Courtesy of:

Scitor Corporation
256 Gibraltar Drive • Sunnyvale, CA • 94089
800/533-9876

Simplifying Project Management Software Selection:

A Balanced Approach
Over the past fifteen years, I have written, taught, lectured and consulted on the topic of
Specifying, Evaluating and Selecting Project Management Software. This includes a
formal seminar, a book Project Management Using Microcomputers (Osborne/McGraw-
Hill-1986), a chapter on that topic in David I. Cleland's Field Guide to Project
Management (VNR-1998) and a video produced by IBM's Skill Dynamics.

Information Overload
In my enthusiasm to fully cover the topic, I more often than not got carried away with
details. This compulsion to cover every aspect of software selection and use may have
been technically sound, but, in retrospect, may have failed to have the desired effect. I
may have made my audience more knowledgeable and aware (an important objective),
but did I leave them in a better position to make a selection decision (the ultimate goal)?
Looking back, my approach was to use a work breakdown structure to organize the
details. There were about a dozen first level subjects, each with several sub items, making
for about 200 total characteristics and features to consider. I was shocked into
recognizing the error of my ways when a client in one of my seminars came to me at the
end and exclaimed "This is the first time that I attended a seminar where I left with more
questions than I had at the start."
This was not an isolated case, and I recognized that I was causing "information
overload". Actually I had some cognitive dissonance, as I wanted the audience to become
aware of all of the things that should be considered in selecting software, while also
trying to make the process easier. I had to correct the common belief that "all PM
software was essentially the same" while avoiding the possibility of scaring people away
from getting these tools. I fear that my extensive coverage of the topic was as
intimidating as it was helpful.

The Simplified, Balanced Approach

Now that I have fully berated myself in public, I want to offer a more simplified approach
to PM software selection. Recognizing the growing popularity of "The Balanced
Scorecard" philosophy, we will reduce the 200 items to just four major categories, and
look for comprehensive and balanced support for these four areas.
It is important that the software that is selected will meet the needs of the entire
community that is involved in projects, and that it provides complete and quality support
for the firm's project management process.

Written by Harvey Levine Page 1 of 8

 Courtesy of:
Scitor Corporation
256 Gibraltar Drive • Sunnyvale, CA • 94089
800/533-9876

The Four Key Categories

Subject to the possibility that we may be moving from an overly detailed approach to an
overly simplified approach, we will consider these four important areas:
! The User Interface
! Data Management
! The Scheduling Engine
! Multi-user Access & Communication
If we can satisfy all four categories, we are likely to have a product that we can use
effectively as part of a computer-based project management system. If any one of these
areas is unduly weak, we can expect failure of the entire system.

The Scheduling Engine

As our software systems get more and more sophisticated, and the user base gets more
widespread (and less computer-savy), the software selection process tends to focus more
on the User Interface characteristics. As a response, the software developers have placed
a special emphasis on making it easy to put the data in and to navigate about the tool.
Likewise, as the user base for the system output expands, the need grows for improved
data management and manipulation. Here too, the vendors have responded with razzle-
dazzle features that slice and dice the data from every conceivable angle.
Chances are that these two items (the user interface and data management) are the ones
that get most of everyone's attention. Yet it is in the scheduling engine where we see
significant differences, and where deficiencies can lead to ineffective computer-based
PM applications. So let's look at the scheduling engine criteria first.
Across the board, the available products keep on getting better in these first two areas and
the user community has benefited from these improvements. But let's think a bit about
why you are using project management software in the first place. Here's a list of typical
objectives:
! Store system calendars
! Store resource pool data
! Store project key dates and milestones
! Store the project workscope
! Store the project work breakdown structure
! Store time and work estimates, resource assignments and costs
! Calculate schedules
! Assist in assigning resources
! Resolve resource scheduling conflicts
! Calculate budgets
! Plot cash flow curves
! Collect and calculate applied time (timesheets)

Written by Harvey Levine Page 2 of 8

 Courtesy of:
Scitor Corporation
256 Gibraltar Drive • Sunnyvale, CA • 94089
800/533-9876

