Integrated project control--the management concept
Every management control system has shortcomings and flaws. For every deficiency, somebody develops supplementary and special solutions. Today, most large organizations can boast numerous accounting, schedule-control, personnel, and cost-control systems.
Multiple systems often hinder an organization more than they help. The care and feeding of several systems increases overhead cost. Incompatible input data ran produce misleading reports. Better cost, schedule, and performance management of large projects requires a return to fundamentals.
The primary purpose of a management control system is to:
Acquire and organize data to facilitate valid decisions by cognizant decision makers.
Almost any system which improves results of a major project even 1 percent through better decisions can justify its cost. But with proper controls, such a system can also serve accounting and other historical functions for payment, billing and legal requirements.
The ideal system would show the manager the best action when internal or external conditions change. It should track all pertinent information inside and outside the organization, and should rapidly produce action guidance for all affected decision-makers (managers or supervisors). This directed action should be best for the organization or the project.
The Babel of existing methodologies and systems actually contains all components to approach the ideal. They only need to be integrated and balanced for effective control of major projects.
Project Control Requirements
Control of a large project is complex. Many decentralized activities must be channeled toward a common goal. Dynamic conditions often demand instant decisions from activity supervisors, but most actions or changes impact on other activities or on the direction of the project. All supervisors need current goal and coordinating information, and managers require reliable and timely data on activity status and available options.
The overall goal is set by management:
- Maximize long-term organizational objectives (profit for most companies).
- Provide the end user (customer) a cost-effective product.
Objective 2 also enhances profit potential, because usually profit cannot be larger than the difference between the product’s value for the ultimate user and its cost to the supplier. This value-cost spread dominates most project trade-offs.
Since major projects last many years, changing user values, new technological options, different resource availabilities, fluctuating material and capital costs, and conflicting demands of competing projects produce continual kaleidoscopic change. Each change affects some major or minor project decision.
An effective manager delegates all but the most central decisions to hundreds of subordinate supervisors. Their independent actions must be coordinated toward the common project goal by communicating user values, management preferences, and impact on related activities. For his own overall or global decisions, he needs timely and reliable data on project status and on available alternatives.
These control requirements call for:
- Organizational schemes that satisfy all data needs from an integrated data base.
- Control schemes that constrain action by lower-level supervisors to the best for the overall project.
The control scheme must be natural to conventional management, and should fit any hierarchical management situation.
Integrated Project Control is a systematic combination of five logical and proven techniques:
- Data Integration. Pre-planning of all data collection for storage in a compatible form.
- Hierarchical Project Structure. A planned formal work breakdown structure (WBS) for orderly data summarization at all project supervision and management levels.
- Incentives. Rewards and penalties representing the value of surpassed or missed goals. Proper incentives communicate to lower-level supervision user needs, management strategy and other external conditions.
- Full-Range Inputs. Joint consideration of all execution alternatives, instead of examining one alternative per activity at a time.
- Optimization. Mathematical procedures to search out and identify the best project actions.
While each individual technique alone might not pay for itself, the combination produces benefits (Table I) that have increased profit of major projects by as much as $57,000,000.
Overview of Paper
Supplementing an extensive description of how Integrated Project Control works (Hollander 1973), this paper focuses on the underlying management concepts. Using techniques from hierarchical control in engineering, the strategic decisions of top management and the external influences are differentiated from the tactical decisions by lower-level supervisors. The result is a comprehensive management concept for controlling major projects or organizations.
The next section describes the relationship of data integration and the hierarchical project organization. Section III develops a universal framework for management control which works with mathematical precision in the most complex environment. Section IV concludes with IPC’s management impact and implementation criteria.
Organization for IPC
Any systems approach to information management and project control depends on a well structured project hierarchy for data management and for control. The information flow must allow all data requirements for decisions, for analysis and for legal requirements to be extracted from a common data base.
Project control requires information on current status and resource availability in various forms. Progress can be measured by expenditure levels or by activity status.
Table I. Benefits from IPC Component Combinations
Inventory systems itemize the availability and actual cost of materials and capital resources, in a common pool or assigned to the project. Financial records identify labor, equipment usage, and material costs by accounting categories. Labor costs are accumulated by crafts, by work orders, by department, by shift, etc. Purchased items can be charged against the project as obligated, as requisitioned, as received, as invoiced, or as incorporated into the end product.
Creative accounting dictates different alternatives to minimize taxes, to maximize progress payments, or to meet conflicting requirements of different regulatory agencies. Since different managers and end users focus on different aspects, most organizations employ specialized systems to accumulate and aggregate data for each specific purpose.
