Professional Documents
Culture Documents
Shipbuilding
Shipbuilding
Support RAND
Purchase this document
Browse Books & Publications
Make a charitable contribution
Monitoring the progress of shipbuilding programmes : how can the Defence Procurement
Agency more accurately monitor progress? / Mark V. Arena ... [et al.].
p. cm.
MG-235.
Includes bibliographical references and index.
ISBN 0-8330-3660-2 (pbk.)
1. Great Britain. Royal NavyProcurement. 2. Defence Procurement Agency (Great
Britain)Management. 3. Shipbuilding industryGreat BritainManagement. I.
Arena, Mark V.
VC265.G7M66 2004
359.8'3'0681dc22
2004018825
iii
iv Monitoring the Progress of Shipbuilding Programs
identify why ships are delivered late and understand why com-
mercial shipbuilders have much better schedule performance.
Preface...............................................................................................................................................iii
Figures ..............................................................................................................................................ix
Tables................................................................................................................................................xi
Summary ...................................................................................................................................... xiii
Acknowledgements ..................................................................................................................xxi
Abbreviations ...........................................................................................................................xxiii
CHAPTER ONE
Introduction...................................................................................................................................1
Major UK Defence Acquisitions Are Typically Behind Schedule.......................2
DPA Often Does Not Realise Projects Are Behind Schedule Until Late in
the Production Cycle .....................................................................................................2
Commercial Ships Are Typically Produced On Time ..............................................4
Areas of Inquiry ............................................................................................................................5
Methodology ..................................................................................................................................6
Organisation of This Report ..................................................................................................8
CHAPTER TWO
How Do Shipbuilders Monitor Progress During Design and
Production? ...........................................................................................................................9
Earned Value Metrics Dominate Primary Progress Measures ............................ 11
The US Department of Defense Has Embraced Earned Value
Management.................................................................................................................... 13
How Do Shipbuilders Use Information That Metrics Provide? ....................... 13
Shipbuilders Employ No Consistent Forecasting Methodology....................... 14
vii
viii Monitoring the Progress of Shipbuilding Programs
CHAPTER THREE
What Progress Information Should the DPA Require of
Shipbuilders? ..................................................................................................................... 17
Implementing Earned Value Management.................................................................. 18
Cost Performance Index ........................................................................................................ 19
Other Metrics ............................................................................................................................. 20
CHAPTER FOUR
What Causes Ships to Be Delivered Late, and Why Do Commercial
Shipbuilders Have Good Schedule Adherence?.............................................. 23
Change Orders and Late Product Definition: The Leading Contributors
to Schedule Slippage.................................................................................................... 23
Commercial Customers Resolve Change Issues Faster and Have Fewer
Changes.............................................................................................................................. 24
Commercial Versus Military Payment Incentives .................................................... 26
Setting Realistic Expectations ............................................................................................. 28
CHAPTER FIVE
Conclusions and Recommendations ............................................................................. 31
APPENDIX
A. An Overview of Earned Value Management.................................................... 33
B. Earned Value Management Calculations........................................................... 41
C. Data Collection Form................................................................................................... 45
References ..................................................................................................................................... 57
Figures
ix
Tables
xi
Summary
Introduction
The Defence Procurement Agency (DPA), part of the UK Ministry
of Defence (MOD), provides services and equipment for the national
security of the United Kingdom. This task requires efficient capital
management. One of the keys to efficient capital expenditure is good
programme management.
Cost and schedule control and estimating are central competen-
cies of programme management. Controlling cost and schedule is the
principal focus of this report. Continually updated knowledge of
project status is important for both operational planning (determin-
ing when the customer will have use of the asset) and financial man-
agement (determining cash flow needed to support the programme).
A good control system can also aid programme improvement by iden-
tifying problem areas before they greatly affect production. Accurate
estimating of changing programme needs allows an organisation to
make best use of limited funding. The importance of estimating and
control has been recognised by the DPA in that two of the five key
targets monitored are related to cost and schedule performance.