! Collect task status and recalculate schedules

! Collect actual costs
! Calculate earned value performance results
In addition, many users may wish to add the following, either in the basic product or in
add-ons:
! Identify risk areas
! Quantify risk items
! Calculate risk mitigation options
! Employ Critical Chain scheduling protocols
! Store typical plan templates and estimates
! Integration with ERP products and data
! Incident tracking
! Change control
Think about these functions. What kind of data do you need to input, calculate, or output
regarding these items? Can the tool that you are considering handle these data, in the
manner that you need for your business? Can you create discrete models of your plan in
the system? Or do you have to "jury-rig" them to make them fit the systems limitations?
Are the algorithms in the program sophisticated enough to calculate correct and efficient
schedules? Can you repeat the calculations and get the same result each time? Can you
preserve defined constraints? Can you define complex assignments?
When you attempt to resolve resource conflicts (via resource leveling) are you left with
gaps where resources are available to work but are not assigned to tasks that could be
worked on? Are the resource schedules consistent with the task schedules (I have
experienced products that could have a task scheduled for two days, but show the
resource as working on it for five days)? Are the earned value calculations correct? These
are not idle questions. These are all items where I have found deficiencies in some PM
software products in the past.
Chances are that you assumed that all products work about the same and will produce
similar results. My evaluation of the popular products show that this is not true.
While I chose to use the term scheduling engine for this key category, you can see that it
goes well beyond basic scheduling. On one hand, almost all of the popular products use
the traditional activity-on-the-node critical path scheduling approach, and the serial
algorithm for resource leveling. However, each vendor has designed its own
characteristics around these two basic calculation models, that may add or detract from
the basic capabilities.
In addition to evaluating for fidelity and consistency, for ability to handle your specific
needs, and for usability, you will also want to consider other performance characteristics,
such as speed of opening files, calculating schedules, leveling resources, saving files, and
manipulating data.

Written by Harvey Levine Page 3 of 8

 Courtesy of:
Scitor Corporation
256 Gibraltar Drive • Sunnyvale, CA • 94089
800/533-9876

And don't forget to examine all of these items from the point of view of multi-project
management. Working with multiple projects places extra demands of the system. These
may include the ability to handle large amounts of data, the ability to identify and
manipulate multiple project data, control of and access to the resource pool, and control
over replacing data in individual projects after working in multi-project mode.
All of the above may be overlooked when evaluating the user interface and data
manipulation characteristics of a product. Yet, they are at least as important to having a
successful system.

Data Management
This category is receiving greater attention every year, as well it should. Here are some of
the capabilities that can be supported by advanced data management functions:

DATA ACCESS AND TRANSFER

! As we move toward multi-project management and project portfolio management,
the data expands exponentially. We also may need to access data from multiple
data sources. For instance, additional, non-project data may be generated by
Enterprise Resource Management (ERP) systems. This data may have to be
coordinated with the project data.
! Hence, we are moving away from fully proprietary databases, toward popular,
open database systems. Whether utilizing ODBC conventions or storing in Oracle
or SQLServer, the process should be close to invisible to the user. The process
should also be fast and secure.
! Where data resides in several locations and in several databases, you may wish to
consider data warehousing systems.

SLICING AND DICING

! As the user base expands, different people need to access different information.
The general projects database will almost always hold way more data than any
one person will need at any one time. We look for a robust data manipulation
capability to allow the user to slice and dice the data as needed to support
individual knowledge requirements.
! In days past, we often needed an external repository for the data and had to learn
an arcane command language to make data queries. Today, it doesn't matter if the
data is internal or external. In the latter case, the link is usually transparent to the
user. Also, today's products provide a simplified query process, often menu and/or
icon driven. Knowing that these capabilities are now available, why settle for
anything less?
! In larger data systems, the process can be improved by using On-Line Analytical
Processing (OLAP) engines. OLAP systems arrange the data into pre-formatted
groups (or cubes) to speed up access.

