Quantifying Engineering Project Scope

for Productivity Modeling

Lingguang Song1 and Simaan M. AbouRizk, M.ASCE2
Downloaded from ascelibrary.org by The City College of New York - CUNY on 12/14/23. Copyright ASCE. For personal use only; all rights reserved.

Abstract: A poor scope definition in an engineering design project disrupts project rhythm, causes rework, increases project time and
cost, and lowers the productivity and morale of the workforce. A quantitative measurement of the project scope is the basis for produc-
tivity modeling that involves the measurement, estimation, control, and evaluation of productivity. This paper proposes a conceptual
model, the quantitative engineering project scope definition 共QEPSD兲, to standardize the measurement of engineering project scope in
construction projects, within a computer aided design environment. The QEPSD quantitatively measures engineering project scope, in
terms of the complexity of design items by defining design categories and complexity functions appropriate to the particular discipline.
The proposed method was originally verified and implemented specifically for steel drafting projects. Actual data was analyzed and used
to demonstrate the benefits of historical data prepared using QEPSD for project scope definition. It was found that the new method led to
increased utilization of previously untapped values in historical data, improving the accuracy of project scope definition, and productivity
modeling. The paper concludes with a discussion of the potential benefits of adopting the QEPSD method, and its implications upon
various project management functions.
DOI: 10.1061/共ASCE兲0733-9364共2005兲131:3共360兲
CE Database subject headings: Project management; Productivity; Construction industry; Personnel management; Cost control.

Introduction be defined in a more detailed manner using work breakdown

structures. A survey conducted by the Construction Industry In-
Project scope management is critical to the success of an engi- stitute 共CII兲 shows that the current practice followed by design
neering design project. The scope management defines the spe- firms is to determine engineering scope and progress by relating
cific design items or services required to meet the project objec- them to the number of design documents for each design disci-
tives. The Project Management Institute 共PMI兲 defines the project pline 共CII 2001; Diekmann and Thrush 1986兲. Essentially, this
scope management process as involving initiation, scope plan- method treats the output of the design process as any paper design
ning, scope definition, scope verification, and scope change con- document such as a drawing or specification. The scope for a new
trol 共PMI 1996兲. Design engineering firms define project scope as project is defined subjectively based on the output produced in the
encompassing different engineering design phases. According to previous design phase and data taken from historical projects. The
American Institute of Architects 共AIA兲 Document B141, these project scope is measured by an estimate of the quantity of docu-
phases include schematic design, design development, construc- ments to be produced, and the progress is measured by the actual
tion document, bidding or negotiation, and construction or con- quantity of documents produced to date.
tract administration 共AIA 1987兲. Each scope management process However, due to the current proliferation of computer aided
usually occurs at least once in each design phase. Scope definition design 共CAD兲 tools, a particular representation of the physical
involves subdividing the overall project deliverables into smaller design deliverables as documents is no longer relevant. In the
and more manageable components, resulting in better project CAD environment, a product model is created and verified on a
planning and control. Scope definition plays an important role in computer and then the model and any of its components can be
the management of design processes. After some degree of pre- selected and printed to the desired drawing size on a plotter. This
project planning and project authorization, the project scope can renders the measurement of the quantity of drawings or paper size
irrelevant. There is no standard definition for the contents and
1
Research Associate, Dept. of Civil and Environmental Engineering, complexities of design documents. Armentrout 共1986兲 argues that
Univ. of Alberta, Edmonton AB, Canada T6G 2G7. E-mail: these units of measure, such as drawings, procurements, and
[email protected] specifications, do not truly reflect the total service rendered by an
2
Professor, Dept. of Civil and Environmental Engineering, Univ. of engineering organization. As projects become much larger, they
Alberta, Edmonton AB, Canada T6G 2G7. E-mail: [email protected] are almost impossible to compare in any orderly or consistent
Note. Discussion open until August 1, 2005. Separate discussions way. Nonetheless, engineering firms still must use this system of
must be submitted for individual papers. To extend the closing date by measurement to quantify the project scope for the sake of project
one month, a written request must be filed with the ASCE Managing
planning and control, despite the obvious drawbacks associated
Editor. The manuscript for this paper was submitted for review and pos-
sible publication on July 17, 2003; approved on April 1, 2004. This paper with it. Engineers normally evaluate the content and complexity
is part of the Journal of Construction Engineering and Management, of design documents subjectively and account for this bias
Vol. 131, No. 3, March 1, 2005. ©ASCE, ISSN 0733-9364/2005/3-360– through a simple document count. Knowledge utilized during the
367/$25.00. evaluation process is generated from personal judgment based on

360 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MARCH 2005

J. Constr. Eng. Manage., 2005, 131(3): 360-367

 historical data, and its validity resides with an experienced indi- methods are limited by the level of detail at which they can be
vidual. An implicit measurement, which does not use consistent applied and the amount of project information which they can
quantitative methods, makes a full understanding of the project represent. Therefore, they have limited use for project scope man-
scope difficult, and causes misunderstandings regarding project agement, such as progress measurement, schedule control, and
scope during project planning and control. The current practice of cost control at the discipline level.
project scope definition in the design industry is much less effec- Studies focusing on performance evaluation and improvement
tive as a medium for carrying project scope information regarding at the postproject stage normally do not explicitly and quantita-
project planning and control. tively measure the project scope. The Construction Industry Insti-
Poor scope definition is recognized by industry practitioners as tute 共CII 1986兲 proposed a system for evaluating design effective-
one of the leading causes of project failure, adversely affecting ness. The method is based on combining the weights and ratings
projects in the areas of cost, schedule, and operation 共Cho and of seven evaluation criteria into a single performance index which
Downloaded from ascelibrary.org by The City College of New York - CUNY on 12/14/23. Copyright ASCE. For personal use only; all rights reserved.

