found's employ of space. Forklifts require wide aisles, are expensive to operate, increase pollution, and encourage unnecessary movement of materials. In the best plants, if materials need to be moved a brusque distance, employees use paw-propelled gyre carts; if the materials are too heavy to motion by hand, garden tractors pull the carts in linked trains.

Space is a valuable commodity in any institute, and some companies make generating new space a productivity objective. Ane of our establish managers elevated this concept to an fine art form. He would regularly gratis up manufacturing space, polish the floor, cordon information technology off with stanchions, and and so claiming our sales group to observe new business to fill up the space. (Questions 7, 12, 13, and xx)

Category six

Levels of Inventory and Work in Process

Internal operations seldom require high inventories, then the observable number of any component role is a skilful measure of a constitute's leanness. You can get a quick read on inventory by watching a production line and counting the inventory at each work centre. For example, if ane widget comes off the line per minute, you lot know the line produces 60 per hour. If you lot count approximately 500 widgets by the piece of work center, and so yous know that over eight hours of output is just sitting there. In most cases, you want no more than a few minutes' worth of inventory by a work heart at one fourth dimension; each function should go directly to the side by side process to exist used fairly quickly. (Questions vii, xi, and twenty)

Category 7

Teamwork and Motivation

In the best plants, people consistently focus on the plant'south goals for productivity and quality, know their jobs well, and are eager to share their knowledge with customers and visitors. Motivated employees are hands discerned during a brief tour, equally are surly, unkempt, or indifferent ones; even a brusque talk with an operator tells you lot a lot.

Come across if there are clearly posted safety and environmental measures, pictures of the plant's softball team, posters boasting of quality and productivity improvements, charts showing contributions to charitable organizations. Yous might likewise look for posters or charts that describe problem-solving and employee empowerment procedures. These are visible indicators of teamwork—and if you can't spot such signs, chances are the plant hasn't truly embraced squad-piece of work. But yous tin supplement your observations with questions to the manager and plant staff about these activities as you bout. (Questions 6, 9, 14, 15, and 20)

Category 8

Condition and Maintenance of Equipment and Tools

In the all-time plants, equipment is clean and well maintained. The purchase dates and costs are stenciled prominently on the side of machinery, and maintenance records are posted. Such details ensure that workers know equally much equally possible about the machines and can plan for preventive maintenance. But perhaps more important, by posting price and maintenance records, the company signals to employees that management cares near the production, that they've invested in keeping the found running smoothly, and that they care about the work people exercise. Those are important factors in maintaining morale.

You lot tin can besides learn a great deal by request people on the factory flooring how things are working. When some of my students toured a new production line at an automotive supplier, one of them asked a worker how things were going. "Pretty well," he said but pointed to one critical sensor that wasn't performing consistently or accurately and then that manual, not automatic, inspection of a office was required. Thus, a significant investment in technology was undermined by an increase in person-hours to perform this task by paw when the sensor malfunctioned—a source of waste uncovered by a simple question.

Another question to ask employees is whether operators and product development personnel are involved in purchasing tools and equipment. People on the manufacturing plant flooring and others directly involved with the product are in the best positions to empathise the strengths and weaknesses of new equipment and the needs of the line.

Finally, look at the equipment yourself. Machines don't have to be new, but a recently purchased machine that's dirty and falling out of repair is a signal of poor preventive maintenance. Conversely, if a machine looks new simply was purchased long ago, you know the constitute's taking intendance of its investments. And many bug are easily visible to the naked eye—if you know to look. When I visited a petroleum refinery in Haifa, State of israel, in 1970, constitute managers told me about a trouble with two of four temperature charts, one at each cease of ii pipes carrying raw petroleum. The two "problem" charts displayed pregnant temperature variations while the other ii were flat, leading managers to assume the latter were under command. I asked to climb to the elevation of the furnace to accept a look and discovered that merely ii of the temperature sensor cables were connected—the two with significant variations. The other two charts were flat because the sensors that led to those charts had been cut. (Questions sixteen and 20)

Category 9

Management of Complexity and Variability

This category judges how the functioning manages, controls, and reduces the complication and variability it faces in its industry. It can be difficult during a tour to judge how a constitute performs in this category, just y'all tin can watch for certain indicators. For instance, many companies collect (so must procedure) much more than data virtually their operations than they demand; if y'all notice many people manually recording data and a large number of keyboards for data entry, the company may be doing a poor job of handling complication, especially if the information collection is done by paw.

