The Economic Payback Technology Assessment Group

of 3D Mice for CAD Design Engineers 2008

Technology
Assessment
Group



The Economic Payback of 3D Mice
for CAD Design Engineers
Research Findings
















Abstract

Technology Assessment Group (TAG), an independent product consulting
firm specializing in product evaluation and productivity measurement,
conducted this research to assess the economic impact of 3D mouse use by
CAD design engineers.

User interface research by GE, IBM, and the University of Toronto suggests
that substantial productivity gains should result from using well-integrated
6-degree-of-freedom (6DoF) devices for complex 3D applications such as
3D CAD.

This resulting report incorporates market data and independent research to
provide a framework in which companies can estimate their economic
results.

Key Findings
- More than 84% of CAD design engineers report a noticeable or
significant improvement in their product designs and their ability to
detect design problems as a result of using 3D mice.
- The average productivity gain reported by CAD users while using 3D
mice is 21%.
- The payback period for 3D mice is very short, typically less than one
month.

The Economic Payback - 1 - Technology Assessment Group
of 3D Mice for CAD Design Engineers 2008
1. INTRODUCTION
Delivering high-quality, defect-free products to the
marketplace faster than the competition is central
to any companys success. Both factorsquality
and time to marketare critical. Companies can
quickly riseand fallbased on their perform-
ance.
Examples of this abound in the business news. For
instance:
Automobile companies are racing to deliver
next-generation fuel-efficient cars in response
to changed customer economy requirements
and government emissions regulations.
- Reuters reports that as the race to bring a
mass-market, rechargeable electric
vehicle to the market heats up, GM
executives have said the Volt is crucial to
the largest U.S. automakers efforts to
snag the environmental technology crown
from Japanese rival Toyota Motor Corp.
Cell phone companies are scrambling to
deliver new offerings to lure customers.
- Motorola, the category leader in 2006
with its hot Razr product, failed to deliver
compelling encores and has slipped to
third place in 2008.
Airplane manufacturers are pushing to deliver
new airplanes that will constitute a substantial
percentage of their future revenues. Getting to
market a few months faster than the compe-
tition can make the difference between
winning or losing billion-dollar orders.
In the product development chain, one key element
to delivering high-quality, defect-free products
quickly to the market is the performance of CAD
design engineers. If they can improve their
product designs, catch problem areas earlier, and
do all this in less time, they can contribute to
improving their companies market performance.
Fundamental user interface research by GE
Research, IBM, the University of Toronto, and
others has documented the performance
improvements resulting from user interface devices
that enable the CAD design engineer to navigate
3D objects intuitively and to work with both hands
simultaneously.
3D mice are user interface devices that provide
both intuitive navigation of 3D models and the
ability to work with two hands simultaneously.
CAD design engineers and companies who have
adopted 3D mice for their product design work
have reported impressive performance gains.
But no careful quantitative research has been done
to determine just how much difference these 3D
mice make. And because 3D mice represent a
company investment, its important to understand
the economic results, which companies can use to
assess the appropriateness for their organization.
Technology Assessment Group (TAG) designed
the following research to help answer these
questions:
A 14-question survey was created to collect
responses from 190 existing 3D mice users.
This survey was fielded by MarketLab, an
independent market research group, in May
2008. The survey asked users about their
experience with 3D mice with regard to:
- Perceived improvements in product
design and early defect detection
- Productivity gains (how much faster they
were in performing their work)
- Length of time it took them to become
comfortable and productive with 3D mice
- Amount of time they spent using their 3D
CAD applications
This report presents the results from this research,
as well as the underlying user interface research
that explains the reasons for the results.
The report then addresses these fundamental
management questions:
What is the economic payback of investing in
3D mice for CAD design engineers?
How can we determine the economic payback
for our company?

The Economic Payback - 2 - Technology Assessment Group
of 3D Mice for CAD Design Engineers 2008
2. USERFINDINGS
One hundred and ninety CAD design engineers
who use 3Dconnexion 3D mice were surveyed in
the U.S. They worked in companies with fewer
than 10 CAD design engineers up to companies
with more than 500 CAD design engineers.
These design engineers most commonly used
familiar 3D CAD applications such as CATIA,
Inventor, NX, Pro/ENGINEER, and SolidWorks.
They represent the full range of 3D mice
experience, from less than three months to more
than two years. Of these design engineers, 53%
used their 3D mouse for less than one year, and
88% used it for less than two years, with the
breakout as shown below.


Note that for the sake of brevity in this report,
percentages are presented with no decimal part. As
a result, presented percentages will sometimes vary
1% due to rounding.