Special information systems lead to multiple accumulation of the same or similar information in incompatible form. Multiple data entry increases input errors, the largest error source in data-processing. But even worse are the wrong conclusions drawn from misleading reports with different cut-off dates for different aggregation rules. (Can your organization instantly reconcile project labor charges with payroll records?)
The problems of multiple specialized systems are overcome by an integrated data base which maintains data in a non-redundant structure for multiple applications. The data categories provide all applications appropriate information from common source data.
Data integration requires planning to ensure that all existing and potential needs can be satisfied with minimum changes and processing. The discipline of advance planning and tight change control pays off in lower data-collection cost, much lower computer processing and storage cost, and better decisions.
Data integration depends on the ability to aggregate the same data for different users. Thus, each major project has multiple concurrent data hierarchies: by organizational units; by location; by end item, by labor, material, and capital-resource classes; and by other legal and regulatory structures. Regardless through which hierarchy data are aggregated, identical measures — say, project cost — must have identical values. Conventional accounting systems provide little help for project control.
Information needs for project management led to work breakdown structures (WBS). The WBS is a formal hierarchy for controlling cost, performance, and schedule of end items, of their components, and of supporting services. The WBS has been a contractual requirement on large U. S. military and space projects since the 1960’s. This rational process has been adopted for other major projects. The principle is simple, but the design of the structure requires time and thought to make the categories at each level exhaustive and mutually exclusive. Fortunately, a WBS applies to an entire end-product class, such as aircraft, ships or nuclear power plants. Thus, the U. S. Government has published specimen WBS for its major military procurements (U.S., 1968).
The WBS defines a uniform relation of users and suppliers. Between different corporate entities, the supplier is a vendor and the user a buyer. Within a company, the supplier produces a sub-assembly or a service for a user department that integrates it into an end product. Departments of an organization can be considered suppliers to the management that sells the product.
For example, when a school board enters into a turnkey contract with a constructor-engineer firm, which appoints a project manager who coordinates the various internal and sub-contractor activities, typical user-supplier pairings might be:
|School Board||Constructor's Mgmt.||Completed building|
|Constructor's Mgmt.||Project Manager||Completed building|
|Project Manager||Engineering||Specifications & Drawing|
|Project Manager||Schedule Control||Milestone or CPM reports|
|Project Manager||Subcontractor||Various construction services|
Regardless of the specific structure in any organization, every project consists of a functional hierarchy in which each user is also a supplier.
Communication and Control for IPC
Large projects employ computers, which work best with uniform and rational procedures. Since decisions are made at many levels, a universal scheme must aggregate status information up the hierarchy and must disseminate strategic guidance downward. It must allow each supervisor to respond quickly to sudden changes in his activity (tactical decisions). It must also process the objectives of the end user and the vendor’s management (strategic factors). Finally, the scheme should use simple concepts that work with mathematical precision in the most complex environments.
This section develops such a scheme by first examining the decision process of a single manager, then extending the process to a hierarchy. Incentives emerge as powerful value indicators for specifying objectives from the user to the supplier.
Simplified Model of Decision Process
When a manager makes a decision, he compares the cost and the value of alternative outcomes, and tries to select the best. The best is usually the largest difference between value and cost, a measure of profit. While he can ascertain cost figures objectively, the value of alternative outcomes depends on his superior’s or his user’s valuation, which he perceives as an incentive. Usually, these incentives encompass many dimensions of performance, schedule, and cost. For illustration, let us examine a specific one, schedule.
Fig. 1. Decision Model
Figure 1 combines the schedule incentive and cost function from a prior paper (Hollander 1973a). The cost function shows the supplier’s cost for different completion times; the incentive, the user’s valuation. If the user pays his value to the supplier as an incentive, the supplier can choose the most profitable completion time, the time with the largest incentive-minus-cost difference.
The incentive guides the supplier to the best schedule. As conditions under his supervision change the cost function, he finds a new optimum by locating the new time that maximizes incentive minus cost. He also finds the new maximum incentive-cost spread after occasional incentive changes.
Figure 1 also shows the importance of setting the incentives properly. Typical project or activity slippages move the cost function to the right (delay) and upward (cost increase), but do not change its shape. Since the supplier seeks the largest incentive-cost spread, the incentive’s slope (e.g., $/day) and horizontal displacement (when the slope is effective) determine his action. A vertical displacement (the actual incentive amount) has no impact on his performance.
The wrong slope gives wrong signals and leads to wrong action. For example, if (the slope of) a schedule incentive is too small, it has no effect: if too large, the supplier sees or receives too much reward for early completion, and the user must pay more than it is worth to him.
