Project Management Tools

Factors that make projects spin out of control-from application size and complexity to the number and skill level of developers-can put your organization at a competitive disadvantage. Here are four tools that can help you manage projects more effectively.

Remember three to five years ago, when all the strategic MIS plans had a prominent declaration? It usually was something like, "We will be in the forefront of efforts to deliver a strategic competitive advantage to the corporation." However, somewhere between lip and cup, Year 2000 problems, and client/ server birthing pangs, that strategic competitive advantage got lost in the shuffle.

But wait, not so, said Dr. Howard Rubin in his MIS Worldwide Benchmark Report (Rubin Systems Inc., 1994). His study stated that a small group of organizations are able to consistently outperform average MIS productivity (measured in function points produced per information technology professional per year) and average MIS reliability (measured in defect rates per information technology professional per year). And Rubin's database is not small-it spans thousands of organizations worldwide, across government and industry lines. The results are startling. In 1991, the productivity advantage of the best MIS shops was 18 to 1, vs. the average (of approximately 6,000 lines of code per information technology professional per year). By 1993, the ratio was 50 to 1, and by 1996 it was 200 to 1.

Let's be clear about what I'm saying. There are organizations whose MIS software productivity in 1996 is nearly 200 times better than average for all MIS shops (the reliability ratio is nearly 100 times better than the average). Clearly, a small group of organizations has been able to achieve a strategic competitive advantage through software development productivity. Imagine having to compete against one of these organizations. As computing becomes essential not just for accounting and manufacturing operations, but also for creative design, engineering, planning, and marketing one can ill afford to concede such a wide advantage. What are these top-tier MIS organizations doing right? Well, they aren't telling. A more illuminating question might be, "What are the other 90% not doing." Here, Capers Jones, in his book Applied Software Measurement, 2nd edition (McGraw-Hill, 1996) agrees with Rubin: the bulk of MIS organizations do not do process management, project management, and system measurement well. Nearly 80% concede they do not achieve the SEI's Capability Maturity Model (CMM) Level 2 status.

Among other things, organizations that have reached Level 2 status consistently apply project measurement and management tools to all their systems projects. Yet nearly 80% of MIS organizations say they don't do this. Is there something wrong, some missing functionality with project management tools?

It is with this question in mind that I shall examine four top project management tools: Scitor's Project Scheduler 7, Primavera's Project Planner (P3) 2.0, ABT's Project Workbench, and Microsoft Project. I am presuming that these products are all flawed in some way, because 80% of organizations do not use them (or any other project tool) on a consistent basis-despite the handsome returns achieved when project management and measurement tools are used.

Overhead Bane

The bane of project management tools is that they are perceived as overhead (that is, not contributing to the completion of work and even as a negative drain on productivity). Other cited flaws with project management tools follow:

• They are hard to learn or to remember how to use.

• They're expensive regarding time, effort, and initial cost.

• They cannot contribute to team and group management processes.

• They're too costly for effective deployment to all staff members.

• They're not well-integrated with other systems (time sheet, mail, fault reporting, personal information managers, and so forth).

• They're good for process, but fall down on content (what, when, and how much) in project planning.

Table 1 shows some of the key benefits project management tools should deliver to the various stakeholders in an MIS project. It also lists the costs and risks the tools should minimize to satisfy a broad user community.

Means of Communication

If you want to use project management tools to communicate project plans to all affected stakeholders, any of the four tools reviewed in this article provide excellent features for doing so. This is no small feat. Countless postmortems of failed projects have cited lack of overall vision or ongoing project status information as contributing to the project's demise.

First, all four project management tools provide four basic views into a project: the Gannt chart, which diagrams a breakdown of tasks over time, as shown in Figure 1 on the previous page; the PERT chart, which shows project dependencies in the form of a graph of predecessor and successor tasks, as shown in Figure 2; resource loading graphs; and spreadsheet views of tasks, resources, and assignments of resources to tasks.

In addition, all the project management tools provide flexible multiple document interface windowing capabilities, so project followers can customize their screens to get their own "perfect project performance" monitor. But be warned, all the tools thrive on a lot of screen space-a 1028 by 740 display is almost essential. All offer extensive zoom control (from 25% to 1,500% in at least six steps) for the PERT and Gannt charts, so you can easily move from the big picture to details of specific activities and tasks. They also offer extensive customization of these views including layout, font, and color coding; diagramming symbols; fields shown in nodes or graphs; and so on. So if your organization has developed specific PERT or Gannt chart diagramming standards, you can likely modify any of the tools to accommodate them.

