Paper Lamb productivity

SAVE OFFLINE

THE SOCIETY OF NAVAL ARCHITECTS AND MARINE ENGINEERS 601 Pavonia Avenue, Jersey City, NJ 07306 Tel. (201) 798-4800 Fax. (201) 798-4975 Paper presented at the 2001 Ship Production Symposium, June 13 - 15, 2001 Ypsilanti, Michigan

PAPER 7 A Shipbuilding Productivity Predictor Thomas Lamb1(F) and Aasmund Hellesoy1(St) ABSTRACT The first author has been working on the development of a Shipbuilding Productivity Predictor for many years. His first attempt was presented as a SNAME Section paper in 1998. It used only two parameters, namely size and Technology Level of the shipyard. However, it was clear that other parameters, such as throughput, vertical integration, range of ship types, and the ratio production workers versus non-production workers also affected productivity, and since that time he has been working to determine what parameters should be included and to develop an equation including all the significant parameters. This paper describes the research performed by the authors and the resulting Shipbuilding Productivity Predictor Equation. It is not the intent that the Productivity Predictor presented should be used by any shipyard to predict a "Target Productivity" based on its values of the parameters, but rather to show an approach how it could be done, and validate it. If the approach was supported by shipyards around the world it would become more acceptable. Shipyards could then benchmark their actual productivity against the target productivity. The research was supported through the NAVSEA Professor in Ship Production Science Cooperative Agreement. INTRODUCTION

NOMENCLATURE BP CAPP CC CGT DP GT K1 MH NSCC OECD PD PE PR ST STWT TE TP VI VLCC

1

Shipbuilders have used some form of productivity metric as a basis for estimating and planning over their long history. Such metrics have also been used to compare their performance with other shipbuilders since shipbuilding began. In some cases, the fact that they were price competitive was enough. On the other hand the shipyards that lost orders certainly performed analysis of the bids so that they could see where they had to improve to win future orders. Often used metrics were man-hours per ton of steel weight and outfit weight. This appeared satisfactory when there were only a few ship types and they were all about the same size. As the number of ship types developed, and ship sizes increased, these metrics have become less adequate measures of the relationships between productivity, the ships and the shipyards. Countries reported the total output of their shipbuilding industry in terms of Deadweight or Gross Tonnage. These were also found to be unsatisfactory metrics, as they could not show the effect of ship types and size.

Best Practice Rating Computer Aided Process Planning Compensation Coefficient Compensated Gross Tonnage Dual purpose Gross Tonnage (International) Gross Tonnage Coefficient Man Hours Naval Ship Compensation Coefficients Organization for Economic Cooperation and Development Productivity Production employees (Production /Non-production workers) Total Number of ship delivered/Number of different types (over 3 years) Steel Weight Total number of employees Total throughput Vertical integration Very Large Crude Carrier

University of Michigan 1

In the 1960’s, Britain developed the concept of the Compensated Gross Tonnage (CGT) as a means to take into account the differences in ship type and size. The main purpose was to try to show that they were still the leading output shipbuilding country based on work content. It was not until Sweden lost its position as shipbuilding output leader to Spain and then Spain to Japan, that other countries paid any attention to the concept. The OECD adopted this concept in the 1970’s as a more reliable aggregate metric for comparisons of shipbuilding productivity between countries. It is surprising, as the U.S. is a member of OECD, that the approach was never adopted by the U.S. shipbuilding industry. This may be because the approach has only been applied to commercial ships and, until recently, The U.S. have not been building commercial ships for a long time. The method’s acceptance is reflected by the fact that it was used by EEC for studies comparing their shipbuilding industry to other world leaders (Arthor Andersen 1993), and as the metric used to identify noncompetitive shipyards forpossible closure (KPMG 1992). A & P Appledore developed the concept further in the 1980’s and it became an essential Productivity tool in their shipbuilding benchmarking kit. Since then they, First Marine International, and others have used this approach for individual shipyard as well as national analyses (NSRP 2001). There are two inadequacies that limit the application of the CGT Productivity Metric in a truly global way. The first, as mentioned above, is that the approach is currently only applicable to commercial ships, as no internationally agreed Compensation Coefficients exist for naval ships. This prevents shipyards and countries that predominantly build naval ships from using the methods to compare their productivity with shipyards building commercial ships. The second inadequacy is that there is currently no way to tie the best practices used by a shipyard to the productivity. This paper describes an approach where an equation for shipbuilding productivity has been developed. It was believed that a shipbuilding productivity predictor could be developed based on readily available shipyard characteristics in the following form: PD = a x TEb x BPc x PRd x STe x VIf x DPg where: PD Productivity (MH/CGT) TE Total number of employees BP Best Practice Rating PR Total/Production Employees ST Number of ships delivered over 3 years divided by the number of ship types delivered over the same 3 year period. VI Vertical Integration DP Dual purpose (commercial versus naval)