2.1 JobCharacteristics
CAD design engineers are different from casual
computer users in that they use job-specific CAD
applications many hours a day to perform their
work functions.
Accordingly, 74% reported that they spend at least
three hours a day using their CAD applications.
Fully 41% spend at least seven hours a day. The
following diagrams show the distribution of usage
by group and cumulatively.




2.2 CADApplicationsand3DMice
As stated earlier, corporate and academic research
has shown that two key 3D mouse factors signifi-
cantly improve the performance of people using
intensive 3D applications:
6DoF devices for quickly orienting 3D objects
or views
Devices that enable working with both hands
simultaneously (for example, a 3D mouse in
one hand and a traditional 2D mouse in the
other hand)
The survey wanted to determine whether 3D
mouse users experienced these two factors in their
work and whether they thought these factors
enabled them to produce higher-quality designs,
detect errors better, and create designs faster.
Of these users, 83% reported (on a five-point
scale) that the 3D mouses 6DoF navigation was
very useful or extremely useful, and nearly
half (49%) found it extremely useful. Virtually
all users (95%) found this feature useful or
better. The detailed response percentages are
shown below.

The Economic Payback - 3 - Technology Assessment Group
of 3D Mice for CAD Design Engineers 2008

Concerning working simultaneously with both
hands, 75% found the 3D mouses enabling of
two-handedness very useful or extremely
useful, and again, nearly half (49%) found it
extremely useful. Virtually all (93%) found this
feature useful or better. The detailed response
percentages are shown below.

How then did these factors affect the product
design process? The Introduction described high-
quality, defect-free products as being key to a
companys success; can 3D mice actually improve
design quality and reduce errors?
According to the surveyed users, a 3D mouse
enabled them to much more easily rotate, inspect,
and explore their designs. As a result:
85% saw a noticeable or significant
improvement in their product designs
84% thought that they could noticeably or
significantly improve their detection of
errors



These are very high percentages, indicating that
companies adopting 3D mice for their CAD design
engineers should confidently expect similar results.
And what about design speedthe time it takes
design engineers to create their design? Are they
faster (more productive) using a 3D mouse?
Improving CAD designer productivity will directly
contribute to faster time to market, which can have
an enormous impact on a products success in the
marketplace.
CAD designers reported that they were, on
average, 21% more productive using 3D mouse
than they were without a 3D mouse. More than
86% of the users reported productivity increases,
ranging from under 10% to over 50%. The
following chart details the responses.

What about the learning curve for using 3D mice?
If it takes three months to become comfortable
with a 3D mouse and another three months to
become productive, are these productivity gains
worth the learning curve?
In order for users to embrace a new way of work-
ing, its critical that they can quickly become
comfortable with the new style. If they find the
new approach awkward or cumbersome, theyll
abandon it, even if it might pay dividends down-
stream.

The Economic Payback - 4 - Technology Assessment Group
of 3D Mice for CAD Design Engineers 2008
With 3D mice, more than half the users (58%)
were comfortable within the first four hours, and
the vast majority (80%) were comfortable within
two days.

Next, how long does it take for users of 3D mice
to feel not only comfortable but proficient?
According to the survey, 3D mouse users move
quickly from feeling comfortable to feeling
proficient: 66% felt proficient within the first
week, and 78% within two weeks.

How quickly does a 3D mouse user become more
productive? This is the ultimate goal of any
changed work style.
Users reported that nearly half (45%) were more
productive within two days, and 68% were more
productive within the first week of using a 3D
mouse.

3. UNDERLYINGUSERINTERFACERESEARCH
It is important to understand the fundamental user
interface concepts that underlie these productivity
improvements. This provides an understanding
both for CAD design engineers who experience
these improvements as well as non-CAD
professionals who might wonder why 3D mice
would make such a difference.
This section first explains how a CAD design
engineers computer use varies from casual
computer user. It then addresses the unique user
interface demands presented by 3D CAD
applications. The user interface bandwidth concept
is introduced along with two major UI bandwidth
accelerators.
References for the research cited in this section can
be found in the References section at the end of
this report.

3.1 CADDesignEngineersvs.CasualComputer
Users
CAD design engineers commonly:
Work at a core job function that depends
heavily on job-specific, complex CAD
applications
- The most frequently used 3D CAD
applications are CATIA, Inventor, NX,
Pro/ENGINEER, and SolidWorks.
Often spend more than half of their day using
their CAD applications
Require very high-performance computers in
order to increase job productivity
Spend between 1000 and 50,000 on
application software
More than one million 3D CAD users worldwide
share this profile.
In contrast, casual computer users:
Work at a core job function that may involve
using general-purpose applications (e-mail,
Web access, word processing, spreadsheet,
and so on) but that typically does not depend
on job-specific applications
Spend, on average, less than half of their day
on a computer
Have less need for high-performance com-
puters
Spend less than 1000 on application software

The Economic Payback - 5 - Technology Assessment Group
of 3D Mice for CAD Design Engineers 2008

The table below summarizes the core differences
between these two classes of computer users.
3D CAD
User
Casual
User
Applications
Complex,
jobspecific
General-
purpose
Computer
Use
48 hours/day 04 hours/day
Computer
Performance
High
performance
Medium
performance
Application
Purchases
1000
50,000
< 1000

These differences provide a context for examining
the characteristics of 3D CAD applications and
their unique user interface challenges.