In addition, since the product in lean plants flows through chop-chop and inventory is kept to a minimum, workers don't demand to proceed rails of a lot of parts. Furthermore, the all-time plants are able to use the same types of parts in the manufacture of dissimilar products. And finally, some companies—Toyota and Dell, to proper name two—build complexity handling into their production processes, designing systems that aid operators in picking the right parts out of a broad pick. If a worker reaches for the wrong valve, for instance, he or she might break an electronic beam, which would turn on cerise lights and possibly trigger an audible bespeak. Ask workers if such systems are in identify. (Questions 8, 17, and 20)

RPA Data for Plants with at Least 10 Bout Reports Each Categories 4, 5, and half-dozen (scheduling; space, materials movement, and production flow; and inventory) in the rating canvas consistently receive the lowest ratings. Few plants have a discernible strategy for how they movement materials, for example, and it's not unusual to run across forklifts trundling large containers with small parts—an inefficient use of space and equipment. Manufacturing plants besides make poor or infrequent apply of visual management systems; managers underestimate the power of such tools. The associated questions for categories 4 and v generally receive noes in average manufacturing plants.

Category 10

Supply Chain Integration

The best operations keep costs low and quality loftier by working closely with a relatively small number of defended and supportive suppliers. You can become a rough approximate of the number of suppliers by looking at container labels: Which supplier names announced on containers? Exercise the containers appear to exist designed and labeled specifically for customized parts shipped to this plant? If a company uses multiple suppliers for the same part or family of parts, it's unlikely that the suppliers were straight involved in the evolution procedure.

A best practise for plants is to pay suppliers based on completed, shippable product: Payment is made automatically when the product comes off the line. This cuts downwardly on paperwork and reduces the number of people involved in settling accounts. Ask plant personnel how suppliers are paid; information technology seems like an innocuous question, so people are often forthcoming with the answer. The presence of lots of paperwork on the receiving dock is some other indication of loftier costs in the supply chain. The best plants pull the materials from their suppliers every bit just another link in the pull organisation for each product line. (Questions 18 and 20)

Category eleven

Delivery to Quality

The best plants are always striving to meliorate quality and productivity, and it shows. Think that initial tour of the Hoover Universal institute by Japanese executives in the early 1980s? Subsequently we received the study from that bout, our managers came to conspicuously sympathise what Toyota expected from its suppliers and began to brand changes appropriately. In 1985, Hoover Universal won the contract to supply Camry seats for Toyota's new Georgetown institute, based not on the quality or productivity in Hoover's plants only on our highly visible commitment to continuous improvement.

Attending to quality is commonly easy to spot. If employees are proud of their quality programme, they usually give it a proper noun and post banners displaying the plant's vision or mission statement, business concern objectives, and metrics showing achievements to appointment. Both curt- and long-term goals for the plant and team—too every bit statements most internal and external customer requirements, production schedules, piece of work instructions, productivity levels, incoming and outgoing quality, scrap and rework levels, attendance, vacation schedules, safety, and levels of employee training—should be displayed at each piece of work center. (This overlaps with category 3.)

You should besides find out what the plant does with scrap. Amend plants call attention to scrap rather than hide it—by shining a calorie-free on it, for example, or mark it with scarlet record—because they want to know right away if fleck is building up or if a portion of the process is producing defective parts. One fashion to observe out: Ask people what they do when a faulty part comes off the line. Discarding it or discreetly putting it out of the way is a sign of inefficiency. Finally, ask about production development. Are cost and timing goals set during development? Are commencement-ups well managed and low in cost? (Questions fifteen, 17, 19, and 20)

Rating the Institute

Immediately following the tour, team members should meet to share their observations and impressions and to develop a report assessing the plant'southward leanness and estimating its price of sales. This meeting should happen right away, since visual data go out a vivid, but fleeting, impression.

The squad should utilize both the RPA rating sheet and the RPA questionnaire to charge per unit leanness. Rate each of the 11 categories on a scale from "poor" (ane) to "excellent" (9) to "best in class" (11). "Best in form" is meant literally: But one institute in each industry, worldwide, deserves this rating. Then total the ratings; the sum will exist between xi and 121, with an average plant scoring 55. The questionnaire is completed at the same time. (The sidebar "The RPA Process Database" explains the scores for some of the plants toured past my students.)

Since 1998 alone, my students and I take used the RPA process in more than than 400 tours of over 150 different manufacturing and service industries ranging from automotive suppliers, aerospace companies, and book manufacturers to cinema complexes, automobile dealers, and microbreweries. The table at the right shows the range of ratings obtained in the RPA rating sheet and questionnaire during some of those tours. These results are for manufacturing operations that were toured at least ten times each betwixt January 1998 and May 2001; the reports in this table correspond only a small fraction of the total reports in our database.

Different teams touring the same plants produce remarkably consequent results, simply we've constitute that the most consistent results are generated when visiting an exceptionally skilful or an exceptionally poor operation; the variance increases with an average plant, particularly when the team is inexperienced. The majority of squad members touring the plants rated here had little or no manufacturing feel.