Gibson 2001兲. This is due to the fact that the project scope defi- describes the design effectiveness. Armentrout 共1986兲 discussed a
nition is the reference point for developing estimates and sched- method of measuring performance by tracking several indices af-
ules, coordinating teamwork, applying control strategies, and fecting specific aspects of the engineering organization, in order
evaluating engineering performance. An interesting observation is to evaluate design effectiveness.
that the fundamental information underlying all these project The lack of quantitative information in those methods is a
management functions is regarding engineering productivity. The serious deficiency for engineering project planning and control.
lack of a quantitative and reliable method for defining the project The project scope definition component in several reported engi-
scope has been a major obstacle for modeling engineering pro- neering project control systems was based on the delineation of
ductivity, and therefore causes collateral ineffectiveness in the work hours according to a cost accounts code, instead of the
management of the design process. Thus, a measurement of the products of the design process 共Thomas et al. 1999兲. Thomas et
project scope for the productivity modeling process, which in- al. 共1999兲 created a conceptual model for measuring engineering
volves productivity measurement, estimating, control, and perfor- productivity during the construction document phase. The mea-
mance evaluation, is of great importance and interest to engineer- surement of the design output accounts for the differences among
ing firms.
all design documents, such as detail drawings, specifications, and
This paper proposes a conceptual model for measuring the
other documents, by using conversion factors. However, the ac-
project scope of construction engineering projects for the purpose
curacy of any measurement is compromised due to the use of
of productivity modeling. This scope definition method, the quan-
CAD tools and because of the lack of a standard definition of
titative engineering project scope definition 共QEPSD兲, measures
content and design document complexity, as previously men-
project scope in terms of complexity of each design item at a
tioned. The CII Engineering Productivity Measures Research
design discipline level, within a CAD environment. The potential
Team 共CII 2001兲 identified that there was no standard measure-
benefits and implications of applying this technique in engineer-
ment of productivity in the engineering phase for internal im-
ing productivity modeling are explained. The scope of this paper
provement and external benchmarking. The research team pro-
is limited to steel drafting as a pilot to verify the applicability of
posed a model focused on measurable, installed quantities, for the
this new approach. Steel drafting is an engineering function
measurement of the design output, such as the length of a pipe or
within the detailed design level of steel structural design. Based
the weight of steel as per the design. This method was applied to
on the architectural and structural design, the drafting process
produces detail drawings for fabrication and erection in compli- the discipline level during the detailed design phase of a project.
ance with the project requirements, fabricator standards, and erec- Raw productivity, which is measured based on installed quanti-
tor standards and specifications. The benefits of applying the ties, is subjectively adjusted by three influencing factors: input
QEPSD method to steel drafting are demonstrated with a focus on quality, scope and complexity, and design effectiveness. However,
project scope definition and estimating respects using historical our experience with steel drafting projects showed that the in-
data. The method is illustrated using actual data collected from stalled quantity could be misleading due to the unaccounted de-
the drafting department of our collaborating company. sign complexity and weight values. The evaluation of the scope
and complexity factor on a project level lacks too much accuracy
to be used for productivity modeling.
The review of available literature resulted in additional obser-
Literature Review
vations. First, a project scope definition method must be devel-
oped before many problems in productivity modeling, project
In spite of some awareness of problems in the definition of engi-
planning, and control can be addressed. Second, the scope defini-
neering project scope, there have been only limited studies in
tion, quantitatively, is required to measure the design outputs con-
response to the industry’s growing need. This need continues to
sistently, rather than the design inputs. Finally, historical project
increase with further changes in design methods, tools and project
data contain a large amount of untapped values and should be
planning requirements.
used to extract predictive information for project scope definition.
Traditional cost modeling methods, such as the unit method,
cube method, superficial area method, and approximate quantities
method, measure the project scope by function unit, square meter
of area, or cubic meter of volume 共Jaggar et al. 2002兲. These units Quantitative Engineering Project Definition Method
measure project scope at the project level for the purpose of cost
estimating only. However, according to a survey conducted by the It is necessary to clarify the concept of project scope definition
CII, 91% of the surveyed companies focus on the discipline level due to confusion arising out of design input and output measure-
for project control, due to the fact that most design firms drive ment methods. A decision can subsequently be made regarding
accountability to the department or discipline level on projects the measurement of project scope and the level of detail that
共CII 2001兲. These measurement units based on cost modeling should be measured.

