A decision-making matrix model for estimating construction project overhead

Caroline T. W. Chan
City University of Hong Kong, People’s Republic of China

Construction project overhead—the expenses for onsite construction-related services and facilities (including staff and offices)—comprises a significant percentage of the costs involved in completing construction projects. The recent advancements in information technology enable construction companies to more efficiently and economically prepare project estimates. As a result, these companies can produce requisitions for costs estimates and product specifications effortlessly via the Internet. However, it is important to understand that the new technology mainly benefits the estimation of trade work items, not the project overhead. Past studies have demonstrated the difficulty and uncertainty in estimating project overheads (Chan & Pasquire, 2002; Hegazy & Moselhi, 1995). For construction companies, this means they must perform work when estimating project overhead. And despite this extra effort, the accuracy of their estimates is still comparatively lower than their estimates for other direct cost items. Since estimating is an important first step during project inception, construction companies need to allocate their resources wisely; they must divide their efforts among different tendering operations while achieving the highest level of accuracy. Because of this, researchers must more thoroughly study the current estimation methods and the actual expenditure of project overhead.

Importance of Estimating Project Overhead

Although the literature shows that project overhead is not a major project cost center (Assaf, Bubshait, Atiyah, & Al-Shahri, 1999; Hegazy & Moselhi, 1995), this area can significantly affect a contractor’s success in winning the project bid. Taylor (1994) presented a scenario that if a contractor sublets all or most of a project’s work to subcontractors, they can obtain from these subcontractors equal or offsetting work quotes. Such a scenario is easily adapted to the present-day e-trading market, where virtual communications negate geographical boundaries. In this market, bidders seek out and select the most competitive quotes to build up their tenders. Likewise, materials suppliers and subcontractors monitor the product or service’s market price to ensure they establish a competitive pricing structure.

Surveys by Assaf et al. (1999) and Oteifa and Baldwin (1991) showed that most practitioners ranked the estimating of project overheads as an important estimating and tendering activity. Chan and Pasquire (2002) found, via their survey, that the risks of wrongly estimating project overhead are most difficult to transfer. Very few project overhead items can be sublet. Thus, it is the construction company that has to bear the risks of an inaccurate estimation. These findings demonstrated the importance of project overhead estimation, particularly since estimating affects a construction company’s project performance.

Current Practice of Estimating Project Overhead

In theory, estimates for project overhead are prepared by calculating the time-related costs, such as rental charges, salaries, and fixed costs, including installation and dismantling expenses (The Chartered Institute of Building, 1997). In practice, however, it is the estimator’s professional judgment and intuition that most significantly determines the calculation’s outcome. Tah, Thorpe, and McCaffer (1994) surveyed contractors working in the United Kingdom (UK) and found that project overhead estimates are highly subjective and shaped by management decisions. Chan and Pasquire (2002) studied large contractors working in Hong Kong and found that 57% of the respondents had their project overheads estimated subjectively by senior estimators. Hegazy and Moselhi (1995) reported similar findings in their study of the United States construction industry, as did Fayek, Ghosal, and AbouRizk (1999) in their examination of the Canadian construction industry. Thus, the extent of personal bias in the estimation process is difficult to detect or to eliminate—but it most likely results in estimates containing much inaccuracy.

Hegazy and Moselhi (1995) found that project overhead estimates exhibit a much higher variability among different tenders when these were compared with direct costs estimates, which indicates that overhead estimates are most likely inaccurate. Chan and Pasquire (2002) reported—as detailed in Figure 1 below— that estimators using the Relative Accuracy Index identified 23 project overhead items and calculated each item. The estimators found that many of the project overhead items were difficult to estimate. However, as Chan and Pasquire (2002) and Steward, Wyskida, and Johannes (1995) report, it is the process of estimating these items that consumes most of the estimators’ time and expertise.