and a is the coefficient and b, c, d, e, f, and g are all exponentials. It is not the intent that the Productivity Predictor presented should be used by any shipyard to predict a "Target Productivity" based on its values of the parameters, but rather to show an approach how it could be done, and validate it. If this approach was supported by shipyards around the world it would become more acceptable. Shipyards could then benchmark their actual productivity against the target productivity. A much larger database and corresponding analysis would be required before it could be used by any given shipyard as an accurate measure. Never-the-less the authors feel that it does correctly show the trends and can be used for preliminary analysis. It was also hoped that the study and the resulting productivity predictor would help identify why there are large differences in the shipbuilding productivity of different countries and even shipyards within the same country. This in turn may suggest ways for shipyards to plan how to improve their productivity. The research used statistical analysis tools to derive the equations, check the correlation and the standard deviation of the differences between actual and predicted productivity. PRODUCTIVITY Before describing the actual research and the results, it is necessary to define productivity. Productivity is output divided by input. Having made this simple definition it suddenly becomes very complex. How do you measure output and input? This depends on the users and their intentions. A metric that has been used by the global shipbuilding industry for a number of years is Man Hours/Compensated Gross Ton, (MH/CGT). This raises another problem. Based on the number of requests that the authors receive when they present the MH/CGT metric, it appears that many people do not know the difference between Deadweight Tonnage and Gross Tonnage and what the compensation Coefficient (CC)is compensating. Add to this that the concept of CGT is not readily in use in the US and it is therefore worthwhile to describe both before proceeding with the rest of the paper. GROSS TONNAGE The International Gross Tonnage is simply the molded volume, in cubic meters, of the enclosed hull and deckhouse of a ship multiplied by a coefficient; GT = K1 x V

performance is it's average labor hours for producing a CGT based on a period of 3 to 5 years. How effective is the CGT approach? If it was precise, for different ship types, sizes and complexity constructed in the same shipyard, the man-hours per CGT would be the same. Table I shows a comparison between man hours/tonne of steel and man hours/CGT. It can be seen that there is significant improvement in the CGT approach but it still is not precise.

The coefficient is to convert volume to admeasurement tons (0.35), and to keep the new Gross Tonnage as close to the average of the Gross Tonnage for existing ships as possible. The coefficient varies from 0.22 for very small ships to 0.32 for very large ships. Compensated Gross Tons is used in the equation to measure throughput (TP). COMPENSATED GROSS TONS

TABLE IV - COMPARISON OF PRODUCTIVITY MEASURES

The concept of Compensated Gross Tonnage was developed from the need to have a basic measurement that could take into account the differences in ship type, complexity in design and construction, and size. People had been working on this since the mid 1960s and it has developed into a world wide accepted approach. The Compensated Gross tonnage Coefficients (CGTC) have been developed internationally and accepted by the OECD. The coefficients are for commercial ships and are provided in a number of the references ( Bruce & Clarke 1992 and Lamb & Knowles 1999)