We've found that typical scores for the sum of the ratings of the 11 categories range between 30 and ninety, with an average of 55. In this table, only four of the operations scored lower than 30 on whatsoever tour report, and just three operations accept e'er scored above xc. In improver, the average number of yeses in the questionnaire for all operations we've toured is seven, and the standard deviation is shut to two. In this table, we find that the mean number of yeses was often higher than the average number for all operations because these plants made improvements between tours. The standard deviation remained low, meaning that ratings from different teams remained quite consequent.

The plant'south total score on the rating sheet and the number of yeses on the questionnaire give you a fairly accurate assessment of a plant's efficiency: It's almost impossible to imitation a lean operation. Scores in 6 of the 11 categories and xvi of the twenty associated questions are based virtually solely on highly visible elements in a found's surroundings. The assessments on the rating sheet may be particularly useful considering the 11 categories highlight broad areas of strength and weakness. Categories with low ratings are instantly visible opportunities for improvement and should be the kickoff steps on a company'south journey to leanness.

By the mode, the RPA tool also includes a template for gauging a plant's price of sales (COS), besides as a set of typical operations measures ranging from yearly sales per employee to overhead to the number of hours needed to assemble a personal estimator. Nonetheless, measuring COS generally requires more experience than judging leanness, so I have omitted this ciphering in this article. For an overview of this aspect of the tool, please see the sidebar "Estimating Cost of Sales."

An estimate for cost of sales is a bonus from a quick institute tour. Simply the COS portion of the RPA tool is adequately complicated, so I'll outline merely the theory here. You'll discover work sheets with detailed instructions on the Spider web at www.bus.umich.edu/rpa.

I break downwardly COS into four wide areas—cloth; people; property, institute, and equipment (PP&East); and other—rather than the traditional accounting cuts of material, straight labor, and overhead. These groupings give more insight into comeback opportunities because of their relative independence. These iv areas are also useful when assessing acquisitions prior to detailed due diligence and in determining the competitive position of the performance.

Briefly, material costs include only the price of raw materials including freight in, with no other allocations such as purchasing price, scrap, and rework. People costs include all salaried and hourly wages and fringe benefits such every bit medical insurance, workers' compensation, and overtime. PP&East costs include costs for depreciation of all plants and equipment, taxes and insurance, scheduled maintenance, and utilities. "Other" costs cover the myriad expenses that commonly occur in whatsoever operation, from contracted services to copy paper. The costs in this category are depression (mostly 5% to 10% of COS) in a lean performance and loftier (often 15% to xx% of COS) in a traditional plant.

These four types of costs tin be estimated rather easily from the data you lot pick upwardly on a tour. Conversations with employees and the plant manager can yield a surprising amount of information—such as the number of employees, number of yearly production units, charge per unit of employee turnover, and complexity of the product mix. To approximate COS for these four categories, at a minimum, yous'll need to gather the following information: the total number of salaried and hourly people in the plant, the average hourly wage, and the facility'south square footage.

The full people cost can exist estimated from these data. Boilerplate salaries and fringes are usually well known, merely if you need to judge those numbers, our Web site includes a chart called Typical Operations Measures that contains data nigh many types of manufacturing costs. These measures show depression, average, and high measures for 45 unlike categories, such as sales per square foot, product hours per year, average salaries, and days outstanding on accounts receivable and accounts payable. Following your cess of a constitute's performance in other categories, yous'll be able to place it in the average, below boilerplate, or above boilerplate category.

Information regarding the material portion of COS is normally divulged during the tour. The PP&E can be estimated by the equipment practiced on the tour if he or she has some knowledge of local costs. The "other" cost is estimated from the plant ratings. All of your data tin exist recorded in a work canvass called RPA Tour Data, also bachelor on our Spider web site. In turn, the data from that table can be used to calculate a constitute's COS in a terminal work sheet.

Experienced tour teams can get a strikingly accurate picture of a plant's total COS, and the COS estimated by this technique will generally be within 5% to 10% of the bodily figure. While not a precise tool, this estimate of COS is however valuable in comparison similar plants—whether your own, your competitors', or your suppliers.

• • •

The fact is, if an functioning looks adept to the trained eye, it normally is. I've used this tool many times in conjunction with due diligence when evaluating an acquisition target. I've often fabricated substantial improvements to my own operations after brief tours of my competitors' plants. And I've taught the RPA procedure to several hundred students and managers, who have in plow looked anew at their own operations and those of their competitors. The tool is pretty uncomplicated—piece of cake to acquire, quick to put into practice—but information technology'southward proved very powerful in application.

A version of this commodity appeared in the May 2002 outcome of Harvard Business Review.