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MARCH 2005 / 361

J. Constr. Eng. Manage., 2005, 131(3): 360-367

 Design Input, Output, and Project Scope
Engineering design creates and transforms ideas and concepts
into a product definition that will satisfy customer needs. A civil
engineering design is accomplished through the collaborative ef-
forts of a number of different design disciplines, such as architec-
tural, civil/structural, mechanical, electrical, piping, and project
management. Engineering work hours represent a major resource
for design inputs in the design process. Many other resources are
consumed, such as computer time and other equipment use. Also,
certain materials are correlated to engineering work hours, so
Downloaded from ascelibrary.org by The City College of New York - CUNY on 12/14/23. Copyright ASCE. For personal use only; all rights reserved.

their cost can be accounted for by assigning an appropriate engi-

neering hourly rate. Work hours are, thus, an appropriate measure
of design input and are traceable. Most engineering companies
have a cost accounting system or time-sheet system that keeps
track of work hours. However, the input measured by work hours Fig. 1. Work breakdown structure for project scope definition
should not be interpreted as the project scope. This confusion
results in a project scope measured by hours or monetary value.
such as crew qualification and the quality of preproject planning.
The disadvantage of such an approach is that it tends to use inef-
Additionally, overall project complexities 共e.g., structure type,
ficiencies to build future projects.
type of construction, and climatologic design considerations兲 are
The design output can be viewed as information, or as design
not included in this detailed item-level measurement method. In-
documents. For example, the output of steel drafting is a complete
stead, the environment variables and overall project complexity
set of fabrication drawings and erection drawings. From the own-
are considered as factors affecting engineering productivity and
er’s point of view, the output is complete technical information,
were studied separately from this study 共Song et al. 2003兲.
allowing steel fabricators and erectors to accomplish their assign-
A work breakdown structure 共WBS兲 is a frequently used tech-
ments. When a project has been completed, the project scope can
nique in project scope definition to decompose the project into
be precisely defined using the design output. For new projects,
measurable elements. The WBSs used specifically for the pro-
scope definition is normally obtained from an expert who relies
posed method decompose a project to the design item level using
on his or her own judgment and similar past projects. Therefore,
the project and the product breakdown structure, as shown in Fig.
analyzing historical projects and their outputs is extremely impor-
1. The WBS in Fig. 1共a兲 divides the project at the design disci-
tant to project scope definition for future projects. It is easy to
pline level first. This is the level most design companies focus
describe the design output, but quantifying the design output is
upon for project control. In order to uniquely quantify a disci-
difficult in practice. This difficulty has driven our research in this
pline’s work scope, more levels of decomposition may be re-
paper.
quired. Within the structural discipline, for example, further divi-
sion of structural concrete design and structural steel design is
Scope Definition Based on Complexity of Design Items possible. The product breakdown structure was designed to rep-
resent the final product model in order to facilitate the quantita-
In a construction project, the project scope is defined by the quan- tive measurement of design items. Product models are designed
tities of construction items within each labor discipline that can for each discipline to represent its final product. Fig. 1共b兲 is an
be easily measured, such as the volume of earth hauled or con- example of structural steel design. The structural steel design sub-
crete poured or the length of pipe installed. Design information project can be further divided into divisions representing different
from different design disciplines is carried through to the con- physical locations, each of these divisions containing more steel
struction work itself, and finally synthesized and materialized by pieces with certain material requirements.
the constructed facility. The engineering output can be measured This WBS model is a structured approach to manage the
naturally based on the quantity of design items, such as a beam or project scope, but does not necessarily result in a quantitative
a window, within a discipline. For example, rather than using the measure of work scope that can be easily communicated to all
total quantity of concrete drawings as a measure, the project project participants. As mentioned above, design items vary con-
scope for structural concrete design can be measured based on the siderably in terms of complexities. Complexities are evaluated
quantity of concrete needs that will be designed. However, a con- based on two functions in the QEPSD method: design category
sideration should be given to the configuration and complexity of and category complexity.
the design items in terms of design efforts required. A simple
count of the physical quantities would be misleading. For ex-
Design Category
ample, the design of a concrete wall, slab, or column represents
different degrees of complexity to engineers. The project scope The first step toward a measurement of design item complexities
can be measured by the sum of design items in terms of their is to group similar design items within a design discipline. This
relative complexity when compared to a particular design item as grouping process defines a list of design categories. A design
a standard unit. Applying this method to various design items, the category is a variable that describes distinct groups of design
design output can be measured uniformly into an abstract unit of complexities. Design items in one category will share the same
measure. This is analogous to a “unitization” scheme used in attributes with regards to complexity. Design item descriptions
quantifying industrial fabrication shop work 共Alfeld 1988兲. used by an engineer to describe item functions within an overall
It is necessary to differentiate between the complexity of de- product, provide a good starting point for the definition of design
sign items and the environment where these items are produced. categories. For example, design categories with the HVAC design
This proposed method does not measure environment variables, discipline may include ducts, air devices, access doors, dampers,

362 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MARCH 2005

J. Constr. Eng. Manage., 2005, 131(3): 360-367

 Table 1. Design Categories for Steel Drafting
ID number Description Complexity variable a b
1 Column Number of fittings 1.00 0.67
2 Beam Number of fittings 1.13 0.53
3 Girder Number of fittings 2.20 1.33
4 Bracing Number of fittings 1.40 0.53
5 Girt Number of fittings 1.13 0.67
6 Purlin Number of fittings 1.13 0.53
7 Hanger Number of fittings 1.53 0.53
8 Support Number of fittings 2.93 0.53
Downloaded from ascelibrary.org by The City College of New York - CUNY on 12/14/23. Copyright ASCE. For personal use only; all rights reserved.