Incentives force management to formulate overall values for strategic guidance to lower levels. By communicating user values and constraining the supplier’s decisions, incentives become a powerful management tool.
Structuring Control Policies (Proper Incentives)
Incentives are easy to structure. The user merely penalizes his cost if the goal is missed; or alternatively, rewards his gain if the goal is surpassed. This makes incentives precise value indicators of alternative results, rather than the conventional carrot and stick.
For example, a school board contracts completion of a new school on September 1, when the new school year starts. Since early completion allows the inevitable bugs to be remedied, the board should give a small reward for each day that the building is available during August. Prior to August 1, the maintenance cost would exceed any benefit.
Gains and losses due to late completion can be quantified. If the teachers return for pre-semester preparation on September 1st to an unfinished building, the board loses their salary value. If the building is still incomplete when the children return on September 15, their education value is lost. A community that taxes itself annually $900 per child for 180 school days, values each day of education at least $5.00 per student. On the other hand, the school board saves the interest on the final payment. Each of these factors defines a simple incentive; but since all three apply to the same goal, they can be added into a single compound incentive.
To motivate each supplier to act in the user’s best interest, the incentives must be structured unilaterally by the user. User conditions determine the value of the supplier’s product. “Negotiated” incentives, used in many U. S. government contracts, fail as precise indicators of user values.
The incentives should be the full value of the user’s gain. Some argue that the supplier can often be motivated to achieve the user’s goal with less incentive (U.S. 1969). This implies that the user knows the supplier’s cost function. They ignore that:
- The user may misjudge the supplier’s internal conditions.
- The supplier’s internal conditions may change.
For example, what delivery acceleration would you expect if the user value is $5000/day; incentive, $2000/ day; and acceleration cost, $3000/day? By making the incentive smaller (or larger) than full value, the user creates a false optimum for the supplier (Fig. 1). Many failures of early U. S. government incentive contracts can be traced to wrong incentives (U. S. 1972b).
Extending Decision Model to Project Hierarchy
At each hierarchy level, individuals trade off their perception of value versus cost. To maximize the product’s value-cost spread, end-user values must reach the working levels via the project hierarchy. Whether hierarchy levels are in the same organization or outside contractors and subcontractors, communication and control should be uniform from end user to the working level.
End users can establish values for various levels of cost, product performance, or delivery schedule. For example, car buyers may pay more for better gas economy or for earlier delivery. Since the supplier can make the car more economical or deliver earlier at additional cost, working levels must understand the trade-off for their activity.
In Figure 2, strategic guidance (incentives) from the top down represents the value of achieving alternative results. Status and action alternatives aggregated upward through cost functions show the cost of different actions. At any level, a decision maker finds his best action by comparing values versus cost.
Fig. 2. Decisions in a Hierarchy
Lack of performance at one level can usually be compensated for, at some cost, at the next level. For example, late drawings can be made up by overtime in the shop. The user shows these costs to the supplier by incentives for local trade-offs. Optimizing an entire project involves multiple interactions between incentives and cost functions that require conventional mathematical techniques (Hollander 1973a).
A user at an intermediate level is also a supplier to the next higher level. When that intermediate user develops the incentives for the level below him, he wants to maximize both his internal and his incentive benefits. For example, the constructor’s management must add to the incentives from the school board, the impact of interest rates, of resource balancing, of minimizing overhead — functions that do not concern the school board.
Incentive distribution through the hierarchy becomes mechanical after the top-level has been established. Each user divides among his suppliers, the incentives that are imposed on him plus incentives representing needs of his own operation. Exact mathematical procedures exist in commercial computer programs. Typical incentive examples for the construction industry have been published (Hollander 1973b, Schrier 1971).
Incentives Solve Other Operating Problems
In many cases, such as major construction projects, the owner (user) accepts most risks through a cost reimbursement contract. This reduces the supplier’s usual motivation to produce a cost-effective product. For self-protection, the user will insist on some operational control in the supplier’s organization.
With proper incentives, the user is indifferent to the outcome, because he is compensated for losses if the target is missed, or he pays out any gains from exceeding the target. This makes the user independent of the supplier’s action and keeps the user out of the supplier’s operation. In turn, the supplier can make unilateral trade-offs that maximize his own objectives (profit). Truly an ideal relationship!
Incentives can also replace priorities. When resources are limited, priority service is provided either to entire projects or departments (blanket priorities) or through case-by-case authorizations. Blanket priorities expedite all project tasks, even if they do not need it. Case-by-case priorities are an administrative burden and introduce extraneous factors, such as personal relationships. In contrast, proper incentives reflect the value of time delays for each task, an objective priority criterion mechanically derived.