3.2 Characteristicsof3DCADApplications
3D CAD users have substantially more demanding
computer working styles than casual users. Their
job-specific applications typically require them to
work in the following unique ways:
More frequent navigation of the work
(models, views)
More complex (degrees-of-freedom)
navigation (panning, zooming and rotating
much more common)
Dramatically more commands/minute and
navigations/minute than a casual computer
user
Much greater number of frequently used
commands
To illustrate, imagine a casual user reading e-mail,
the most frequently used application. The user
would start reading an e-mail message and perhaps
scroll down vertically to finish reading it. Then
they might reply or forward the message, and
then select the next email to read. In this typical
scenario:
The navigation (vertical scrolling) is typically
limited to one degree-of-freedom (1DoF), as
is the selection of the next e-mail message to
read.
The number of commands actually used is
fairly limited.
The user input bandwidth requirement is
quite low, for both navigation and commands.
If you were to watch this users hands from above,
the pace would be measured and slow. In contrast,
the 3D CAD users hands appear like those of a
concert pianist racing through a fast passage, the
right hand rapidly moving the mouse and mouse
wheel while the left hand repeatedly selects keys
(often Ctrl, Shift, Alt, and Esc) on the keyboard.
Based on observation of and interviews with 3D
CAD users, TAG estimates that 3D CAD users
issue 5 to 10 times more navigations/minute and
commands/minute than casual users. This demand
to push a large number of navigations and
commands per minute is the core requirement of
high-bandwidth user interfaces, as discussed in the
next section.

3.3 UserInterfaceBandwidth
3D CAD applications performance can be
throttled by three distinct bandwidth channels:
Computer bandwidth
Graphics bandwidth
User Interface bandwidth
To illustrate, lets take the example of a
mechanical engineer designing a new faucet using
3D CAD software such as Pro/ENGINEER or
SolidWorks.
The computation bottleneck is the ability of
the software/computer to keep a 3D model up-
to-date. As products become more complex,
the computation requirements increase
rapidly.
The display bottleneck is the ability of the
software/graphics card to render the 3D model
accurately in real time.
The user interface bottleneck is the ability of
the user to directly move the object to the
desired position and then issue various com-
mands, with the least number of interruptions
and context shifts, in the shortest amount of
time.
Whereas computer bandwidth and graphics band-
width have increased at a Moores Law pace, 3D
CAD user interfaces have not kept up. As a result,
user interface bandwidth has emerged as one of the

The Economic Payback - 6 - Technology Assessment Group
of 3D Mice for CAD Design Engineers 2008
principal bandwidth throttles for 3D CAD
applications today.
A conceptual framework developed by academic
researchers provides a useful visual representation
for understanding user interface bandwidth.
Navigation Command
Time
Navigation Command Command
Switch
Time
Switch
Time
Switch
Time
Left
Hand
Right
Hand

User Interface Bandwidth Framework
(Source: Buxton, W., Billinghurst, M., Guiard, Y., Sellen, A., and
Zhai, S. 2002)
This framework illustrates that user interfaces
(today and in the near future) are driven by the
activity of the right and left hands, generating both
navigation and commands. User interface
bandwidth is simply the time it takes to execute a
series of navigations and commands to perform a
particular application function.
3.4 InputStreams
The first user interface bandwidth limitation in 3D
CAD applications has to do with input streams.
As just discussed, all user input is driven through
the right and left hands; in reality, however, the left
hand is typically doing very little except periodic-
ally invoking a mode key press (for example,
Ctrl, Shift, or Alt). From the user interface band-
width framework shown below, we see that the
right hand (assuming a right-handed user) is doing
almost all the work, essentially constituting a
single input stream.
Navigation Command
Time
Navigation Command Command
Switch
Time
Switch
Time
Switch
Time
Left
Hand
Right
Hand