SHIP TYPE

VLCC 16 SuezMax Tanker 19 Product Tanker 27 Chemical Tanker 46 Bulk Carrier 19 Container ship 4400TFEU 19 Container ship 1800TFEU 28 Reefer 43 General Cargo 56 Ferry 51 Ocean Tug 105

CGT: A BASIS FOR PRODUCTIVITY METRIC To overcome this problem of a lack of an internationally accepted Productivity Metric, the CGT approach was further developed in 1967 by the Association of West European Shipbuilders and the Shipbuilders Association of Japan as an aggregate productivity metric by developing a metric based on man hours/CGT. Man-hours per CGT has been internationally accepted since then as a potential measure of productivity. A comparative aggregate productivity measure used for assessing an individual shipyard's 100

MH/ST. WT.

*100 to 150

80 Man Hours per CGT

70

U.S. *

60

*50 to 80

Germany

50 40

$1,000 / CGT

30

Denmark

Korea $500 / CGT

20

1992 Japan

10 0

0

10

20

32 22 20 36 20 22 22 34 29 39 31

From the CGT productivity metric, the concept of a competitive constant cost curve was developed. Such a curve is shown in Figure 1, which is reproduced from the author’s paper (Lamb et al, 1995). It is based on plotting the Productivity metric, MH/CGT, against fully burdened cost of labor. It is then possible to enter country averages or individual shipbuilders on the plot and compare them for the constant cost curve that passes through the lowest entry, which will be the price setter.

CHINA* CHINA** $2/Hr

90

MH/CGT

30

40

1995 Japan 50

Total Cost (Fully Burdened)per Man Hour (U.S. dollars)

Figure 1 - Iso-cost Curves

60

The cost in dollars to produce a Compensated Gross Ton (CGT) can be used to provide a common yardstick to reflect relative output of commercial shipbuilding activity in large aggregates such as countries, regions or shipyards. The Cost/CGT is not based on the total cost of building a ship, as material is not included. This is because labor cost is under the direct control of a shipyard whereas material is not, and when purchased on the global market, material should be about the same for similar ships. Material was thus considered to be relatively constant from one shipyard to another. CGT coefficients are not available for naval ships. In order to attempt to derive a rough order of magnitude productivity measure for U.S. shipbuilders in the recent Global Shipbuilding Competitiveness study (7), CGTCs were estimated for naval ships. These were then applied to data from a number of U.S. shipyards for naval ships, and the resulting productivity ranged from 120 MH/CGT for a Navy amphibious ship to 180 MH/CGT for a destroyer. The Global Shipbuilding Competitiveness study (Storch et al. 1995) also developed an overall measure of U.S. shipbuilding productivity by deriving the total output over the past five years of the shipyards visited, in terms of CGT and the man-hours expended. The average productivity was 185 MH/CGT. This probably presents a worse case than actually existed, as some of the shipyards had "planning yard" and other "white collar" Navy support activities that expend man hours without producing additional ship output. APPROACH Shipyard Questionnaire A questionnaire was distributed to 33 shipyards in Asia, Europe, and the US, attempting to gather the information needed to develop the productivity equation. Only 10 shipyards responded and only 1 from the U.S. The shipyards that responded provided information for a period of 4 years, most of them from 1996 to 1999. The distribution is shown in Table II along with the responses. TABLE II - QUESTIONNAIRE DISTRIBUTION Country U.S. Asia Europe TOTAL

Number Sent 6 13 14 33

Responses 1 7 2 10

As can be seen from the table the response from U.S. and Europe was disappointing. Additional data collected during recent shipyards visits were used to supplement the information provided by the