Even more important, the four tools have several ease-of-use and navigation features that make moving among views and organizing data fairly simple tasks. Primavera's P3 and Scitor's Project Scheduler also export views and reports directly to the World Wide Web (that is, they publish project results as HTML pages). Thus, if users are comfortable with a browser, they can access project data. Again, all four tools make it simple to see the overall task layout in Gannt or PERT chart form while viewing all the specifics of any one task, as shown in Figure 3. This feature is vital in tracking down developing problems. Finally, Microsoft Project and Scitor's Project Scheduler provide users with extra views and customizations that allow a project leader added flexibility in setting up views, graphs, and reports for all possible clients and contingencies.

Overall, the project management tool vendors have done a superb job of providing a standard language for communicating project results. If I were a user investing in a high-risk or long-term software project, I would insist the project proposal include a Gannt chart showing major milestones and deliverables including resource usage and costs. Likewise, as an executive manager, my sense of control and insight into a software project under my stewardship would demand the ongoing use of a project management tool that is at least equal in caliber to the four reviewed here.

Ease of Setup and Ease of Use

From a benefits viewpoint, project management tools bring many important stakeholders on board, as shown in Table 1. However, if you're not comfortable or not willing to invest the time necessary to set up a project schedule, then the tools will be used grudgingly at best, or not at all. Likewise, for timely and up-to-date reporting, project management tools depend on results and data input provided by members of the project team. If those tasks are burdensome (for example, the same data is entered two or three ways in different systems) or clumsy (for example, people are not sure how to allocate time spent on various tasks), then project data may become tardy, incomplete, or inaccurate.

In setup time and costs, only one of the four systems, ABT's Project Workbench, provides a direct add-on program that can significantly reduce planning time and effort. Think of ABT's Project Bridge Modeler as a repository of project templates and an estimator for task durations and costs. A project leader can use a template as a starting point to model a particular project. Next, he or she adds or modifies tasks and, using estimator routines, calculates time and resource usage for the project. When satisfied with the results, you can use Project Bridge Modeler to generate all the milestones, phases, tasks, resources, and resource allocations in file form ready for use by Project Workbench. This meta-project or process modeling tool significantly reduces project setup time and makes project planning much more focused and accurate.

Fortunately, a number of excellent third-party process modeling tools are available: Ernst and Young's Navigator; LBMS's Process Engineer; and SHL Systemhouse's SHL Transform. Process Engineer has a library of dozens of project templates-many of which are provided by leading consultants in the areas of three-tier client/server development, Year 2000 projects, intranet development, and so forth. You can use these templates as the starting prototype for a specific MIS project plan. You can also use Process Engineer throughout the planning process, then output the final schedule and resources either directly to Primavera's P3 or in *.MPX format to the other tools. Thus, for savvy project leaders, initial setup time should not be a problem in using project management tools.

Just less than two years ago, one could say that even Windows GUI project management tools were hard to use. Not anymore. All four programs I reviewed provide carefully thought-out interfaces. All products had context-sensitive menus and toolbars. Specific objects have specific menus activated by right-clicks on the mouse. Views and toolbars are movable and customizable. Online data entry is a snap. Setting up calendars and resources is simple, and you can set inter-task dependencies in multiple ways. There are also scores of unique ease-of-use features.

ABT's Project Workbench provides extensive drag-and-drop operations. Microsoft Project and Scitor's Project Scheduler provide tips of the day and stay-on-top cue cards for guidance in stepping through processes that are a bit rusty in memory. Primavera's P3 provides wizards (for example, fill in the dialog questions and project reports are fully ready for World Wide Web usage). The bottom line is that they're all easy to navigate through. Where the going gets rough, help is available in a variety of formats.

However, there are some gaps when you integrate the tools with support programs or when you're dealing with large-scale projects.

Integration and Large-Scale Operation

Project management systems are highly codependent, especially for large projects. They derive input from a variety of sources (time sheets, job costing reports, purchase control, and so forth). Likewise, project management tools should send output directly to a number of systems such as performance measures databases, work reports, other project systems, and so on.

Only with the emergence of project management systems based on standard databases (Primavera's P3 uses Btrieve, Mantix's Cascade uses Oracle, Welcom's Open Plan uses .DBF file format, and so on) or the provision of ODBC retrieval (available with Microsoft Project and Scitor's Project Scheduler), has access to project data become more open to users. True, Primavera and ABT provide time sheet and measurement data storage programs that directly interface with their project management tools. ABT publishes an API to its data files' format, but the integration is just not there yet. So, data entered into a project management system often has to be reentered into other systems as well (or vice versa). Thus, project staff members have a legitimate concern about redundant data entry.