9 Monorail–straight Number of fittings 1.53 0.53

10 Monorail–curved Number of fittings 3.47 0.53
11 Crane rail Single piece 2.80 0.00
12 Stiffener Single piece 0.67 0.00
13 Gusset Single piece 4.00 0.00
14 Sag Rod Single piece 1.53 0.00
15 Truss Number of fittings 0.67 1.40
16 Frame Number of fittings 0.67 1.40
17 Conveyor gallery Number of fittings 0.67 1.40
18 Utility bridge Number of fittings 0.67 1.40
19 Platform Number of fittings 0.67 1.40
20 Walkway Number of fittings 0.67 1.40
21 Stair Number of fittings 0.67 2.67
22 Stair tread Number of fittings 0.00 0.67
23 Handrail–straight Number of fittings 0.67 0.53
24 Handrail–sloping Number of fittings 0.67 1.00
25 Handrail–circular Number of fittings 0.67 1.67
26 Ladder no cage Number of fittings 0.67 0.40
27 Ladder with cage Number of fittings 0.67 0.33
28 Checker plate Number of fittings 6.00 1.33
29 Toe plate Single piece 2.00 0.00
30 Safety gate Single piece 4.13 0.00

fan units, and architectural features. A design item should be clas- ables. For each category, complexity variables can be identified
sifiable into one of the defined categories. The design item cat- and the relationship between the variables and the complexity can
egories and classifications may be expressed mathematically as be formulated:
follows. Let x jk, k = 1 , 2 , . . . , s, be all complexity variables for the de-
C1 , C2 , C3 , . . . , Cm is a set of mutually exclusive and exhaus- sign category C j. The complexity qi of a piece pi in this category
tive categories so that Ci 艚 C j = ⌽ and C1 艛 C2 艛 C3 艛 ¯ 艛 Cm is is given by
the entire design item space. The design item classification is to
assign a design item, p j 兵p j : j = 1 , . . . , n其, to one of M categories
qi = f j共x j1,x j2, . . . ,x js兲 共1兲
兵Ci : i = 1 , . . . , m其, so that p j 苸 Ci.
In our study, the steel drafting design categories grouped simi- For example, the number of fittings is a complexity variable
lar steel pieces together based on their function within a steel within the column category in steel drafting. A complexity func-
structure. The developed drafting category consisted of 30 catego- tion f must be defined for each design category. Experienced
ries, as shown in Column 2 of Table 1. An effort in attempting to engineers may help in defining these functions based on their
standardize the naming convention used by draftspersons is cur- experience.
rently underway in the collaborating company. In practice, more than one variable may affect a design cat-
egory’s complexity. However, for each design category, if it is
properly defined, one dominant variable may adequately describe
Category Complexity Functions
the complexity associated with all items in that category. For
Considerable variability with regards to design complexity may example, within the handrail category in steel drafting, the type of
still exist within each category. This requires a more in-depth handrail and number of fittings affect the design complexity. To
evaluation of design complexity, resulting in a definition of cat- reduce the dimension of the relationship, the type of handrail is
egory complexity variables and complexity functions. Category considered as one of the definitions of the design categories.
complexity variables are factors describing the complexity of de- Handrails are classified as three categories, as shown in the first
sign items within a design category. Complexity functions evalu- column of Table 1: “Handrail–straight,” “Handrail–sloping,” and
ate a design item’s complexity based on category complexity vari- “Handrail–circular.” A dominant complexity variable is identified

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MARCH 2005 / 363

J. Constr. Eng. Manage., 2005, 131(3): 360-367

 Table 2. Complexity Factor Table Table 3. Results of Correlation Analysis
Complexity variable value Measure unit Drawing Piece Weight Drafting unit Hours
共number of fittings兲
Drawing 1 — — — —
Category a b 0 1 2 3 Piece 0.48 1 — — —
Column 1.00 0.67 1.00 1.67 2.34 3.01 Weight 0.50 0.45 1 — —
Beam 1.13 0.53 1.13 1.66 2.19 2.72 Drafting unit 0.81 0.51 0.79 1 —
Gird 2.20 1.33 2.20 3.53 4.86 6.19 Hours 0.75 0.53 0.67 0.88 1
Bracing 1.40 0.53 1.40 1.93 2.46 2.99

due to the large quantity of design items and the difficulty of

Downloaded from ascelibrary.org by The City College of New York - CUNY on 12/14/23. Copyright ASCE. For personal use only; all rights reserved.