Single-Stream User Input
As summarized aptly in Zhai, Smith, and Selker
(1997):
One basic feature of the existing mainstream
user interfaces is that the user communicates
with the computer system via a single
stream of spatial input, physically driven by
a 2 degree of freedom input device, typically
a mouse, and graphically displayed as a
cursor. The universal cursor travels around
the entire interface, switching its functions
from pointing, to selection, to drawing, to
scrolling, to opening and to jumping,
according to what virtual devices (widgets),
such as the main document/window, a menu,
a scrolling bar, an icon or a hyperlink, has
been acquired and engaged. Such a single
stream operation, needless to say, has
offered the users many advantages such as
the ease of understanding and learning the
interaction mechanism. The disadvantage,
however, is the limited communication
bandwidth (Buxton 1986) and the costs in
time and cognitive effort of acquiring
widgets and control points (Buxton and
Myers 1986, Leganchuk, Zhai and Buxton
1996).
In observing both 3D CAD and casual computer
users, TAG estimates that the 3D CAD user issues
5 to 10 times more navigations/minute and
commands/minute than the casual user. When
these have to proceed largely through a single
stream (albeit with some use of the keyboard for
buttons or modifiers), the bandwidth is severely
restricted.
The first opportunity for improving user interface
bandwidth is thus to increase the number of
streams through which the user can drive the
application.

3.5 Navigation
The second user interface bandwidth limitation is
navigation. Navigation involves getting to the
place of interest to perform a task. This could be
scrolling to read an e-mail message, panning to a
location in Photoshop, or rotating a model to view
the back side of a part in CATIA.
Although navigation is a frequent activity in most
applications, the nature of the navigation varies
dramatically depending on the application type.
The following table provides a description of
common navigation operations, together with the
number of degrees of freedom (DoF) they require
and some example applications.

The Economic Payback - 7 - Technology Assessment Group
of 3D Mice for CAD Design Engineers 2008


#
DoF
Description
Common
Applications
Scrolling
(Vertical)
1
Moving a
document
up/down
E-mail, Web,
Word
Scrolling
(Horizontal)
1
Moving a
document
left/right
Excel
Panning 2
Moving a
drawing
simultaneously
horizontally and
vertically
AutoCAD,
Photoshop
Zooming 1
Moving a
document/model
in or out
AutoCAD,
Photoshop
Rotating 3
Moving a model
simultaneously
around any of
three rotational
axes
3ds Max,
CATIA,
Pro/ENGINEER,
Maya,
SolidWorks
DoF Requirements for Different Types of Navigation

These DoF numbers are additive. For example, to
pan and zoom you need 2 (pan) + 1 (zoom) =
3DoF. To pan, zoom, and rotate you need 2 (pan) +
1 (zoom) + 3 (rotate around three axes) = 6DoF.
Different applications vary dramatically in their
use of these various types of navigations, as shown
in the following table.
A
p
p
l
i
c
a
t
i
o
n

S
c
r
o
l
l
i
n
g

(
V
e
r
t
i
c
a
l
)

S
c
r
o
l
l
i
n
g

(
H
o
r
i
z
o
n
t
a
l
)

P
a
n
n
i
n
g

Z
o
o
m
i
n
g

R
o
t
a
t
i
n
g

E-mail
*****
Word
***** * *
Excel
**** *** **
Photoshop
* * *** ****
CATIA and
3D CAD
applications
*** ***** *****
Navigation Frequency by Application
(* = low; ***** = high)
The salient fact is that most 3D CAD applications
frequently navigate using pan and zoom (3DoF) or
pan, zoom, and rotate (6DoF). Accordingly, this
presents another important user interface band-
width opportunity.


BeingintheFlow
Before turning to research regarding high-
bandwidth opportunities, its worth noting that the
three bandwidth limitations break an inherently
creative process called being in the flow.
Being in the flow is a term used by artists,
athletes and designers to describe activities where
they are fully engaged and in control. Another
phrase used to describe this state is being in the
zone. All of these activities involve substantial
concentration and outlay of mental and/or physical
energy.
For 3D CAD computer users working with
complex and cognitively demanding applications,
being in the flow translates to higher quality and
faster performance. Often, however, theyre
distracted from their flow by user interfaces that
siphon off cognitive bandwidth and require the
user to slow down in order to drive tedious
aspects of the user interface (Bederson 2002).
Significantly, one of the most common interrup-
tions to being in the flow is a low-bandwidth user
interface in which users cannot engage in their
tasks as quickly as they can think.
In contrast, high-bandwidth user interfaces allow
3D CAD users to stay in the flow; well turn now
to these bandwidth opportunities.

3.6 HighBandwidthUserInterfaceOpportunities
In the previous section, two significant user
interface throttles were identified:
Limited input streams
Limited navigation
For both of these throttles, research provides
approaches that can significantly increase the
bandwidth.

HigherBandwidthInputStreams
We introduced the problem of the single stream of
input when we observed that 3D CAD users are
trying to push 5 to 10 times more commands per
minute than a casual user. Whereas a casual user
might not be as greatly affected by having a single
stream, the 3D CAD user has much higher band-
width requirements.