Table 1. Summary of rankings of project overhead items (ranked by the likelihood of inaccurate estimation)

Understanding Project Overhead Cost

Although it is worthwhile to conduct more comprehensive studies related to the estimating project overhead, in an effort to improve estimating accuracy, the research in this area is very limited (Chan & Pasquire, 2004; Tah et al., 1994; Taylor, 1994). Solomon (1993) completed a tender cost analysis of the U.K. construction market. In this study, Solomon identified four major project overhead cost centers (amounting to about 80% of the overall project overhead): staffing, mechanical plant, access and scaffolding, and site accommodation. Solomon also found that companies consistently distributed their project overhead costs, regardless of project scope or type. Figure 2 summarizes the contributory percentages of each project overhead item.

Figure 2. Analysis of major project overheads cost by Solomon (1993)

Although this analysis serves as a good guideline for distributing project overhead, these findings may be affected by the estimating strategies used by the examined companies, strategies which are hidden in the cost data. In other words, the tender figures for each project overhead item may not truly reflect the actual cost of such an item in relation to a project.

Objectives

As the above-cited literature reveals, estimating project overhead consumes much estimating resources. However, the processes for estimating different overhead items comprise varying degrees of inaccuracy. Since these processes rely on professional judgment, it is an estimator’s personal bias behind the problem of inaccurate and inefficient estimates. Therefore, the field needs a more systematic approach for estimating project overhead, an approach that will enable construction companies to better utilize their estimating resources as they maximize their estimating accuracy. Because of this, this study aims to achieve two results:

1. To identify the cost distribution of project overhead costs on large construction projects executed in Hong Kong;
2. To devise a systematic matrix model that enables top management to direct their estimating resources with the highest efficiency.

Since the construction practices in Hong Kong are very similar to those used in many Western and Pacific Rim countries, the results of this study can serve as a reference for comparative studies on other international markets.

Data Collection

In Hong Kong, there is an official list of Approved Contractors and Suppliers for Public Works. This list is maintained by the Works Bureau of the Hong Kong Special Administrative Region (SAR) Government for Public Works. The contractors identified on this list as approved are classified into three groups—A, B, and C. The C group contractors are allowed to tender for public works of HK$50 million or more. This group comprise the large—both local and international—contractors that are working in Hong Kong. According to the list, as of January 2005, there are 119 Group C contractors. To generate results that I could compare with the above-cited studies, I studied only those contractors comprising Group C.

I sent invitations to each company, to the attention of the contracts manager. I request access to cost data for one or more of their projects. Due to the confidential nature of project cost data, most of the construction companies initially refused to disclose this information. However, after telephoned the companies, and through referral contacts, I was able to collect information on twenty completed construction projects, information pertaining to project overhead expenses as well as other project details.

Data Analysis

I studied the project overhead cost data together with the contract conditions and bills of quantities. From this, I was able to interpret the project overhead cost by referring to the special requirement of the client or site condition. The contract sum of the twenty projects ranged from HK$120 to 550 million. All involved the construction of new building. Four involved public housing projects; nine, private housing; and seven, commercial buildings. These were all constructed between 1997 and 2001.

Project Overhead Distribution

In general, project overhead accounted for between 11% and 19% of the total project cost, which—at below the rate of 20% of total project costs—is in line with the literature (Assaf et al., 1999; Solomon, 1993). Regarding the cost breakdown structure for the project overheads, I found that the sampled construction companies adopted slightly different strategies for control project costs. For example, some contractors combined the costs associated with the contractor’s site office and the consultant’s site office; others separated these items. Therefore, I combined some cost items to gauge my comparison. Altogether there were twenty-one items consolidated; seven of these (site management, mechanical plants, site cleaning, insurances and surety bond, temporary works, scaffolding) accounted for more than 80% of the total project overhead cost. On the other hand, half of the items listed accounted for a total of 4% of project overhead. Because the companies recorded costs for sundry items, costs for varying purposes (e.g., entertainment, photo processing, etc.), I created the line item titled miscellaneous to capture these costs. Figure 3 details the percentages of these itemized costs.