for each category. Column 3 of Table 1 shows the complexity evaluating complexity. A manual count is inefficient, if not im-
variable defined for each drafting category. “Number of fittings” possible. Currently, a variety of CAD software is used in almost
refers to the quantity of detail materials, or steel fittings, on a steel every engineering design discipline. The proposed QEPSD
piece. “Single piece” indicates that the complexity of the piece is method is designed to work in a CAD environment. The CAD
measured by a single design item. system captures vast amounts of data in an electronic format. This
Based on the draftsperson’s experiences and the accumulative creates a unique opportunity in automating the measurement of
nature of drafting design, the relationship between a dominant design outputs from past projects. Data required for measuring
complexity variable and the complexity of a single piece was design output is normally recorded in a project CAD model. Most
assumed to be a linear function, which is commercial CAD systems have the capability of interfacing with
other software systems; exporting design data in a text format is a
f j共x j兲 = a j + b jx j 共2兲 minimum requirement. Data exchange interface can be imple-
where a j⫽base complexity value for category C j and mented to transfer the design data from a CAD model to a data-
b j⫽coefficient for the complexity variable x j. A standard design base system. The complexity evaluation algorithm can be encap-
item can be defined as an abstract unit of measurement. The quan- sulated within a software module to automate the quantification
tification is a conversion based on weighting other design items process.
for their degree of complexity compared to the standard unit. For Our collaborating company uses specialized CAD drafting
steel drafting, a simple steel column with no fitting is defined as a software, StruCAD 共StuCAD user manual 2003兲, for its steel
standard unit, called a “drafting unit.” To assist the weighting, the drafting work. File exporting, data exchange interfaces, and data-
degree of complexity can be compared at the design process level. base systems have already been used by the company for material
The design process involves multiple stages of development, re- listing, and they were used in this study. The complexity evalua-
view, and revision. To facilitate the definition of a j and b j in Eq. tion algorithm was built using structured query language, and was
共2兲, the steel drafting process is broken down into wire frame integrated into the database system and an interface for users.
modeling, bill of material, two-dimensional drawing, electronic Over 1 million steel pieces from projects in the last 5 years were
drawing, check and administration 共Allouche and Song 2003兲. a j quantified in drafting units over the course of our study.
and b j are defined by the sum of the complexity evaluated at the
process level. The systematic decomposition of a project into
clearly defined design items and the use of process modeling Quantitative Engineering Project Scope Definition
makes the definition of complexity functions easier and more ac- Validation
curate. Additionally, the user can gain confidence in using this
method by securing the quantification procedure. The proposed conceptual model aims at quantitatively measuring
To illustrate the result from the unit measure, a sample com- the engineering project scope for construction projects. Experi-
plexity factor table is shown in Table 2. A bracing in the bracing enced engineers define the design category and complexity func-
category with two fittings is 2.46 units of work, according to the tions. To verify its capability and accuracy, the model must be
table. The total adjusted quantity of a project output, or project tested on actual projects for each discipline in the engineering
scope, in drafting unit is given by: design. However, this method has not been used previously.
Therefore, in this pilot study, only steel drafting was studied for
n
the QEPSD validation. Historically, the weight of steel, the quan-
Qtotal = 兺
i=1
qi 共3兲 tity of drawings, and the quantity of steel pieces were used to
measure steel drafting project scope. These records will be com-
where Qtotal⫽project scope measured by drafting unit; pared to the newly developed drafting unit. The criterion of the
qi⫽complexity of a piece pi measured by drafting unit, which is comparison is that a good measurement of project scope has a
defined in Eqs. 共1兲 and 共2兲; and n⫽total number of pieces in a high correlation to the input, which is in work hours. An analysis
steel drafting project. correlating the different measurements with the inputted work
hours is performed to compare the relative effectiveness.
Data from a total of 69 steel drafting projects has been col-
Automation of Quantification Process
lected for the correlation analysis. Scatter diagrams were con-
The quantification procedure is defined by the design categories, structed for each measured unit and correlation coefficients were
category complexity variables, complexity functions, and a stan- calculated and compared, as shown in Table 3. The correlation
dard design unit. Precisely quantifying historical projects using analysis shows that the drafting unit outperforms other commonly
the standard unit of measurement can help accumulate knowledge used measures. The correlation value for the drafting unit R is
in project scope definition for future projects. However, the quan- 0.88, which is the highest value. The value rankings following
tification process can be extremely tedious and time consuming this are the quantity of drawings, the weight of steel, and the

364 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MARCH 2005

J. Constr. Eng. Manage., 2005, 131(3): 360-367

 quantity of steel pieces. The major reasons behind this ranking are et al. 1995兲, and neural network models 共Creese and Li 1995兲,
the use of CAD tools, and the irrelevancy of draftspersons’ work may be used to measure a new project’s scope. A comprehensive
regarding the physical weight. A t test at the 95% level shows that discussion of these estimating techniques falls outside the scope
the correlation between the drafting unit and work hours is statis- of this paper. Estimating based on historical data is an alternative
tically significant. Thus, the drafting unit is considered to be the to the existing method that is based on personal judgment. One of
best measure of project scope, and the most accurate predictor of the applications of QEPSD in project scope definition and esti-
drafting work hours based on other units of measure. mating for steel drafting project using historical data is illustrated
By definition, the coefficient of determination 共R2兲 represents with a case study.
the proportion of variation in the dependent variable that has been
explained or accounted for by an independent variable. The quan-
tity of drafting unit accounts for about 77.4% of the drafting man Case Study
Downloaded from ascelibrary.org by The City College of New York - CUNY on 12/14/23. Copyright ASCE. For personal use only; all rights reserved.