The Economic Payback - 8 - Technology Assessment Group
of 3D Mice for CAD Design Engineers 2008
One very promising user interface approach takes
advantage of humans ability to use both hands
simultaneously in a cooperative fashion. As noted
in Buxton (2002):
A student turns a page of a book while
taking notes. A driver changes gears while
steering a car. A recording engineer fades
out the drums while bringing in the strings.
By equipping both hands with tools to drive the
application (typically a 3D mouse in the left hand
and a standard 2D mouse in the right hand),
substantial bandwidth increases can be achieved.
First, lets look again at how single-stream
interfaces work today.
Navigation Command
Time
Navigation Command Command
Switch
Time
Switch
Time
Switch
Time
Left
Hand
Right
Hand

Single-Stream User Input
Note that the user incurs a switching time by going
from one mode to another. The near-universal
example of this is navigation and selection. The
right hand first navigates to the point of interest,
say on a model, using the mouse. Then the user
switches modes, whereby the mouse now
becomes a selection tool to issue a command. This
process repeats itself endlessly.
Also observe the lack of parallelism: the user is
either navigating or selecting at one time, but not
both.
A bimanual stream would change the activity
profile as illustrated below.
Navigation
Command
Time
Navigation
Command Command
Left
Hand
Right
Hand

Bi-manual Input Streams
Because each hand has a tool to perform tasks, the
user doesnt need to switch the right hand from a
navigation mode to command mode and back
again. Removing the unnecessary switches
essentially reduces the bandwidth requirement.
In addition, the human physiology allows for
parallel activities that can be synchronized with
each other, providing additional bandwidth
headroom. This parallelism is depicted in the
preceding illustration by the partial overlap of
navigation and commands: the user can start the
command with the right hand while the left hand is
completing the navigation.
The resulting comparison of unimanual and
bimanual performance is shown below.
Navigation
Command
Time
Navigation
Command Command
Left
Hand
Right
Hand
Navigation Command
Time
Navigation Command Command
Switch
Time
Switch
Time
Switch
Time
Left
Hand
Right
Hand
Bandwidth
Increase

Unimanual (top) vs. Bimanual (bottom) Bandwidth
The conceptual framework illustrated above was
validated in a study conducted by IBM (Zhai
1997), in which they found that a bimanual inter-
face (in this case, a joystick in the nondominant
hand and a mouse in the dominant hand) was 1.36
times faster than using the mouse alone, in tasks
involving navigation and selection.
Bi-Manual vs. Uni-Manual Performance
72
69
53
51
Test 1
Test 2
One Handed
Two Handed

Unimanual vs. Bimanual Performance
(Source: IBMZhai 1997)
Furthermore, in a study conducted at the Univer-
sity of Toronto (1997), as the tasks became more
cognitively demanding (larger, more complex
models) two-handed interfaces produced an even
more significant performance gain than the Zhai
research.


The Economic Payback - 9 - Technology Assessment Group
of 3D Mice for CAD Design Engineers 2008
HigherBandwidthNavigation
As described earlier, navigation in 3D CAD
applications as compared to more traditional 2D
applications is much more frequent and requires
more DoFs for efficient performance.
The following diagram shows how many simul-
taneous DoFs are required by various types of
navigation, from no rotation (scrolling) to pan and
zoom and finally to pan, zoom, and rotate.
P
a
n

&

Z
o
o
m
High Low
High
Low
Rotation
6DoF
1DoF
3DoF

Application Navigation and DoF
3D CAD applications typically fall squarely in the
6DoF quadrant, as shown in the following
diagram.
P
a
n

&

Z
o
o
m
High Low
High
Low
Rotation
3D CAD
Animation
3D GIS
Graphics Arts
EDA
2D CAD
Email
Web
WP
Spreadsheets

Navigation by Application Type
This introduces the potential for devices that offer
more simultaneous DoFsup to 6DoF to address
applications with high zoom and pan and high
rotation, which are typically 3D applications.
The following table lists common input devices
and their characteristics, notably the number of
simultaneous DoFs.
D
e
v
i
c
e