Written by Harvey Levine Page 4 of 8

 Courtesy of:
Scitor Corporation
256 Gibraltar Drive • Sunnyvale, CA • 94089
800/533-9876

ADMINISTRATION & CUSTOMIZATION

! Individual users should not be required to set up the OLAP cubes or the slice and
dice queries. The software should come with an administrator capability, to
facilitate design of the data system, including retrieval and security.
! The administrator should be able to determine the information needs of each user
and tailor outputs, queries, spreadsheets and graphical presentations as
appropriate to meet these needs.
! Data must be able to be presented in user-defined time-phased segments, arranged
to system-defined hierarchies. Graphical presentation is very popular.

SUMMARIZATION
! Caution is advised when reviewing the software for summarization capabilities.
Most products provide a large variety of data fields, for text, dates, costs, etc.
Today, we are provided with almost unlimited user-defined fields. However, you
should be aware that sometimes the data in these fields can only be sorted or
filtered, but not summarized. Do not take it for granted that all data fields can be
summarized or rolled-up to a defined hierarchy. Yet, this capability is paramount
to advanced data manipulation and reporting.

User Interface
The first thing that anyone notices is the user interface. We all like the graphical user
interface style, such as supported by Windows and Macintosh. Here, again, there are
some things to look for and some things to avoid.
There are several ways that the user interface can facilitate access to the system
functions.
! The most popular mode is via graphical icons. I find these to be useful only if
there is text along with the icons or there are pop-up tool tips (text appears when
cursor is mover over icon).
! Drop-down menus is my preferred mode. This allows me to move through the
menus and learn what functions are available.
! Hot-key codes (usually combinations of Ctrl or Alt with F1 through F12) can
reduce key strokes are mouse movement, but I usually can't remember the codes.
! Several programs can only call up specified functions by clicking the mouse in a
particular place on the screen. The call may require a left-click, or a right-click, or
a double-click. Sometimes these will immediately initiate a function and other
times it will open a window that presents a selection of functions. This capability
is fine -- when available in addition to accessing these functions from the menu
bar. I find it to be unacceptable for such functions to be available only via these
"hot-spots" as there is no way to learn that they are there unless you already
know.

Written by Harvey Levine Page 5 of 8

 Courtesy of:
Scitor Corporation
256 Gibraltar Drive • Sunnyvale, CA • 94089
800/533-9876

Under this category, I also include those features that help me to input and view my data.
These would include:
! Flexible screen arrangement
! Drop-down or "pick" lists
! In-cell editing
! Custom, saved views
The User Interface can be considered to be part of a larger group, called Usability. To this
group, we can also add:
! Ease-of-learning
! Ease-of-use
! Macros
! Context-sensitive Help
! Tutorials

Multi-user Access & Communication

It is over forty years since critical path scheduling was developed and made available to
computers. Initial computer systems assumed that single projects would be managed by a
co-located group. Over the next twenty years, mainframe and minicomputer systems
changed as the developers recognized that project participants were scattered. The time-
sharing systems of the 70's allowed data input and analysis to take place in multiple
locations.
It is about twenty years since this technology was ported to microcomputers. Here too, at
first, the system design favored the individual user, on single projects. Remember -- we
called them "personal" computers. But it soon came apparent that the user base was in
multiple locations, working on multiple projects. Again, the industry responded.
However, they were hampered, somewhat, by technology that was really not designed to
optimize multiple, decentralized users.
The initial developments were centered around computer networks. The concept of
client/server systems was a key basis of emerging designs for multi-user project
management. These eventually developed into multi-tier client/server systems which
maximized the efficiency of the computer equipment while extending the user access and
security.
During the past few years, this entire world of multi-user computing was turned upside-
down by a tornado of sorts -- web-based computing. The web has revolutionized the way
that we use the computer and what we can accomplish with it. And if there is any
application area that can fully benefit from this new technology, it is project
management. The improved access and collaboration capabilities, enabled by web
technology, has totally changed the way that we use project management software, and
the things to consider when selecting such software.