The DPA measures its annual performance against five key tar-
gets. Key Target 2 relates to programme slippage, i.e., the delay
between the promised in-service date and the actual or projected in-
service date. The MOD indicates average programme slippage results
xiii
xiv Monitoring the Progress of Shipbuilding Programs
in product delivery approximately one year later than the date origi-
nally anticipated at Main Gate (MOD, 2002a). Moreover, for mili-
tary shipbuilding, slippage is often recognised very late in the pro-
gramme, making it more difficult to overcome. Although measures of
slippage cover MOD programmes broadly, recent shipbuilding pro-
grammes such as the Landing Platform Dock, Astute, and Auxiliary
Oiler have been documented as suffering slippage as well (Scott,
2004).
These issues led the DPA to ask the RAND Corporation to
Methodology
Figure S.1
Shipbuilder Use of Metrics at Various Production Phases
Earned Earned
Milestones Milestones Milestones Milestones value value
related related
2/3 to
1/3 Task Task Task Task Milestones Milestones
Figure S.2
Causes of Schedule Slips Reported by Shipbuilders (percentage)
Other
Productivity 6%
8%
Material
availability
8%
Weather, unexpected
equipment failures, etc.
9%
Change orders/late
Lack of technical product definition
information 46%
23%
RAND MG235-S.2
Summary xix
tives. Still, the practicality for such an approach for a warship is un-
clear.
This report would not have been possible without the contributions
of several firms and individuals. First, we would like to thank Andy
McClelland of the DPA for guiding this research and providing con-
tacts within the MOD. We also would like to thank Muir Mac-
Donald (Astute program) and Keith Prentice (Type 45 program), the
IPT team leaders who sponsored this research. We would like to
thank Philip Koenig (Office of Naval Research) and Harry Thie
(RAND) for their review of the report and the many improvements
and suggestions they made. Professor Thomas Lamb (University of
Michigan) participated in the studys data collection and made several
helpful suggestions for the analysiswe thank him for his time and
help. Finally, we are deeply indebted to the shipbuilders that partici-
pated in the study survey and interviews. Without their assistance,
this research would not have been possible.
xxi
Abbreviations
xxiii
CHAPTER ONE
Introduction
1
2 Monitoring the Progress of Shipbuilding Programs
Some slippage shows up early, but the bulk of it is often not recog-
nised until the latter stages of the procurement cycle (NAO, 2002).
Overall, schedule (or time) variance occurs very early, levels off mid-
way through the procurement cycle, and increases again late in the
cycle (see Figure 1.1). Cost variance, conversely, does not occur early,
increases at the greatest rate midway through the procurement cycle,
and levels off late in the cycle.
Introduction 3
Figure 1.1
Procurement Life Cycle Cost Variation Versus Time Variation
Cost variation
____________
1 One of the reviewers offered an additional explanation:
With poor accuracy control . . . various stages of production can be churning out on time
(but dimensionally varying) intermediate products. Every process is indeed meeting its sche-
dule, and everyone is happy. Then when it comes to final assembly (late in the programme), all
of a sudden things just dont fit together and all sorts of time-consuming, handcrafted rework
must be done. There goes the schedule.
4 Monitoring the Progress of Shipbuilding Programs
Figure 1.2
Representative Commercial Shipbuilders Delivery Performance,
19852001
4
1343
AHEAD OF SCHEDULE
Delivery variation
1322
1278
1279
(in months)
2
1302
1331, 1337
1327, 8, 9
1323
1336
1309
1316
1297
1344
1290
1289
1291
1286
1292
1340
1341
1339
1295
1296
1294
1293
1345
1346
1282
1283
1284
1287
1288
1303
1310
1304
1315
1312
1317
1318
1320
1319
1325
1324
1334
1335
1321
1285
1319
1330
1332
1298
1301
1333
DELAYED
2
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Year
Areas of Inquiry
Methodology
We used a four-part methodology to address our areas of inquiry.
First, we surveyed major shipbuilders in the United Kingdom,
United States, and European Union 3 (the survey form is reproduced
in Appendix C). The survey included questions about such pro-
gramme areas as schedule tracking methods, change control, causes of
slippage, and forecasting.
Second, after receiving responses to these surveys, we had
follow-on conversations with the shipbuilders to better understand
the survey information, ensure that the forms had been completed in
a consistent manner, and allow the shipbuilders to discuss relevant
issues not captured in the survey. Most shipbuilders used this oppor-
tunity to show us a sample progress report. From those surveys and
discussions, we determined metrics commonly used by shipbuilders.
For comparison purposes, we also interviewed representatives from
the project management department of a major oil firm in regards to
their practices for cost and schedule control.