Proper incentives are also a tool for finely dividing risk and responsibility between user and supplier. Often, a weak supplier cannot accept the full consequences of the user’s losses — one reason for cost-related contracts. Multiple incentives can shift to the supplier part or all of those costs, performance, and schedule risks that he can or should shoulder.
Implication for Management
Integrated Project Control (IPC) clearly separates the strategic from the tactical decisions. At the project level, incentives represent strategic constraints imposed on the project manager by management and the customer. The cost functions describe options communicated from below. Aggregation and optimization algorithms produce the tactical decision data at the project level; and the activity managers make tactical decisions below.
Management moves out of the tactical-decision loop, but must give better strategic guidance. Since policy choices are broader and made in advance rather than in response to an immediate crisis, the value of alternative results can and must be clearly expressed. With such advance guidance, few changes or mishaps will require management’s attention. Instead, the incentives specify the new optimal action.
IPC reduces management risks through reliable reports generated from consistent data. Overruns due to misleading reports have threatened even large conglomerates with bankruptcy. The U. S. government now demands integrated reporting on its major procurements (U. S. 1972a).
The IPC system or its subsystems can be managed at the line or staff levels currently responsible for similar functions. To ensure continued data compatibility, the discipline must be maintained by a single data coordinator.
Implementation of Integrated Project Control
Any decision-maker can apply, formally or informally, some of the IPC concepts at little cost. However, formal introduction of IPC for a major project requires thorough planning of the control hierarchy and the accounting structures, and positive expression of management objectives.
Planning for IPC is comparable to planning any new undertaking. On-going operations may have to change existing procedures and data bases. Such a drastic change can be expensive and disruptive. For example, briefing each supervisor of 20,000 activities only one hour requires 20,000 man-hours.
The best time to introduce IPC is at the start of a major undertaking. But how can existing operations make the transition economically? First, management must establish IPC as a future requirement. Staff or line functions can then design an ultimate system structure with due consideration for the transition. This planning effort is reasonable.
Any subsequent system or procedure changes must conform to this new, approved plan. The organization will evolve steadily toward Integrated Project Control without extra disruption. New data bases are created in the proper format. Existing data bases phase out as they become obsolete, and the remainder can be translated when the new system predominates. The gradual revision of forms and procedures is now focused toward a single goal, IPC.
Integrated Project Control (IPC) is the systematic integration of customer requirements, supplier-management values, and status information from all data sources in compatible form for rational decision-making. Data integration reduces the cost for data entry and processing and ensures compatible reports for decision-makers. IPC demands advance planning, but produces operational economies and better decisions. Better results repay the initial planning effort many-fold.
Most of the benefits summarized in Table I rely on:
- Formal description of alternatives by incentives and cost functions.
- Treating incentives as precise indicators of customer and management values.
Such a universal framework for management control works with mathematical precision in the most complex environments, which facilitates computer implementation.
Management of major projects with IPC provides:
- Greater profit through identification of best decisions and actions by rational rules or by computer program.
- Fewer crises by pre-specification of the more constant strategic constraints.
- Better utilization of management talent by separating strategic and tactical decision processes.
- More reliable reports by use of compatible source data.
- Lower information cost through less data entry, storage, and processing.
The references (Hollander 1973b, Schrier 1971) describe actual procedures and applications with TOPS™, a program that implements Integrated Project Control.
Hollander, Gerhard L., (1970), Sequential Decision Making for Complex Projects, Proceedings 1970 System Science and Cybernetics Conference, 70C43, p. 243.
Hollander, Gerhard L., (1973a), Integrated Project Control — Control to Maximize Profit, Project Management Quarterly, IV: 1, p. 6.
Hollander, Gerhard L., (1973b), TO PS/Schedule: A Module for Integrated Project Control, Project Management Quarterly, IV:2, p. 6.
Morse, Robert V., (1971), Control Systems for Better Project Management, Computer Decisions, 111:7, p. 28.
Schrier, R. J. and Tilley, E. A., (1971), TO PS/Schedule: A Construction Industry Example, Proceedings, 3rd Annual PMI Symposium.
U. S. Govt., (1968), Department of Defense Military Standard 881: Work Breakdown Structure for Defense Materiel Items.
U. S. Govt., (1969), Department of Defense/National Aeronautics and Space Administration, Incentive Contracting Guide, U. S. Govt. Printing Office: 1969 O-364-685.
U. S. Govt., (1972a), Department of Defense Instruction 7000.2: Performance Measurement for Selected Acquisitions.
U. S. Govt., (1972b), Office of the Assistant Secretary of Defense, Report on the Interservice Audit of DoD Procurement-Incentive Contracting, Comptroller Report No. 336.