questionnaire recipients. This resulted in 27 data points. A database, as shown in Table III, was set up using the data received and in addition data was derived for a 3 year period from Lloyds records for 1997 through 1999. A number of the responding shipyards provided the throughput in CGT. However the Lloyds data provided throughput in GT and they had to be converted to CGT. The Lloyds based CGT was then compared to shipyard provided data as a check. Table IIIshows the type of information collected, but certain information is not shown in the table in order to protect the individual shipyards. Next the appropriate Compensation Factor was derived from the 1974 Table (Bruce 1992) and the CGT determined for each ship and totaled. The derived values were compared to the data received directly from the shipyards and any significant difference was reconciled before proceeding with the analysis. The total employment man hours for the same 3-year period was obtained either directly from the shipyard or by estimate. This gives the average productivity in Man Hours/CGT over the 3-year period. Next the Best Practices Rating was assigned to each shipyard based on the authors’ ongoing research. The Best Pratice rating (BP) is almost the same as the Technology Level, used in the early paper on this subject (Lamb 1998). Best Practices is a better term as it includes more than technology. Developing the Productivity for the shipyard database An average annual throughput (TP) in CGT was derived as well as the average annual total employees (TE) for each shipyard in the database over the 4 year period. The Productivity was calculated by determining the total manhours (MH) for the year and dividing that by the average annual throughput in (CGT). That is: PD = TE x Number of hours in year/TP MH/CGT Definition of Parameters An explanation and description of the different parameters of the productivity equation is provided below. Total Number of Employees (TE) The questionnaire recipients provided information about total number of employees (TE) as well as the number of production employees (PE). PR is the ratio of total number of workers divided by the number of production workers. Best Practices Rating (BP) The Best Practices Rating of the shipyards were developed based on the Technology Level approach used in the Global Shipbuilding Competitiveness studies, mentioned above (3). The approach assigns one of five technology levels to eight elements, namely; A. Steelwork Production B. Outfit Production C. Other Pre-erection

D. E. F. G.

Ship Construction and Outfit Installation Layout and Environment Amenities Design, Drafting, Production Engineering and Lofting H. Organization and Operating Systems

Vertical Integration (VI) Vertical Integration (VI) is the ratio of value added by the shipyard versus the total ship value and is defined by the percentage of labor cost to total cost. Dual purpose (DP) The DP=1 if a shipyard is building commercial and naval ships only, and the value is DP=2 for a yard producing commercial as well as naval ships. Ships Delivered/Ship Types (ST) ST is a parameter that takes into account the number of total ships built compared to number of “series” ships built over a given time, such as three years. Developing the Productivity Predictor Equation The statistical tool SYSTAT was used to analyze the data and eventually obtain the coefficient and exponentials (a, b, c, d, e, f, g and h) in the productivity predictor. To validate the derived values for both Best Practice rating and Productivity the values for the different shipyards were superimposed on the Curves developed by First Marine International as part of the Benchmarking Study that they have just completed for

the NSRP ASE (3) and the result is shown in Figure 2. It is clear that the actual productivity of some of the shipyards in the research database is not as good as the European Large Shipyard curve with most of the data points lying well above the Curve.. To compare the data, equations were derived for the two curves shown in Figure 2. They are: PD = 480 BP-2.3 for Large Shipyards, and PD = 252 BP-2.3 for Small Shipyards. The fit for the large shipyard equation with the shipyards in the database is shown in Table IV. It was, understandably not very good. Developing a similar equation from the study database results in the following equations: PD = 4058 BP-3.5 including US Shipyard data PD = 18982 BP-4.64 excluding US shipyard data At first an attempt was made to derive the productivity for U.S. shipyards using the Naval Ship Compensation Factors (NSCF) derive by one of the author's (8). However, during the analysis the prediction equation was not satisfactory and it was found that the U.S. data was on a different level. So the analysis was performed for all the shipyards excluding the U.S. shipyards and the result was an improvement, as can be seen from Table IV. The fit for these equations is also shown in Table III and as it is based on the collected data the fit is better but still not good enough.