Third, we reviewed relevant literature to assess the advantages
and disadvantages of the metrics we had identified. For example,
____________
2 Overoptimism is not just limited to the DPA. The recent National Audit Office report
(Ministry of Defence: Major Projects Report 2002) indicating significant cost overruns in sev-
eral preSmart Acquisition programs has raised questions about the soundness of contractors
bids on defence procurements. A Times Online article (PA News, 2004) suggests that con-
tractors quote unrealistically low prices to win contracts.
3 For simplicity, throughout this report, the authors use the term European Union, or
EU, to refer to those non-UK European shipbuilders surveyed (even though the United
Kingdom is an EU member). Specifically, EU countries that participated consist of Finland,
France, Italy, the Netherlands, and Spain (see Table 1.1 for the full list of shipbuilders).
Introduction 7
Table 1.1
UK, US, and EU Shipbuilders Surveyed
____________
4 For example, see references and links provided on the Office of the Under Secretary of
Defense (Acquisition, Technology & Logistics) Acquisition Resources & Analysis/Acquisi-
tion Managements Web site on EVM at www.acq.osd.mil/pm/.
8 Monitoring the Progress of Shipbuilding Programs
9
10 Monitoring the Progress of Shipbuilding Programs
Table 2.1
Schedule Metrics Categories
Figure 2.1
Shipbuilder Use of Metrics at Various Production Phases
Earned Earned
Milestones Milestones Milestones Milestones value value
related related
2/3 to
1/3 Task Task Task Task Milestones Milestones
Figure 2.2
Percentage of UK, US, and EU Shipbuilders Using Earned Value Metrics
Design
Module block
Assembly
Outfitting
US
UK
Testing trials European
Commissioning
0 20 40 60 80 100
Percentage
a
That is, the percentage for firms reporting the use of a metric in a particular phase.
RAND MG235-2.2
How Do Shipbuilders Monitor Progress During Design and Production? 13
The DPA should begin requesting from shipbuilders the basic infor-
mation needed for earned value management, including core meas-
ures on
17
18 Monitoring the Progress of Shipbuilding Programs
Other Metrics
EVM has some significant limitations for schedule control that must
be considered, particularly when using an SPI to evaluate progress
(Book, 2003). One of these limitations is that EVM metrics are based
directly on cost data (in pounds or worker hours) and have only an
indirect link to programme time or duration (although cost and time
are generally correlated). EVM assumes a linear correlation between
cost and duration for an activity, with tasks or activities weighted by
cost. Therefore, an activity that is more costly contributes more to the
overall programme SPI than does one that is less costly. Thus, a case
of schedule slippage for a low-cost task on the critical path would be
masked in the overall SPI for the programme. One advantage of re-
questing EVM data at the task or activity level is to better understand
whether activities on the critical path are slipping.
Because of the limitations of EVM data for schedule manage-
ment, the DPA should consider some additional progress metrics on
shipbuilding programmes. Some shipyard representatives we inter-
viewed said that some installation quantities, such as piping and cable
installed, were good monitors of progress, although others noted
these metrics were not useful in monitoring progress and were diffi-
cult to quantify in practice. During the testing phase, most shipyards
use the number of test procedures/plans complete as a metric to assess
progress. For commissioning and handover phase of production,
many shipyard representatives we interviewed said the number of
compartments completed or accepted was a useful metric for measur-
ing progress.
Some other information related to schedule should be tracked.
First, the DPA should ask the shipbuilder to provide an updated,
forecasted completion date for each progress report. This information
should include a revised critical path analysis (based on a high-level
schedule described earlier). Such additional information will offset
the weaknesses of EVM in weighting by value and ignoring critical
path issues. Lastly, the DPA should monitor the value of unresolved
(unadjudicated) changes. An estimated value for these changes not in
What Progress Information Should the DPA Require of Shipbuilders? 21
the baseline would help to serve as a check of the EAC and whether
the amount of potential new work would make the schedule slip.
Our discussions with shipbuilders and review of relevant litera-
ture lead us to recommend that the DPA request EVM data from the
shipbuilders to monitor progress. The agency should also track the
critical path and completion milestones. Being able to track progress,
however, is but one part of the problem the DPA faces in improving
schedule adherence. Perhaps a more important issue is to understand
the reasons that ships are delivered late and why commercial pro-
grammes do much better than military ones. We explore that issue in
the next chapter.