With OLE automation, Visual Basic for Applications and other scripting languages, plus ODBC interfaces and APIs, it will become easier to interface and integrate project management tools into the MIS mainstream. This is already reflected in a strong third-party market for all project management tools, especially Microsoft Project.

But large projects and underlying databases present two other problems: limits to project size and multiple projects. Except for Primavera's P3, these limits are around 30,000 tasks, 30,000 resource assignments, and 1,000 resources. These limits are due to either 16-bit code or all-in-memory operation. To compensate (or simply to match subprojects to independent groups), grouping of multiple subprojects has developed. The problem is with the underlying databases and the problems of concurrency control when a person or resource spans several subprojects. The result is some clumsy solutions such as whole file locking. Likewise, you can still perform the critical path calculations, resource levelings, and other what-if calculations, but they are subject to more complex interactions or refined constraints.

Until multiple project and large-scale wizards emerge, Primavera's P3 handles large, multi-user projects much more gracefully; but all four project management tools are more challenging to operate at these high-end limits.

Feedback Mechanisms and What-If Analysis

It's hard to imagine taking on an MIS project without a project management tool, because project management is integral to every step in software development. However, notes and feedback present a conundrum. On one hand, one could easily argue that e-mail is more than sufficient. But when you see notes attached to specific tasks and the rich information displayed in context within project management tools, it's easy to be swayed back.

On the other hand, the sophisticated features for classifying, securing, and authenticating messages in a system like Lotus Notes makes the simple security features in project management tools laughable. Then again, the tight system of project feedback implemented in such systems as SE Companion in SHL Transform has well-specified protocols for message types and status (such as inquiring, clarifying, deferring, accepting, and so forth), plus secure archiving. With this type of sophisticated feedback mechanism, you can see why problems seldom slip between the cracks or finger-pointing contests of who said what, when, where, and why rarely develop. Regardless, either the notes mechanisms in the project management tools should be improved to the SE Companion level, or they should be ignored.

Likewise, what-if analysis in project management tools leaves several things to be desired. True, the tools I reviewed provide critical path calculations, work breakdown structure costing, baselining, earned value, auto-resource leveling, and so on. Each has a special feature-crosstabs in Microsoft Project, tree views in Scitor's Project Scheduler, trace logic in Primavera's P3, and percent complete indices in ABT's Project Workbench. But the world of management science is rich with network optimizers and robust mixed-integer programming solutions that simply haven't been applied to their systems by project management tool vendors. These optimizing systems could supply the provably best solutions. In addition, contingencies and tradeoffs emerge from other algorithms. In short, project management tools, particularly in the upper tier, are due for a breakthrough in what-if analysis to help better manage the thankless tasks of resource balancing and scheduling.

Making the Tools Work for You

At the outset, I asked two questions. First, are project management tools missing some important functional features? Second, are their various overhead costs too high?

Let's take the second question first. The fairly high ease of use and the high benefits of being able to plan and track projects using standard tools surely outweighs the project management tools' dollar and training costs. Support for team use and interfacing to other systems is improving, especially with ODBC, scripting support, and web enablement. However, all the tools need the support of process planning tools like ABT's Project Bridge Modeler or LBMS's Process Engineer; particularly when you're doing large-scale projects. Table 2 summarizes my rules of thumb for choosing project management tools. For small- to medium-scale projects, either Microsoft Project or Scitor's Project Scheduler are the low-cost, robust, easy to use and easy to customize solutions. When the projects are large (thousands of tasks and resource assignments) or involve multiple updaters, then tools like ABT's Project Workbench or Primavera's P3 quickly earn their keep.

Are the project management tools functionally flawed? For large projects, obviously, yes. Is that enough to explain the nonconsistent usage of project management tools in MIS projects? Emphatically, no, because the incremental benefits of using any one of the four project management tools reviewed here is so high in comparison to using no tool at all. Just having the abilities to change the completion time of a critical task and see what effects that has on the project schedule, or to alter resource availability and test new resource balancing options are worth the price of admission. With the simple expedient of purchasing a tool for planning process management, the biggest flaw in current project management tools is removed, especially for large projects. Therefore, as a project leader, your response as to why you chose not to use a project management tool can only be, "No excuse!"