hours required. An explanation of the residual is expected by

other environment variables and overall project complexity fac- The selected project is a unit price contract, involving the drafting
tors as previously defined 共e.g., degree of cloning, draftsperson of structural and miscellaneous steel of an industrial facility.
experience, and engineering standard兲 共Song et al. 2003兲. Under a unit price contract, the contractor must prepare a detailed
cost for each category defined by the owner, based on the esti-
mated quantities given in the contractual documents. An estimate
Project Scope Definition with Quantitative of the total number of hours for internal scheduling use is desir-
Engineering Project Scope Definition able. It is not uncommon that during the bidding stage, architec-
tural and structural design has not yet been completed. Due to the
Engineers must determine a project scope using only the informa- absence of detailed engineering drawings for quantity takeoff for
tion available at the time for project scope definition. The proper this project, quantities assumed based on a survey given by the
approach for quantifying project scope of a new project is a func- owner, are used for defining the project scope and category unit
tion of the availability of usable information. In the light of this cost. The quantity, in terms of weight, is the only information
fact, both project scope definition possessing complete project available for preparing the estimate. The total weight of the
information and scope definition possessing incomplete project project is 120.50 t, in which 41.11 t were drawn by the collabo-
information were studied. rating company, and 79.39 t were subcontracted to two other
drafting companies. Our case study is limited to analyze only the
part of the project drawn by the collaborating company. For con-
Project Scope Definition with Complete Project
fidentiality reasons, productivity data used in this case study were
Information
scaled.
This situation may be encountered by some drafting disciplines In this project, 15 unit-price categories were listed in the con-
during the construction document phase. For example, in most tract document, as shown in Column 2 of Table 4. For the selected
lump sum contracts, the steel drafting begins after the architec- project, the quantities measured by weight in tons are available
tural and structural design, and uses structural arrangements and for each unit-price category. The weight is assumed to be accurate
layout drawings as a design basis. Project scope can be measured in this case study, so the actual weight of each unit-price category
directly, using the described QEPSD method, based on informa- is used, as shown in Column 3. An estimator subjectively pre-
tion from a manual or an automated quantity takeoff from engi- dicted a category-specific drafting productivity level in work
neering drawings or a CAD model, coupled with some estima- hours per ton 共Column 4兲, in which the complexity of each unit-
tions on the quantity specified by category complexity variables price category and profits were accounted for by his or her expe-
共e.g., number of fittings兲. Therefore, this will not be further inves- riences. Thus, the unit cost for a unit-price category is the sum of
tigated. the productivity and a predefined hourly rate.
This estimating problem can be approached alternatively using
historical data to evaluate the complexity and unit cost for each
Project Scope Definition with Incomplete Project
unit-price category. A total of 216 similar types of projects were
Information
quantified and stored in a database system using the QEPSD
To relate scope definition to quantities of design items, the scope method developed for steel drafting. Queries were performed to
of the project must be completely defined. Such is not the case for find out the ratios of the drafting unit quantity to the weight of
most design disciplines in engineering projects. During schematic each unit-price category from the database system. For a specific
design and design development, the scope is described in a vague category, this ratio will vary from project to project. The uncer-
manner that prevents any direct measure of the final product. In tainties of this ratio can be modeled by fitting a standard statistical
this case, historical data and past experiences are the best infor- distribution to historical data. BestFit 共BestFit users manual 1999兲
mation to use to estimate the project scope quantitatively, as far as was used for the data-fitting analysis. Either normal or uniform
these are available and relevant. Obviously, the confidence in any distribution was found to reasonably represent the distributions
estimate will be higher if it is based on relevant past experience, underlying the sample data for a category. The distribution type
particularly if the new project can be defined in some assured and parameters are listed in Column 6. In order to get a point
details. QEPSD can help to quantify historical projects for this estimate of the work scope, the mean value of each category’s
purpose. drafting unit–weight ratio was used. The quantity of work mea-
In the preproject planning phase, if a facility’s capacity infor- sured in drafting units 共Column 7兲 is the product of the mean
mation is all that is available, for example, the capacity of a value in Column 6 and the weight of each category in Column 3.
concrete tank or the area of an office building, then simple statis- A drafting productivity prediction model based on an evalua-
tics, equations, or other advanced models derived from historical tion of various productivity-influencing factors was used to pre-
data prepared using QEPSD, such as the six-tenths rule 共Steward dict the productivity value for this project 共Song et al. 2003兲. The