CHAPTER FOUR
23
24 Monitoring the Progress of Shipbuilding Programs
Figure 4.1
Causes of Schedule Slips Reported by Shipbuilders (percentage)
Other
Productivity 6%
8%
Material
availability
8%
Weather, unexpected
equipment failures, etc.
9%
Change orders/late
Lack of technical product definition
information 46%
23%
RAND MG235-4.1
____________
1 Simpler commercial projects tend to require even fewer changes. For example, a $60 mil-
lion commercial vessel needed about $300,000 in changes. See Buetzow and Koenig (2003).
2 Stakeholders that have some input include the project team, the fleet, the classification
societies, safety groups, inspectors, and so forth.
26 Monitoring the Progress of Shipbuilding Programs
Figure 4.2
Percentage of Total Number of Changes Occurring at Various
Production Phases
60
Commercial
Percentage of total number of changes
Military/government
50
40
30
20
10
0
Design Module block Assembly Outfitting Testing/trials
RAND MG235-4.2
The consequences of late changes are twofold. First, they are more
expensive to implement because already-completed work needs to be
redone. In the design phase, only design time is lost through changes.
Second, changes lead to additional work that is much more difficult
to accomplish in the later stages of construction (e.g., poor access
must be overcome).
costs or ill will for the customer. (Military ships, of course, do not
face such problems.) Commercial shipbuilders often spend their own
money, sometimes through outsourcing some tasks, to avoid liqui-
dated damage penalties and to keep construction on time. Recent
DPA shipbuilding programmes, namely Astute and Type 45, do have
liquidated damage clauses. These are to cover potential additional
costs to the MOD should these programmes slip but are not targeted
as schedule incentives. The possibility that such penalties could bank-
rupt one of the few remaining UK shipbuilders also makes enforce-
ment of such terms problematic. In particular, given the high amount
of late change, liability becomes uncertain.
A recent incentive for on-time delivery by commercial builders
is their full order books. If one programme has a delay, their other
programmes will as well. Because most contracts are fixed-price, com-
mercial shipbuilders want to move production as quickly as possible
to maximise profit.
Payment schedules constitute another incentive for commercial
builders (see Figure 4.3). Some 80 percent of the contract value is
paid on delivery, meaning shipyards carry some of the financial bur-
den, such as construction loan interest payments, of a vessel until the
ship is delivered.
By contrast, the DPA pays for military ships by milestones, typi-
cally ones that are easily verifiable and evenly spaced over the pro-
gramme. The milestones are usually negotiated between the DPA and
the contractor. The intent is that the payment plan is cash neutral (or
slightly negative) with respect to progress (meaning that the shipyard
is paid such that its costs for the programme are covered and it does
not have to resort to financing as do commercial shipyards). There-
fore, the military shipyards do not carry much financial burden for
construction.
Some aspects of the DPA Smart Acquisition process allow for a
more commercial approach to procurement (MOD, 2004). These
approaches include public-private partnership. For example, the Pri-
vate Finance Initiative transfers financial risk to the private sector.
Such approaches permit the DPA, in essence, to contract for a service
28 Monitoring the Progress of Shipbuilding Programs
Figure 4.3
Percentage of Total Value of Production Payments Paid to
Shipbuilders on Commercial Contracts by Production Phase
Contract signing
Start of production
Keel laying
Ship launch
Ship delivery
0 20 40 60 80 100
Percentage of value paid
RAND MG235-4.3
31
32 Monitoring the Progress of Shipbuilding Programs
Earned Value
Earned value is basically the value of the work accomplished up to a
point in time based on the planned or budgeted value for that work.
Project progress is measured against an earnings plan (in dollars or
man-hours) and is a variance analysis method that quantifies the
deviation of measures of actual performance from a standard. There
____________
1 Project Management Institute (2000). For more information on cost control, see Fisher
(1970).
33
34 Monitoring the Progress of Shipbuilding Programs
Table A.1
Notional Budgeted Cost of Work Schedule
Day
1 2 3 4 5 6 7 8 9 10
Pipes
(number) 5 10 15 20 25 30 35 40 45 50
BCWS $100 $200 $300 $400 $500 $600 $700 $800 $900 $1,000
The manager goes to the work site at day 5 and sees that the
project has spent $550 and that 30 pipes have been welded. From
Table A.1, it is clear that the BCWS is $500 at day 5, and the BCWP
is calculated to be $600 (30 pipes at $20 each). The finance depart-
ment gives data to the manager saying that $550 has been spent,
meaning the ACWP was $550.