PRODUCTIVITY (MH/CGT)

100 US

80 60

Europe

40 Large Shipyards

Korea

Small Shipyards

20

Japan

Ongoing Research Database Trendline

0 1

2

3

4

BEST PRACTICE RATING (BP) Figure 2 - Plot of Productivity and Best Practice Rating

5

TABLE III - STUDY DATABASE COUNTRY

NO TE

Asia Asia Asia Asia Asia Asia Asia Asia Asia Asia Europe Europe Europe Europe Europe Europe Europe Europe Europe Europe US US US US US US

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

PE

SHIPYARD PARAMETERS BP TP VI DP 5.0 0.5 1 4.1 0.6 2 5.0 0.5 1 5.0 0.4 1 4.7 0.4 1 4.5 0.4 1 4.5 0.4 1 4.6 0.4 1 4.8 0.4 1 4.7 0.4 1 4.0 0.3 1 3.9 0.3 1 4.0 0.3 1 3.7 0.4 2 4.0 0.3 1 4.0 0.3 1 4.0 0.4 1 4.3 0.4 1 4.2 0.4 1 4.0 0.3 1 3.3 0.5 2 3.6 0.5 2 3.7 0.4 2 3.8 0.4 2 3.0 0.4 2 3.3 0.5 2

ST 12.8 8.0 3.7 4.2 9.3 12.3 13.3 4.4 12.0 14.5 5.5 3.0 1.0 5.0 2.0 3.0 3.3 1.0 4.7 2.5 3.0 1.7 1.5 3.0 8.7 2.0

PR 1.20 1.65 1.16 1.20 1.20 1.31 1.20 1.30 1.20 1.18 1.42 1.25 1.19 1.27 1.40 1.40 1.33 1.14 1.14 1.39 1.51 1.38 1.35 1.30 1.20 1.40

PD=MH/CGT Actual 6.3 20.0 10.1 10.4 13.9 26.0 31.9 19.0 9.8 17.7 22.8 34.2 33.9 48.6 43.3 30.3 21.1 16.1 18.3 29.6 77.9 67.7 50.9 42.5 73.6 102.3

TABLE IV - COMAPRISON OF BEST PRACTICE RATING EQUATION Actual MH/CGT 6.29 20.00 10.11 10.40 13.88 26.04 31.92 19.03 9.83 17.67 22.75 34.17 33.90 48.61 43.34 30.33 21.06 16.07 18.30 29.60 77.95 67.68 50.93 42.47 73.55 102.27

FMI Large Shipyard MH/CGT 11.85 18.70 11.85 11.85 13.66 15.10 15.10 14.35 13.01 13.66 19.79 20.98 19.79 23.68 19.79 19.79 19.79 16.76 17.69 19.79 30.81 25.22 23.68 22.27 38.36 31.91 Standard Deviation

Difference 5.55 -1.30 1.74 1.45 -0.23 -10.94 -16.82 -4.68 3.18 -4.01 -2.96 -13.19 -14.11 -24.93 -23.54 -10.53 -1.26 0.69 -0.61 -9.81 -47.14 -42.46 -27.25 -20.20 -35.19 -70.36 18.32

New incl. US MH/CGT 14.66 29.33 14.66 14.66 18.20 21.18 21.18 19.62 16.91 18.20 31.97 34.92 31.97 41.98 31.97 31.97 31.97 24.83 26.96 31.97 62.61 46.19 41.98 38.24 87.35 66.04

Difference 8.37 9.33 4.55 4.26 4.31 -4.85 -10.73 0.58 7.08 0.53 9.21 0.75 -1.94 -6.64 -11.37 1.64 10.91 8.76 8.66 2.36 -15.34 -21.48 -8.95 -4.23 13.79 -36.23 11.34

New excl. US MH/CGT 10.90 27.33 10.90 10.90 14.51 17.76 17.76 16.04 13.16 14.51 30.65 34.47 30.65 43.99 30.65 30.65 30.65 21.92 24.45 30.65 74.76 49.95 43.99 38.87 116.28 80.24

Difference 4.60 7.34 0.78 0.50 0.63 -8.28 -14.16 -3.00 3.34 -3.16 7.90 0.29 -3.26 -4.62 -12.69 0.32 9.59 5.85 6.15 1.05 -3.19 -17.73 -6.94 -3.60 42.73 -22.03 11.81

The next step was to determine the best equation/combination of the 6 shipyard parameters. This was done by deriving many equations with different combinations of the 6 shipyard parameters. Table V gives a number of equations that present the development to the final selected equation. It also shows the correlation (r2) for each equation.