T
y
p
e

S
i
m
u
l
t
a
n
e
o
u
s

D
o
F
s

R
a
t
e

o
r

P
o
s
i
t
i
o
n
a
l

E
x
a
m
p
l
e

Two-button
mouse
2 Positional
Classic
mouse
Wheel
mouse
2+1 Positional
Microsoft
IntelliMouse
Graphics
tablet
2+1+1+1 Positional
Wacom
Intuos
Joystick 2+1 Rate
Logitech
Wingman
3D motion
controller
6 Rate
3Dconnexion
SpaceBall
Device Types and Characteristics
The conventional mouse offers 2DoF, being able to
move along the plane of a desk. The mouse wheel
separately offers 1DoF (typically for scrolling in
text-based applications, and for zooming in 3D
applications). Users typically do not move the
mouse and spin the wheel at the same time, so a
wheel mouse can be described as a 2+1DoF
device.
A 6DoF device allows the user to move in one
fluid movement to zoom, pan, and rotate the object
to any orientation.
In contrast, the wheel mouses 2+1DoF intrinsic
capabilities require a modal DoF mapping to
achieve 6DoF navigation, typically involving
pressing an additional key. A common approach is
as follows:
Mode A (Ctrl key depressed) + mouse
movement pans the model
Mode B (Alt key depressed) + mouse
movement rotates the model
Mode C (no keys depressed) + mouse wheel
zooms the model
Using the UI bandwidth framework, the following
comparison shows the increased bandwidth
resulting from using a 6DoF device for 3D
navigation.

The Economic Payback - 10 - Technology Assessment Group
of 3D Mice for CAD Design Engineers 2008
6DoF Navigation
Time
Left
Hand
Right
Hand
Pan
Time
Left
Hand
Right
Hand
Bandwidth
Increase
Zoom Rotate Pan
Mode
Switch
Time
Mode
Switch
Time
Reset
Mouse

Mouse (top) vs. 6DoF device (bottom) 3D Navigation
One of the most common activities in 3D CAD
applications is the frequent precise movement of a
model from one orientation to another. In a GE
study of seven users (Salazar and Marteau, 2004),
users had to move from one of eight possible
starting points to reach a precise (+/1) target 3D
orientation, using a classic mouse and a 6DoF
device.
In the GE study, users could achieve the target 3D
orientation almost twice as fast with a 6DoF device
(in this case, a 3Dconnnexion 3D mouse) as
compared to a standard 2D mouse, as noted in
following graph.
3D Navigation Performance
12.0
22.5
18.0
14.0
12.5
13.0
12.0
6.0
10.0
9.5
7.0
8.0
8.0
6.5
#1
#2
#3
#4
#5
#6
#7
Subject
Mouse
6DoF Device

3D Navigation Performance: Standard Mouse vs. 6DoF Device
(Source: Salazar and Marteau 2004)
When using the standard mouse, users took 89%
longer to perform the required 3D orientation.
Moreover, all users were substantially faster when
using the 6DoF device, ranging from 1.56 to 2.25
times faster, suggesting that the results would
apply broadly to all users.

Ratevs.PositionalDevicesforNavigation
Another point worth noting is the distinction
between rate devices and positional devices, and
their respective strengths for navigation. The
earlier table showing device types and character-
istics indicates which devices are rate vs.
positional. According to Zhai (1997):
As shown in recent six degree of freedom
input control studies (Zhai and Milgram
1993, Zhai, Milgram and Drascic 1993, Zhai
1995), position control is better conducted
with isotonic, free moving devices, such as
the mouse; and rate control is better
conducted with isometric or elastic devices.
The key factor to this compatibility issue is
the self-centering effect in isometric or
elastic devices. With self centering, rate
control can be easily done. Without it, rate
control requires conscious effort. Either
position control or rate control can give
users the ability to control all aspects of
movement, including displacement,
movement speed or higher order derivatives,
but each mode corresponds to only one
aspect directly: displacement or speed.
A rate control technique that is compatible
with isometric devices can be particularly
suitable for navigation tasks where you need
very precise movement and also very large
movements (e.g. scrolling long documents,
rotating a model, moving a camera) as no
repetitive release-reengage problem exists as
in the case of a mouse.

3.7 UserInterfaceResearchConclusions
3D CAD computer users require much a higher
user interface bandwidth in order to stay in the
flow of their work and perform at their optimum
level.
3D CAD users issue 5 to 10 times more naviga-
tions/minute and commands/minute than casual
users. 6DoF navigations are common, further
taxing user interface bandwidth. These points,
coupled with the high percentage of time that 3D
CAD users spend using their CAD applications,
present significant opportunities for improving
productivity by increasing user interface
bandwidth.

The Economic Payback - 11 - Technology Assessment Group
of 3D Mice for CAD Design Engineers 2008
Two user interface approaches present substantial
potential for improving productivity:
Bimanual interfaces, using a mouse in the
dominant hand and a rate device in the
nondominant hand (1.36 times fasterIBM
research)
A 6DoF device for the nondominant hand,
particularly in 3D applications (1.89 times
fasterGE research)
Moreover, these approaches should have an
additive impact, further increasing the user
interface bandwidth for 3D CAD users.
The survey of 3D mice CAD users and the time-
measured test designed by a senior CATIA
application engineer indicate that significant
productivity gains can be realized by 3D CAD
design engineers. Fundamental user interface
research further explains the reasons for such
gains.
The productivity increases reported by CAD
design engineers and the productivity time
measurements of CATIA users are concrete
manifestations of this underlying research.
Given these impressive productivity increases, its
now time to address the larger economic question:
what is the economic payback of equipping CAD
design engineers with 3D mice?