Figure 3. Analysis of major project overhead costs

Major Project Overhead Costs

As Solomon (1993) found in preparing a project overhead cost breakdown, I found that staffing cost was the largest cost center among all the project overhead costs. However, the relative percentage of this item is much greater in the current study, up to an average of around 36% (in Solomon’s study, it was only around 26%). Expenses for site staff ranged between 24% and 46%. This reflects the current trend among clients: They require contractors to have site management expertise and skills.

The second highest cost center was mechanical plants. Among most of the projects I analyzed, the cost of mechanical plants comprised between 9% and 13% of the project’s cost. Although this item remains as the second most costly project overhead, I found—among the companies I studied—that its relative percentage (12.4%) was less than what Solomon (1993) observed (22.3%). I attribute this difference to the fact that over the last decade the inflation in staff salaries is much higher than the cost of other factors of production (e.g., materials and plants). The relative rankings and percentages for the rest of the project overhead items were quite different from what Solomon found. One reason for this may relate to the difference between tender estimates and actual expenses and to the fact that companies set tender estimates in line with their tendering strategy.

The third highest overhead item is cleaning and rubbish removal. These accounted for 10% of the total project overhead, almost double the percentage reported by Solomon (1993). This increase reflects the fact that today, construction companies are more aware of safety and environmental issues. And good housekeeping and site cleanliness is not only a concern from the clients, it conveys the public image that the construction company is a safety-conscious and environmentally aware enterprise, a key marketing point to see the company to future clients. Therefore, it is reasonable to expect the construction company to spend more money in keeping their site clean and tidy than they did a decade ago.

The fourth item, insurances and surety bond, comprised more than 7% of the total project overhead. This percentage is much larger that the percentage Solomon (1993) found (1.7%). The project managers I interviewed believed that insurance expenses would continue to climb: The terrorist attacks that have occurred during the past few years have increase insurance premiums globally.

Another interesting overhead cost involves site offices, stores, and latrines. It ranked ninth in my study, contributing an average of 3.2% to the overall project overhead. Solomon (1993), however, ranked this item ranked fourth. In the tender, construction companies could list a high cost for site accommodations so as to obtain more money at the start of the project. The tender costs that Solomon used for site accommodation involved a much higher percentage—than the project’s I studied—of the project’s total overhead.

Relative Percentage versus Estimation Accuracy of Project Overhead Items

Using the Relative Accuracy Indices, as calculated by Chan and Pasquire (2004) in a study of project overheads, I cross-compared the average percentages of project overhead costs with the relative estimation inaccuracies. Since the cost breakdown structures of the construction companies do not match entirely with the significant project overhead items, as shortlisted in Chan and Pasquire, I combined the percentages of some items listed in Figure 3 (e.g., site management and watchman, water and power). As a result, I generated a list of sixteen items; these are listed in Figure 4.

Figure 4. Summary of average percentage and relative accuracy index of each project overhead item

From this, I found several interesting relationships. First, some items that the surveyed estimators described as difficult to estimate accounted for only a small amount of the project overhead (e.g., protection, restrictions to noise and dust nuisance, samples/mock ups). On the contrary, some items described as being easily estimated with high accuracy were quite costly (e.g., setting out, insurances). As mentioned earlier, estimators spend the most of their time in estimating project overhead costs, particularly those items involving much inaccuracy. The above data suggest that estimators are not under-performing, that they are wasting their energy in attempting to minimize risks that contribute to relatively trivial overhead costs.

Decision-making Matrix Model for Project Overheads Items

By better understanding the process of distributing project overhead costs, estimators can more effectively plan their activities and resources. Based on the findings above, I suggest that construction companies use a matrix model to select and prioritize the project overhead items they will estimate. Figure 5 outlines the relationship between relative cost and estimation accuracy for the sixteen project overhead items identified in a two-dimensional matrix described above.