Each equation was used to calculate the productivity for the study database shipyards and the differences between the actual and predicted productivity determined as well as the standard deviation of the differences. The results are shown in Table VI. It can be seen that the final two equations are a substantial improvement over the others. The predicted shipyard productivity for U.S. shipyards in the table, were then used to reverse analyze the U.S. shipyard naval ship compensation coefficients (NSCC) to make the U. S. shipyards fit the equation. The new derived NSCC are compared to those developed earlier (Lamb 1998) in Table VII. It can be seen that there is significant agreement and that the earlier developed NSCCs are thus validated.

TABLE V - DEVELOPMENT OF EQUATION

1 2 3 4 5

r2

Equation

No PD = 3302 BP

-4.73

0.24

TE

.981

PD = 2669 BP-4.68 TE0.25 PR0.30

.982

PD = 3806 BP-4.98 TE0.26 PR0.36 DP-0.13

.983

PD = 203 BP-3.39 TE0.24 PR0.41 DP0.35 VI-0.74

.986

PD = 111 BP-3.00 TE0.27 PR0.60 DP0.41 VI-0.66 ST-0.08

.987

TABLE VI - PRODUCTIVITY PREDICTION COMPARISON Actual MH/CGT 6.29 20.00 10.11 10.40 13.88 26.04 31.92 19.03 9.83 17.67 22.75 34.17 33.90 48.61 43.34 30.33 21.06 16.07 18.30 29.60 77.95 67.68 50.93 42.47 73.55 102.27

1 Diff. MH/CGT 9.05 2.75 21.23 1.23 9.84 -0.27 9.15 -1.25 12.39 -1.49 24.67 -1.37 31.10 -0.82 17.36 -1.68 9.59 -0.24 19.82 2.15 22.18 -0.57 34.31 0.14 32.28 -1.62 48.60 -0.02 28.87 -14.47 31.84 1.52 25.51 4.45 16.48 0.41 18.40 0.10 28.69 -0.91 98.39 20.44 71.46 3.79 51.93 1.01 42.04 -0.43 103.73 30.17 122.37 20.10 Standard Deviation 8.61

2 MH/CGT 8.87 22.67 9.55 8.97 12.15 24.77 30.49 17.39 9.40 19.33 22.75 34.00 31.55 48.34 29.50 32.54 25.73 15.94 17.80 29.28 102.55 72.67 52.55 42.08 101.68 125.04

Diff. 2.57 2.67 -0.56 -1.43 -1.74 -1.27 -1.43 -1.65 -0.43 1.66 -0.01 -0.17 -2.35 -0.28 -13.84 2.21 4.67 -0.12 -0.50 -0.33 24.60 4.99 1.62 -0.40 28.13 22.77 8.97

7

3 MH/CGT 8.61 22.97 9.27 8.70 11.79 24.04 29.59 16.87 9.12 18.76 22.07 33.00 30.62 48.97 28.63 31.58 24.97 15.47 17.28 28.41 103.89 73.62 53.24 42.63 103.01 126.68

Diff. 2.31 2.97 -0.84 -1.69 -2.10 -2.00 -2.33 -2.16 -0.71 1.09 -0.68 -1.17 -3.28 0.36 -14.71 1.25 3.91 -0.60 -1.02 -1.19 25.94 5.94 2.31 0.16 29.46 24.40 9.59

4 MH/CGT 8.76 20.09 9.39 10.45 13.05 25.41 30.86 18.35 10.39 20.69 25.09 35.63 33.92 48.72 32.45 35.78 22.73 15.16 16.42 32.18 77.23 60.52 53.41 44.05 77.48 91.35