Several measures of project progress can be calculated from these
data. To determine project deviation from original cost and schedule
goals, two measures of variance can be calculated. Cost variance (CV)
is the difference between the BCWP and the ACWP, whereas sched-
ule variance (SV) is the difference between the BCWP and the
BCWS. In the above example, we calculate the CV (or BCWP
ACWP) as $600 $550 = $50, which means the value of what the
project has accomplished is $50 more than what was spent. A positive
CV means that the project is under budget, while a negative CV sig-
nifies that the project is over budget. A zero CV indicates that the
project is exactly on budget. The SV (or BCWP BCWS) in the
above example equals $600 $500 = $100, which means the value of
what is accomplished to date is $100 greater than the value of what
had been planned for that date. A positive SV indicates a project is
ahead of schedule; a negative SV signifies it is behind schedule; and a
zero SV means it is exactly on time.
Two other measures are used to gauge project progress or com-
pletion rates. The schedule performance index (SPI) is equal to
BCWP divided by BCWS. In the above example, the SPI is 600 /
500, or 1.2, meaning the project has completed 120 percent of the
work originally planned for completion by that date. The cost per-
36 Monitoring the Progress of Shipbuilding Programs
EVM Assumptions
EVM Limitations
EVM Extensions
Table B.1, on the following pages, shows common EVM terms along
with their definitions and calculations.
41
42
Table B.1
EVM Calculations
43
APPENDIX C
On the pages that follow is the data collection form that was sent and
completed by the participating shipyards.
45
46 Monitoring the Progress of Shipbuilding Programs
RAND Contacts
John Birkler (310) 393-0411 x7607, [email protected]
John Schank (703) 413-1100 x5304, [email protected]
Mark Arena (703) 413-1100 x5383, [email protected]
6. What are the major milestones for which you track progress
(completion of design, start of steel fabrication, launch, delivery,
and so on)? Is the planning process based on a project/programme
life-cycle model? If so, please describe the model.
48 Monitoring the Progress of Shipbuilding Programs
Metrics
8. In the following matrix [Table 1], please indicate which progress
metrics your firm employs to monitor progress. For the primary
metric, fill in the box with a P. For a secondary metric(s), fill in
the appropriate box(es) with an S. For metrics that are not used,
leave the boxes blank. If you employ work subdivisions or meas-
ures of progress other than those listed in this table, please fill out
an equivalent table using your own divisions and metrics. In do-
ing so, please define the terms you use. If your metrics for gov-
ernment and commercial customers differ, please fill out the table
once for each set of customers.
9. Which progress metrics do you routinely report to customers?
Earned value
Percentage of estimated
hours expended
Work packages
Available schedule
float/contingency
Milestone/major activity
Percentage of estimated
Number of blocks/modules
installed
Number of compartments/
areas accepted by customer
49
50
Table 1 [from Question 8]Continued
Number of compartments/
areas completed
Other:
_______________________
Data Collection Form 51
10.Do you use different progress metrics for your own use than you
report to your customer?
YES _____ NO_____
If you answered YES, what metrics do you use, and how do they
differ from those you report to your customers?
Forecasting
12.How do you forecast completion dates using the metrics above?
14.Are your cost and schedule estimates updated during the course of
a project?
YES _____ NO _____
If you answered YES, how frequently are they updated and how
are the updates used?
Results
16.If possible, please provide an example of your primary progress
metric(s) plotted versus time for a recently completed ship.
How does the difference between the forecast delivery date at con-
tract and the actual date vary from ship to ship (i.e., is the differ-
Data Collection Form 53
ence fairly consistent from ship to ship, or does it vary for each
ship)?
18.What are the most common reasons for errors in the forecast
completion date?
20.How would you apportion the source of delays and cost increases
above original estimates for recent projects? Please fill out the
table below [Table 2] and add any other categories you think
appropriate.
Change
21.How are scope changes and change orders addressed with respect
to the schedule plan/control for a government customer?
Percentage of
total delay
Percentage of
total cost
increase
Percentage of
total number
of changes
Percentage of
total number
of changes
55
References
57
58 Monitoring the Progress of Shipbuilding Programs