4. ECONOMICPAYBACKOF3DMICE
Its difficult to precisely quantify the impacts of
higher product quality, fewer defects, and faster
time to market. But with the research results
presented here, the economic return from a design
engineers productivity gains can be calculated.
Its paramount, however, to recognize that product
quality, fewer defects, and faster time to market
represent a much larger financial impact than
simply the cost savings from having a more
productive CAD design engineer.
As Gavin Finn writes in Quality Digest:
Very real costs are associated with inattention
to design quality. If errors or omissions in the
design data are not addressed early, more
costly changes are required later in the product
development process.
This is depicted in Finns early detection
diagram, below.

Thus, if an economic return can be demonstrated
on the design engineers productivity gains alone,
its reasonable to assume a much higher payback
overall.
Three principal factors will drive the ROI of
investing in 3D mice for CAD design engineers:
Cost of the 3D mouse
Loaded salary of the CAD design engineer
Productivity gains as a result of 3D mouse use
Companies use two common metrics to evaluate
such investments: payback period and annual ROI.
Further metrics (NPV, IRR, and so on) will not be
discussed in this report but could be easily derived
from this data.

4.1 PaybackPeriodandROI
The payback period determines how quickly the
investment cost will be fully recovered. The
calculation is as follows:
PaybackPeriodinYears=
3DMouseCost/(AnnualCADDesignEngineer
LoadedSalary*ProductivityGain)
As shown in the following illustration, this
calculation can be depicted visually in a payback
calculator, in which the user can adjust the three
sliders:
CAD Design Engineer Loaded Salary
3D Mouse Cost
Productivity Gain

The Economic Payback - 12 - Technology Assessment Group
of 3D Mice for CAD Design Engineers 2008
The payback calculator would then compute and
display the resulting payback period in months.

The ROI calculation measures the ongoing return
on an investmenttypically on an annualized
basis, which gives a more comprehensive financial
evaluation. This calculation is:
Annual ROI =
(Annual CAD Design Engineer Loaded-Salary
*Productivity Gain) 3D Mouse Cost ) /
3D Mouse Cost
Two of these variables are reasonably straight-
forward: the 3D mouse cost and the CAD design
engineers loaded salary. The critical variable
productivity gainis derived from the survey of
3D mouse users.
These constitute the inputs into determining
expected economic returns on investment in 3D
mice for CAD design engineers.

4.2 3DMiceCosts
3Dconnexions professional 3D mice range from
69 to 275 in price. Many companies select the
higher-end professional devices, SpaceExplorer
(199) or SpacePilot (275), due to their richer
feature set. For the purpose of this analysis, well
use the 275 cost of the SpacePilot.

4.3 CADDesignEngineerSalariesandCosts
Several websites summarize salaries for various
job titles. ITJobsWatch reports the average salary
for a CAD design engineer in 2008 as 37,500.

This will of course vary by all the usual factors,
including years of experience, location, and
industry. In general, 3D CAD design engineers
will make more than 2D CAD design engineers.
Employee benefits (vacation, health insurance, and
so on) are estimated conservatively at 25% of base
salary, resulting in an average benefit-loaded cost
of 46,875 per CAD design engineer.
Fully loaded costs (space, equipment, and so on)
add another substantial cost multiple. In the
absence of solid data, this factor will be ignored in
the analysis.

4.4 3DMiceProductivityGains
The productivity gains from using 3D mice are
calculated by taking the average productivity gain
reported in the survey and multiplying it by the
average percentage of the day that design
engineers spend using their 3D CAD applications.
The average productivity gain reported by the 190
3D mouse users was 21%. The average time per
day users reported that they used their CAD appli-
cations was five hours; a conservative estimate of
50% of their day will be used.
Multiplying these two figures together, we get an
average productivity gain of 10.5%.
Now, using the earlier payback period formula of
PaybackPeriodinYears=
3DMouseCost/(AnnualCADDesignEngineer
LoadedSalary*ProductivityGain)

we get the following calculation:

275/(46,875*10.5%)=.056years(20days)
This means that an investment in a 3D mouse will,
on average, pay for itself in less than one month.