Diff. 2.46 0.09 -0.72 0.05 -0.83 -0.63 -1.05 -0.69 0.56 3.02 2.34 1.46 0.02 0.10 -10.89 5.45 1.68 -0.90 -1.88 2.58 -0.72 -7.15 2.48 1.58 3.93 -10.93 3.90

5 MH/CGT 8.56 21.07 10.11 10.94 12.57 25.18 30.62 19.31 9.73 20.14 23.43 34.83 36.30 48.61 33.57 36.24 22.65 16.38 15.60 32.68 82.12 68.95 59.59 46.30 66.09 99.74

Diff. 2.26 1.07 0.00 0.54 -1.31 -0.85 -1.30 0.28 -0.10 2.47 0.68 0.65 2.40 0.00 -9.77 5.92 1.60 0.31 -2.70 3.07 4.17 1.27 8.66 3.83 -7.46 -2.53 3.67

CONCLUSIONS

TABLE VII - REVERSE DERIVED NSCC Ship Type Aircraft Carrier Destroyer LHD LSD AOE

Old 4.5 12.4 5.1 4.6 2

New 5.8 11.3 4.8 4 2

1.

2.

SENSITIVITY ANALYSIS To determine the influence of the various parameters on a shipyard's productivity a series of sensitivity calculations were performed. This was accomplished by observing the change in the productivity metric for a range of step changes in each parameter, one at a time. The results are shown in Table VIII. It can be seen that the Ratio of Number of Ships Delivered to the Range of Ship Types delivered, has the minimum significance on productivity and probably can be dropped from the equation, even though it has a small improvement in prediction accuracy. The Best Practice Rating has the greatest significance. The actual ranking from greatest to least is shown in Table IX.

An approach developing a Productivity Prediction equation has been shown and the results based on the study database, as well as other recently available data, validate the approach. In addition it was possible to reverse derive Gross Ton Compensation Coefficients for the naval ships involved in the U.S. shipyard data. This showed a remarkable validation of the previously derived NSCCs. TABLE IX - RELATIVE IMPACT OF PARAMETERS Parameter

Best Practice Rating Dual Purpose Total /Production Employees Vertical Integration Total Employees No. of Ships Delivered/No.of Ship types

Impact Ratio 29 21 8 8 4 1

TABLE VIII - PARAMETER SENSITIVITY ANALYSIS TE BP VI DP ST PR Productivity Ratio to % Chg Relative MH/CGT Base Weight 4000 4.50 0.50 1 4 1.4 19.18 1.000 Total 3600 4.50 0.50 1 4 1.4 18.65 1.028 2.8 3.5 Employees 3200 4.50 0.50 1 4 1.4 3.59 18.08 1.061 6.1 2800 4.50 0.50 1 4 1.4 17.45 1.100 10 3.70 Best 4000 4.05 0.50 1 4 1.4 26.32 0.729 27.1 33.88 Practice 4000 3.60 0.50 1 4 1.4 37.47 0.512 48.8 28.71 4000 3.15 0.50 1 4 1.4 55.93 0.343 65.7 24.33 Vertical 4000 4.50 0.45 1 4 1.4 20.57 0.933 6.7 8.38 Integration 4000 4.50 0.40 1 4 1.4 8.06 22.23 0.863 13.7 4000 4.50 0.35 1 4 1.4 24.28 0.790 21 7.78 Dual 4000 4.50 0.5 2 4 1.4 25.53 0.752 24.8 31.00 Purpose 4000 4.50 0.5 3 4 1.4 30.17 0.636 36.4 21.41 4000 4.50 0.5 4 4 1.4 33.96 0.565 43.5 16.11 No Ships/ 4000 4.50 0.5 1 3.6 1.4 19.34 0.992 0.8 1.00 Types Ships 4000 4.50 0.5 1 3.2 1.4 1.00 19.52 0.983 1.7 BASE 4000 4.50 0.5 1 2.8 1.4 19.72 0.973 2.7 1.00 Total/ 4000 4.50 0.5 1 4 1.26 18.02 1.065 6.5 8.13 Production 4000 4.50 0.5 1 4 1.12 8.35 16.80 1.142 14.2 Employees 4000 4.50 0.5 1 4 0.98 15.52 1.236 23.6 8.74 Change Parameter

8

Impact 41% 30% 11% 11% 6% 1%

3.