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MARCH 2005 / 365

J. Constr. Eng. Manage., 2005, 131(3): 360-367

 Table 4. Case Study: Unit Price Contract Steel Drafting Project
ID Weight Productivity Estimated hours Unit per ton Quantity Estimated hours Productivity
number Category description 共t兲 共h/ton兲 共h兲 共unit/t兲 共unit兲 共h兲 共h/t兲
1 Rolled shapes 15– 31 kg/ m ⬍ 2,744 mm 2.44 26.07 63.61 Normal共73.38,35.55兲 178.68 82.19 33.69
2 Rolled shapes 32– 61 kg/ m ⬍ 2,744 mm 3.46 16.53 57.19 Normal共32.82,12.38兲 113.39 52.16 15.07
3 Rolled shapes 62– 100 kg/ m ⬍ 2,744 mm 0.29 14.46 4.19 Normal共12.34,4兲 3.62 1.67 5.74
4 Rolled shapes 32– 61 kg/ m ⬎ 2,744 mm 14.37 14.39 206.78 Normal共11.25,2.71兲 161.69 74.38 5.18
5 Rolled shapes 62– 100 kg/ m ⬎ 2,744 mm 3.32 12.48 41.43 Normal共5.49,1.8兲 18.22 8.38 2.52
6 Rolled shapes 101– 150 kg/ m ⬎ 2,744 mm 5.90 9.83 58.00 Normal共5.46,2.31兲 32.21 14.82 2.51
7 Bracing–WT section ⬍2,744 mm 1.16 27.20 31.55 Normal共86.07,43.03兲 99.50 45.77 39.46
Downloaded from ascelibrary.org by The City College of New York - CUNY on 12/14/23. Copyright ASCE. For personal use only; all rights reserved.

8 Bracing–WT section ⬎2,744 mm 2.43 24.13 58.64 Normal共25.39,13.96兲 61.65 28.36 11.67
9 Girt⬍ 30 kg/ m ⬎ 2,744 mm 0.17 22.54 3.83 Uniform共19.2,27.2兲 4.04 1.86 10.93
10 Girt⬎ 30 kg/ m ⬎ 2,744 mm 0.42 16.14 6.78 Uniform共3.2,7.75兲 2.31 1.06 2.53
11 Web stiffeners W14–W18 section 0.02 58.99 1.18 Normal共120,50.08兲 2.28 1.05 52.44
12 Web stiffeners⬎ W18 section 0.05 58.99 2.95 Normal共92.5,25.2兲 5.00 2.30 46.00
13 Ladder 0.92 40.49 37.25 Normal共90.26,19.39兲 82.68 38.03 41.34
14 Handrail–straight 4.40 44.66 196.50 Normal共104,20.2兲 457.39 210.40 47.82
15 Handrail–sloped 1.76 44.66 78.60 Normal共192,59.9兲 337.34 155.18 88.17
Total 41.11 — 848.50 — 1,560.00 717.62 —

estimated productivity is 0.46 h per drafting unit. The estimated 829.30 h. The 80th percentile of the project completion time is
hours 共Column 8兲 for each category based on the mean value of 792.56 h.
the drafting unit–weight ratio is the product of the quantity in After the completion of this project, QEPSD measured the
Column 7 and the estimated productivity value. For bidding pur- project scope as 1,537.01 drafting units. That is a total of 341
poses, the productivity measured in work hours per drafting unit drawings. The actual drafting hours collected through the compa-
is converted to work hours per ton in Column 9. ny’s office time sheet system was 676.50 h. This is inside the
The Monte Carlo simulation technique was used to evaluate 90% confidence interval. The 80th percentile indicates an overes-
the risk and uncertainty of the estimate 共Ahujia et al. 1994兲. The timate of 17.16% of the total actual hours when using the new
work hours of each category were calculated as the product of the approach that is based on the drafting unit. The model output is
productivity in work hours per drafting unit, the weight, and the considered to be accurate. The results obtained from historical
unit–weight ratio. The experiment was implemented in Microsoft data are different than those obtained from the estimator’s esti-
Excel. Fig. 2 shows the histogram of the project’s total hours and mate. Unlike the existing estimating method, the new approach
the probability density function of a fitted normal distribution for obtains the results by separating the estimate of a project scope
the total hours showing a mean value of 722.00 h, and a 90% using the QEPSD method and the estimate of productivity using
confidence level that the total hours are between 614.67 and appropriate influencing factors. More accurate estimates can be

Fig. 2. Probability density function for project total hours

366 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MARCH 2005

J. Constr. Eng. Manage., 2005, 131(3): 360-367

 achieved using this structured estimating approach than using the steel drafting projects. It leads to increased utilization of untapped
estimator’s subjective judgment. Moreover, the result of this ap- values in historical data for project scope definition and produc-
proach is an estimate of actual productivity and hours to be con- tivity modeling. It improves the common understanding of the
sumed, in which the profit is considered separately. It establishes work scope and the accuracy of estimating through project par-
a baseline for scheduling and project control. ticipants. It holds significant potential as a force in improving the
project management process. This pilot study offers insights into
the implementation of this method for measuring project scope in
Conclusion other engineering disciplines. The QEPSD method will be applied
to different design disciplines, and its applicability will be further
The measurement at the design item level suggested by QEPSD verified.
allows a quantitative indication of project scope in terms of the
Downloaded from ascelibrary.org by The City College of New York - CUNY on 12/14/23. Copyright ASCE. For personal use only; all rights reserved.