The Economic Payback - 13 - Technology Assessment Group
of 3D Mice for CAD Design Engineers 2008
By adjusting the three payback calculator sliders to
these figures, we see the resulting 20-day (= 0.65
month) calculation.


5. CONCLUSIONS
This report purported to evaluate the anecdotal
claims that 3D mice can significantly improve
CAD design engineer productivity. It further
sought to evaluate the user interface research that
also suggested impressive productivity gains.
Based on a survey of 190 3D mice users, it appears
that in fact substantial gains of more than 20% are
being experienced by CAD design engineers while
using 3D mice with their CAD applications.
These users further corroborated the underlying
user interface research observations that 6DoF
navigation and simultaneous two-handedness were
the key factors leading to their improvements.
Finally, it was shown that an investment in 3D
mice can have an unusually fast paybackless
than a monthleading to the conclusion that
companies would be well advised to proactively
consider adopting 3D mice for their CAD design
engineers.

The Economic Payback - 14 - Technology Assessment Group
of 3D Mice for CAD Design Engineers 2008
6. REFERENCES
Bederson, B.B. (2002) Interfaces for Staying in the
Flow, Human-Computer Interaction Lab, University of
Maryland.
Buxton, W., Billinghurst, M., Guiard, Y., Sellen, A.,
and Zhai, S. (2002). Human Input to Computer Systems:
Theories, Techniques and Technology.
Buxton, W. (1986) Theres more to interaction than
meets the eye: some issues in manual input. User
Centered System Design, Lawrence Erlbaum Associ-
ates, Norman, D.A. and Draper, S.W. (Eds.), 3 19337.
Buxton, W. and Myers, B. (1986) A study of two-
handed input. Proceedings of CHI 86: ACM Conference
on Human Factors in Computing Systems, 321326.
Callahan, J., Hopkins, D., Wiser, M., and Shneiderman,
B. (1988) An Empirical Comparison of Pie vs. Linear
Menus, Computer Science Department, University of
Maryland.
Fitts, P. (1954) The information capacity of the human
motor system in controlling the amplitude of movement.
Journal of Experimental Psychology, 46, 199210.
Guiard, Y. (1987) Asymmetric division of labor in
human skilled bimanual action: The kinematic chain as
a model. Journal of Motor Behavior, 19(4) 486517.
ISUR Project: Industry Usability Report (1999) NIST
White Paper.
Kabbash, P., Buxton, W., and Sellen, A. (1994) Two-
handed input in a compound task. Proceedings of CHI
94: ACM Conference on Human Factors in Computing
Systems, 417423.
Leganchuk, A., Zhai, S., and Buxton, W. (1996) Manual
and cognitive factors in two-handed input: an experi-
mental study. Submitted for publication.
MacKenzie, I.S., Sellen, A., and Buxton, W. (1991) A
comparison of input devices in elemental pointing and
dragging tasks (1991). Proceedings of CHI 91: ACM
Conference on Human Factors in Computing Systems,
New Orleans, Louisiana, 161166.
Nielsen, J. (1994) Usability Engineering.
Poulton, E.C. (1974) Tracking skill and manual control.
New York, Academic Press.
Rutledge. J. and Selker, T. (1990) Force-to-motion
function for pointing. Proceedings of INTERACT 90:
The IFIP Conference on Human Computer Interaction,
701705.
Salazar, P. and Marteau, J-M. (2004) Designing a 3D
Input Device for Interventional Radiology GE Health-
care, Global Industrial Design Department.
Smith, D.C., Irby, C., Kimball, R., Verplank, W., and
Harslem, E. (1982) Designing the Star user interface.
Byte, 7(4), 242282.
Venolia, D. (1993) Facile 3D direct manipulation.
Proceedings of INTERCHI 93: ACM Conference on
Human Factors in Computing Systems, Amsterdam, The
Netherlands, 3 136.
Zhai, S. (1995) Human Performance in Six Degree of
Freedom Input Control, Ph.D. Thesis, University of
Toronto. http://etclab.mie.utoronto.edu/people/shumin
_dir/publications.html.
Zhai, S. and Milgram, P. (1993) Human performance in
evaluation of manipulation schemes in virtual
environments. Proceedings of VRAIS 93: IEEE Virtual
Reality Annual International Symposium, Seattle,
Washington, 155161.
Zhai, S., Milgram, P, and Drascic, D. (1993) An
evaluation of four 6 degree-of-freedom input tech-
niques. Adjunct Proceedings of INTERCHI 93: The
IFIP Conference on Human-Computer Interaction,
Amsterdam, The Netherlands, 155161.
Zhai, S., Smith, B., and Selker, T. (1997) Improving
Browsing Performance: A Study of Four Input Devices
for Scrolling and Pointing Tasks. Proceedings of
INTERACT 97.