Project Management Institute

Power of the people

by Kenneth G. Cooper

images

IN A WORLD OF RAPIDLY advancing technology-driven new products and systems, could it be that good old-fashioned people management is the key to project success? One survey of project managers suggests just that.

The initial findings from the “$2,000-Hour” Project Management Survey and Simulation explain why spending time in rework just doesn't work.

In a pilot survey of about 100 managers of several hundred projects, four of the top six sources of productivity loss were about people and how they are managed: low staff experience or skill; low staff morale; low availability/experience of supervisors; and overtime/fatigue. The results were consistent across design and build project stages, and across different project types and industries.

I discussed most of these “people” factors in “The $2,000 Hour: How Managers Influence Project Performance Through the Rework Cycle” [Project Management Journal, March 1994] wherein the marginal cost of each additional hour of real progress made by overtime-fatigued staff was calculated to exceed $2,000. In this survey; however, several factors are cited as being even more costly to projects.

What are these project-killing factors, and why are they doing those terrible things? The results of the survey identify the factors, and the simulation model described in the “$2,000 Hour” illustrates the consequences.

The Survey

Nearly 100 experienced project managers, representing several hundred projects managed, responded to a pilot survey to identify the prevalence of different conditions affecting project productivity. The managers ran projects in industries ranging from IT/software and finance, to transport, to aerospace and electronics, to civil construction. Across these varied backdrops, the projects were multi-month and multiyear (averaging two years in planned duration, nine months more in actual duration), with budgets averaging about $9 million (and actual costs averaging $11.5 million).


Kenneth G. Cooper is president of Pugh-Roberts Associates, a division of PA Consulting Group, with offices in Cambridge, Mass., and headquartered in London, England. In a management-consulting career that spans 25 years, he specializes in the development and application of computer simulation models to a variety of strategic business needs.

The $2,000 Hour” Project Management Survey at a Glance

Initial survey results display the consistency of projects experiencing significant overruns, large efforts on rework, and similar sources of productivity loss

Exhibit 1. Initial survey results display the consistency of projects experiencing significant overruns, large efforts on rework, and similar sources of productivity loss.

With project overruns averaging near 35 percent, it is interesting to note that the responding managers estimated that nearly 35 percent additional effort (25/75) was spent reworking tasks previously thought accomplished. Hmmm…that pesky rework cycle (see Exhibit 1).

Finally, we asked managers to evaluate factors affecting staff productivity and rework. We asked them to think of their typical projects, and the conditions that change over the course of the projects. A list of 17 factors was offered for both design and build stages of work. Respondents were asked to assess each factor according to the strongest impact that it has on productivity (work output per person-hour of input) during their project(s). They could rate each factor as having an impact that is not discernible, low (0-10 percent maximum impact), medium (10-20 percent maximum impact), or high (over 20 percent maximum impact). (Note that in compiling averages, we attached values of 0, 5, 15, and 25 percent productivity reduction to those responses.) The average weighted impacts are reported in Exhibit 2, sorted by degree of maximum impact for each of the design and build project stages.

Among the many implications, what stands out is the prevalence of “people” factors at the top of each set of rankings. Indeed, the same four “people” factors sit among the top third (six elements) of each list—low staff experience/skill, low availability/experience of supervisors, low staff morale, and overtime/fatigue. Clarity of requirements specifications and work scope changes complete the top half-dozen in both design and build. The maximum productivity impacts from each of the top-ranked factors range from 12 percent to 21 percent. (Interestingly, there were no dramatic differences in these responses—or, indeed, other response categories—between IT/software projects and more construction-intensive projects. We intended to test this apparent similarity more completely with a larger survey.)

What does this mean for project performance and the prospects for improvement? In the next section and the sidebar we resurrect a version of the simulation model described in “The $2,000 Hour” to demonstrate and test the consequences of several of these project productivity impacts. More important, we'll see what performance improvement we can expect by fixing them.

The Simulation

In a series of PMI® publications [PM Network, Feb. 1993, “The Rework Cycle: Why Projects are Mismanaged” and “The Rework Cycle: How It Really Works…And Reworks”; Project Management Journal, Mar. 1993, “The Rework Cycle: Benchmarks for the Project Manager”; and Project Management Handbook, ed. J.K. Pinto, 1998, Jossey-Bass, “Four Failures in Project Management”] I described a project simulation model built around the concept of the “rework cycle” (see sidebar).

In order to demonstrate the consequences of the factors reported in this survey, and the benefits achievable by improving them, we have used a version of this model that was set up with the average characteristics of the survey respondent group.

The “average” project simulated in this analysis is planned to span 24 months, with a budget totalling 100,000 staff-hours. In the simulation of the as-surveyed “average” project, the actual cost is just over 130,000 hours, about a 30 percent overrun. However, let us enter, via simulation, a magical world in which the top four cited “people issues”—low skill/experience, low morale, low availability/experience of supervisors, and overtime/fatigue—all have been reduced to “low impact.”