design items’ complexities from a bottom up approach. The ap-

proach presents a number of good characteristics: Acknowledgments
• the complexity of design items has a high correlation to the
work hours; This work was supported by the Natural Sciences and Engineer-
• the complexity can be counted with properly defined design ing Research Council of Canada under Grant No. CRD 226956-
units; 99. The writers wish to thank Dwayne Hunka, Paul Zubick, and
• the measure is quantitative and consistent; and Mark Scott of Waiward Steel Fabricators Ltd. for their input and
• the method is practical to use in a CAD environment. support.
Many problems associated with measuring project scope and
engineering productivity can be alleviated and resolved with the
quantitative measurement of project scope using QEPSD. Engi- References
neering productivity can be conveniently measured by work hours
per unit of design. Both scope measurement and productivity Ahuja, H. N., Dozzi, S. P., and AbouRizk, S. M. 共1994兲. Project manage-
measurement address the need for estimating, scheduling, project ment: Techniques in planning and controlling construction projects,
2nd Ed., Wiley Chichester, U.K.
control, and performance evaluation. These project management
Alfeld, L. E. 共1988兲. Construction productivity–on–site measurement and
functions need to be revisited and updated accordingly with the
management, McGraw–Hill, New York.
new measurements. The following discussion presents some of Allouche, M., and Song, L. 共2003兲. “Analytic approach for productivity
the implications and benefits of adopting the proposed method. modeling.” Internal Reports, Department of Civil & Environmental
Engineering, Univ. of Alberta, Alta., Canada.
American Institute of Architects 共AIA兲. 共1987兲. “Standard form of agree-
Estimating and Scheduling
ment between owner and architect.” AIA Document B141, Washing-
Estimating and scheduling the engineering design process is ton, D.C.
highly subjective, as it is affected not only by the quantity of Armentrout, D. R. 共1986兲. “Engineering productivity management and
work, but also the productivity that can be achieved. The project performance measurement.” J. Manage. Eng., 2共3兲, 141–147.
scope can be measured using the QEPSD method; productivity BestFit user manual. 共1999兲. BestFit, Palisade Corporation, Newfield,
influencing factors can be evaluated using historical productivity N.Y.
Cho, C. S., and Gibson, G. E. 共2001兲. “Building project scope definition
data prepared by the QEPSD method 共Song et al. 2003兲.
using project definition rating index.” J. Arch. Eng. Manage., 7共4兲,
115–125.
Project Control Construction Industry Institute 共CII兲. 共1986兲. “Evaluation of design effec-
tiveness.” Rep. No. 8–1. Univ. of Texas at Austin, Austin, Tex.
During detailed design phase, subjectivities in progress reporting Construction Industry Institute 共CII兲. 共2001兲. “Engineering productivity
can be removed using the QEPSD method to automate progress measurement.” CII publication 156-1, Austin, Tex., 1–3.
measurement in a CAD environment. Quantities can be rolled up Creese, R. C., and Li, L. 共1995兲. “Cost estimation of timber bridges using
to any level in the WBS for project planning and control. This neural networks.” Cost Eng., 37共5兲, 17–22.
allows for the monitoring and control of an engineering project, at Diekmann, J. E., and Thrush, K. B. 共1986兲. Project control in design
a greater detail than only at the project level. engineering, Univ. of Colorado Press, Boulder, Colo.
Jaggar, D., Ross, A., Smith, J., Love, P., and Ross, A. 共2002兲. Building
design cost management, Blackwell Publishers, Oxford, U.K.
Performance Evaluation Project Management Institute. 共PMI兲. 共1996兲. A guide to the project man-
The proposed approach measures productivity quantitatively in agement body of knowledge, Upper Darby, Pa.
terms of work hours per unit of work. However, it does not mea- Song, L., Allouche, M., and AbouRizk, S. M. 共2003兲. “Measuring and
sure the design effectiveness, which may be measured by con- estimating drafting productivity.” Proc., 2003 Construction Research
structability, rework rate, and other field complications arising Congress, Honolulu.
Steward, R. D., Wyskida, R. M., and Johannes, J. D. 共1995兲. Cost esti-
from the engineering design. At the completion of a project, the
mator’s reference manual, 2nd Ed., Wiley, Chichester, U.K.
quantitative productivity measurement can be combined with StruCAD user manual. 共2003兲. StuCAD AceCAD Software Ltd., Derby,
evaluation factors, measuring time, cost, quality, and safety per- U.K.
formance, to give a more comprehensive evaluation of the engi- Thomas, H. R., Korte, Q. C., Sanvido, V. E., and Parfitt, M. K. 共1999兲.
neering performance. “Conceptual model for measuring productivity of design and engi-
The QEPSD method has been implemented and verified on neering.” J. Arch. Eng. Manage., 5共1兲, 1–7.

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MARCH 2005 / 367

J. Constr. Eng. Manage., 2005, 131(3): 360-367