However, as Figure 5 shows, many of the points are concentrated along the x-axis. This is because most project overhead items comprise a small percentage of the total project overhead cost. In this regard, I applied a logarithm function to the percentage values of each project overhead item. The adjusted matrix in Figure 6 provides a well-scattered diagram for better differentiation and decision-making.

Figure 5. Percentage of total project overhead cost against relative accuracy index

Figure 6, Decision-making matrix – log
(percentage of total project overhead cost against relative accuracy index

From the decision-making matrix above, I developed four basic approaches for tackling project overhead estimation. They are listed according to level of priority.

1. Apply full capacity (1^st quarter project overhead costs): High percentage contributed to total project overhead cost and estimation inaccuracy.
   This quarter includes: site management; watchman and attendance to nominated sub-contractors; facilities to be provided by Main Contractor (plants, scaffolding); cleaning and rubbish removal; and temporary works (e.g., hoardings, temporary roads, signboard). These items accounted a large proportion of project overhead cost and were the most difficult to estimate. Therefore, full estimating capacity should be allowed to these items.
2. Exert reasonable effort (2^nd quarter project overhead costs): High percentage contributed to total project overhead cost and low percentage contributed to estimation inaccuracy.
   This quarter includes: insurances; power and water supply; site offices, stores, latrines; setting out; and fees and levies. These items were costly but their estimates exhibited a low likelihood of inaccuracy. Since expenditure on these items is generally high, an accurate estimation is necessary. Besides, the uncertainty involved in these items is less and a reasonable level of estimation effort to be paid on them is justifiable.
3. Don’t ignore them (3^rd quarter project overhead costs): Low percentage contributed to total project overhead cost and high percentage contributed to estimation inaccuracy.
   This quarter includes: safety precautions; drawings to be prepared by the Main Contractor; protection of adjacent/existing finished works; and samples/mock ups. These items represented the trivial, uncertain project overhead items. Although the values are considerably low, the actual cost impact can be greater than expected. Hence, these were put in the 3^rd priority and should not be ignored, especially when the estimating resource is available.
4. Estimate with professional judgment (4^th quarter project overhead costs): Low percentage contributed to total project overhead cost and estimation inaccuracy.

This quarter includes: testing; and contract conditions and amendments. These items were neither significant nor uncertain in nature. To save estimating resources, estimates to these items can be simply relied on the estimators’ past experience.

Limitations and Discussions

Due to the sensitivity of project cost information, I limited my data collection to twenty projects. Therefore, these results may not fully represent the spread of project overhead costs. However, the findings provide some new insights into project overhead estimation, particularly in relation to the strategy of allocating limited estimating resources. Construction companies can apply this methodology to their analyze project overhead data so as to improve their decision-making processes when allocating their resources.

In this paper, I show that there exists a clear misalignment between the resources input in the estimation process and the likely cost impact of the items. Although the decision-making matrix that I have outlined aims to assist companies in allocating their estimating resources, I have not yet resolved the root problem in estimating project overhead: to enhance the accuracy and reliability of the estimates. One solution could involve developing an artificial intelligent model.

Conclusions

In this study, I examined the project cost data collected from twenty completed building projects located in Hong Kong. In doing so, I analyzed the project overhead costs after identifying the cost data and project information. Similar to the past literature, I found that the total project overhead cost ranged between 11% and 19% of the total project cost. Amongst the sixteen project overhead items that I studied, I found that site management was the most significant contributory item, accounting for 36% of the total project overhead cost. Other major items included mechanical plants (12.4%), cleaning and rubbish removal (10.0%), insurances and surety bond (7.4%), setting out (6.2%), temporary works (5.9%), and scaffolding (5.9%). When I cross-compared the cost level with the relative accuracy index of the project overhead items, I found that the estimating resources input and the likely cost impact did not align. To effectively allocate estimating resources, I developed a decision-making matrix model and prioritized the allocation of estimating resources according to cost level and the degree of estimating uncertainty/inaccuracy. In doing so, I suggested four levels: Apply full capacity; Exert reasonable effort; Don’t ignore them; and Estimate with professional judgment.