Maximum Impacts on Productivity

Staff skills shortages and poor specs top the survey list of design productivity problems, followed by staff morale, scope changes, supervision, and overtime fatigue. The same two factors—specs and skill problems—hurt build productivity most, followed by work-scope changes and a host of “people” factors images

Exhibit 2. Staff skills shortages and poor specs top the survey list of design productivity problems, followed by staff morale, scope changes, supervision, and overtime fatigue. The same two factors—specs and skill problems—hurt build productivity most, followed by work-scope changes and a host of “people” factors.

Simulations of Hours Spent With Typical and Lowered “People” Impacts

Simulations of the “typical” project showed cost savings of nearly 40 percent from improving staff skill, supervision, morale, and overtime fatigue

Exhibit 3. Simulations of the “typical” project showed cost savings of nearly 40 percent from improving staff skill, supervision, morale, and overtime fatigue.

We can run the project simulation model with just these four changes, effectively asking, “What if…What would the project's performance be like if our managerial actions reduced those four problems to insignificant levels?” The staff-hour expenditure in the as-surveyed case is contrasted with the expenditure in the lowered “people” impact case in Exhibit 3. The project is executed for much less than the as-surveyed case (even less than the planned budget)—costing about 85,000 hours; a savings of some 50,000 hours in this magical condition. So, even if we cannot imagine eliminating these problems to this degree, the leveraged impact of making improvements would seem to justify some significant changes in staff management practices. With nearly 40 percent of the project's labor costs as the potential pool of savings, some notable morale-boosting, supervision-enhancing, skill-improving, fatigue-reducing measures would be worth trying!


Reader Service Number 059

The Rework Cycle

Introduced in the February 1993 issue of PM Network, the Rework Cycle structure portrays flows of project work, be it feet of cable, tons of steel, lines of code, design drawings, etc. The structure is the core of a simulation model developed for and applied to dozens of complex design and construction projects (power plants, aerospace developments, electronics, software systems, shipbuilding, etc.).

At the start of a project or project stage, all work resides in the pool of Work To Be Done. As the project begins and progresses, changing levels of People working at varying Productivity determine the pace of Work Being Done. But unlike all other program/project analysis tools and systems, the rework cycle portrays the real-world phenomenon that work is executed at varying, but usually less than perfect, “Quality.” A fraction that potentially ranges from 0 to 1.0, the value of Quality (as well as that of Productivity) depends on many variable conditions in the project and company. The fractional value of Quality determines the portion of the work being done that will enter the pool of Work Really Done, which will never again need redoing.

The rest will subsequently need some rework, but for a (sometimes substantial) period of time the rework remains in a pool of what we term Undiscovered Rework—work that contains as-yet-undetected errors, and is therefore perceived as being done. Errors are detected by “downstream” efforts or testing; this Rework Discovery may occur months or even years later, during which time dependent work has incorporated these errors, or technical derivations thereof. Once discovered, the Known Rework demands the application of People, beyond those needed for completing the original work. Executed rework enters the flow of Work Being Done, subject to similar Productivity and Quality variations. Even some of the reworked items may then flow through the rework cycle one or more subsequent times.

“Traditional” methods can overlook the role of rework and the ever-dangerous <i>undiscovered</i> rework

“Traditional” methods can overlook the role of rework and the ever-dangerous undiscovered rework.

The full simulation models of these development projects employ thousands of equations. They explicitly portray the time-varying conditions that cause changes in productivity, quality, staffing levels, rework detection, and work execution, as well as the interdependencies among multiple project stages. All of the dynamic conditions at work in these projects and their models (e.g., staff experience levels, work sequence, supervisory adequacy, “spec” stability, worker morale, task feasibility, vendor timeliness, overtime, schedule pressure, hiring and attrition, progress monitoring, organization and process changes, prototyping, testing, etc.) cause changes—some more directly than others—in the performance of the rework cycle. Because our business clients require demonstrable accuracy in the models upon which they will base important decisions, we have needed to develop accurate measures of all these factors, especially those of the rework cycle itself. images

What Next?

If these survey results are truly representative of most projects, there are significant implications for performance improvement. Can we find ways to improve our staff's skill, morale, supervision, and overtime/fatigue? We who manage projects and who run organizations with (and depending on) projects would achieve cost savings of up to 40 percent by substantially improving these aspects of staff management…if the survey is correct.

The initial survey findings are sufficiently tantalizing that our next step is to conduct a more comprehensive survey. This summer PA Consulting Group will conduct a survey of much of the PMI membership and publication subscribers and compile and then report the results in an upcoming PM Network. In addition to the “paper” survey, readers will have the option of participating via the Web. Web survey participants will be able to view their survey responses versus the benchmark data of other respondents. (We will provide a customized benchmark report to all valid survey respondents, regardless of the response mode used.) In addition, those who participate by Web will be able to simulate the conduct of their own projects, under their own conditions, and with improved characteristics, so as to see the cost and schedule performance consequences for their projects.

IT SEEMS THAT everywhere you turn these days, someone is saying: “It's the human factors….” Our research seems to add a voice to that chorus, but more in-depth examination is needed. Stay tuned. images

This material has been reproduced with the permission of the copyright owner. Unauthorized reproduction of this material is strictly prohibited. For permission to reproduce this material, please contact PMI.

July 1999 PM Network

Advertisement

Advertisement

Related Content

Advertisement