4.

5.

6.

7.

8.

9.

up to 8 major categories and it is probable that a better Productivity Prediction equation would result from a study in which the major categories were treated as individual parameters.

Regarding the impact of the different parameters on a shipyard's productivity the biggest hitter, by a significant amount, is the Best Practices Rating. This simply confirms what is shown in Figure 2. The other parameters were moderators that provided better prediction and thus explained why there can be significant differences in productivity in different shipyards. The next significant parameter is Dual Purpose and it clearly shows the adverse impact of trying to perform both government and commercial shipbuilding in the same facility. Next is the Total to Production Employee ratio and this shows the obvious adverse impact when this increases. The next one, Vertical Integration is very interesting in that it goes against current organization recommendations to concentrate on core competences and outsource everything else. The study showed that shipyards with greater vertical integration were more productive. The impact of total size, based on Total Employees also had a relatively small impact on productivity. Thus it appears that there is only a small disadvantage to having a very large shipyard and large work force. It should be noted that this conclusion is based on looking at shipyards that were all above the large shipyard line in Figure 2. There were no shipyards in the study database that fit the small shipyard curve in the figure. It is essential that this aspect be further investigated as the reasons for the significant difference shown in Figure 2 should be understood and utilized to the benefit of larger shipyards. Finally it does not appear that the number of ships built relative to the number of ships of the same design has any impact on the productivity. It is suggested that this is because world-class shipyards do not focus on ships but rather the interim products that make up the ships, thus being able to benefit from similarity in interim products rather than ships. This should be of interest to all those who promote the learning or experience curve approach. It perhaps also explains why the largest series ship builders in the world, that is the U.S. shipyards currently building the DDG51 (up to 26 ships each) are not the most productive. The Best Practices rating is a composite of

RECOMMENDATIONS 1.

2.

Continue to work on the development of the equation but only if a larger number of shipyards (200), of all sizes, from around the world are interested in participating. Perform a similar analysis using the components of the Best Practices Rating as individual parameters in order to determine the influence of each one on the productivity of a shipyard. This would then provide a tool for a shipyard to examine a number of alternative approaches to improving productivity.

ACKNOWLEDGEMENTS The authors would like to acknowledge the help from many shipyards throughout the world, in the form of answers to the questionnaire providing the information necessary to populate the database from which the equation coefficients could be derived and also for the help and openness of the shipyards visited over the past 6 years. Without such willingness to contribute their effort and data the outcome could not have been achieved. However, the authors fully accept full responsibility for the work described and the results presented in the paper. REFERENCES Arthur Anderson 1993, EEC Shipbuilding Industry Study on Costs and Prices, November KPMG Peat Marwick 1992,Report of a Study into the Competitiveness of European Community Shipyards, , October NSRP ASE 2001, Benchmarking of U.S. Shipyards: Industry Report, , January Bruce, G and Clark, J. 1992, Productivity Measures as a Tool for Performance Improvement, Transactions RINA Lamb, T. and Knowles, R. A. 1999, Productivity Metric for Naval Ships, 1999 Ship production Symposium, Washington, D.C. 9

Lamb, T. 1998, A Productivity and Technology Metric for Shipbuilding, SNAME Great Lakes & Great Rivers Section Meeting, Cleveland, Ohio, January

Lamb, T., et al. 1995, A Review of Technology Development, Implementation, and Strategies for Further Improvements in U.S. Shipbuilding, SNAME Annual Meeting Storch, R., Clark, J and Lamb, T. 1995, Requirements and Assessments for Global Shipbuilding Competitiveness, NSRP

10