Written by Harvey Levine Page 6 of 8

 Courtesy of:
Scitor Corporation
256 Gibraltar Drive • Sunnyvale, CA • 94089
800/533-9876

For many people involved in project management software selection, web support has
moved to the top of the list of key selection criteria. Again, I would advise caution. Web
support must be provided on top of solid scheduling capabilities. It should not be given
such a weigh factor so as to allow it to dictate the selection process and subrogate the
essential scheduling and calculation capabilities.
When we talk about web-based systems, there are several configurations. These include:
! Web-based services - A web-based application service, hosted by an external
vendor. The user pays a usage fee to access the service.
! In-house total web-based system. The user owns the system and places the
software on internal hardware.
! A hybrid system, utilizing client/server and web-based components.
! Web-based output (not really a web-based system, but rather web publishing)
Today, you can expect all modern project management systems to embrace the web to
some degree. However, many of the early entries into the totally web-based segment of
the market have sacrificed scheduling functionality, as compared to the more established
products. The exception to this would be companies that have already established strong
client/server systems that are redesigning the system to be fully web-based.
With every new release, web-based functionality increases. As this is being written, I am
learning of new web-based products and product updates with new web capabilities. If
web-based functionality is important to you, then you must specify these needs and
search for products that support that criteria. Today, just about every product
configuration is available. We had expected that, as the industry matured, several vendors
would fall by the wayside. Surprisingly, the opposite has occurred. We are seeing a
growth in vendors and products, and a greater variation in what they have to offer. For
you, as a software selector, the greater choice will let you find what you want, but the act
of choosing is more difficult.

Final Comments
As the user, you have to put all of the various benefits and deficiencies in perspective and
weigh all of the attributes against your specific environmental preferences and needs.
Another tradeoff area is track record vs. freshness. In general, the newer products are
using more of the newer and advance technologies. You have to weigh the attractiveness
of this against the benefits of a more established product, with proven performance. Also,
another caveat; often the newer technology can exhibit more razzle-dazzle than
substance. Check it out thoroughly.
Check out the vendors, also. Give them a call. Does someone answer the phone? Do they
return phone calls promptly? Are they knowledgeable about project management? Some
of the new products may sound exciting. But are these tested, shipping products, backed
up by a vendor with an office and qualified staff? Checking this out is part of the
software selection process.

Written by Harvey Levine Page 7 of 8

 Courtesy of:
Scitor Corporation
256 Gibraltar Drive • Sunnyvale, CA • 94089
800/533-9876

Implementing a Computer-based Project Management

Capability
One last thing. Selecting project management software is but a part of the process of
developing a computer-based project management capability. There are five key phases
of this process:
! Determine your project management methodology
! Select your project management tools
! Train for project management
! Establish an implementation plan
! Audit the implementation
I'm going to leave this for a follow-up article. Look for it soon.

Harvey A. Levine, with 38 years of service to the project management industry, is founder of The
Project Knowledge Group, a consulting firm specializing in PM training, PM software selection,
evaluation & implementation, and PM using microcomputers.

He has implemented or enhanced the project management capabilities of numerous firms, often
combined with the selection or implementation of computerized project management tools. Mr.
Levine is considered the leading consultant to the project management software industry and is
recognized as the leading expert in tools for project management.

He has been an Adjunct Professor of Project Management at Rensselaer Polytechnic Institute

and Boston University. And has conducted numerous project management public seminars for
ASCE, AMA, IBM, and PMI.

Mr. Levine is the author of the book "Project Management using Microcomputers", and has been
published extensively in other books, periodicals and videos.

Mr. Levine is a past president of the Project Management Institute and the recipient of PMI's 1989
Distinguished Contribution to Project Management award. Recently, he was recently elected as a
Fellow of PMI.

Mr. Levine has offices in Saratoga Springs, NY and San Diego, CA and can be contacted via e-
mail at: [email protected]

Written by Harvey Levine Page 8 of 8