The importance of accurate estimating cannot be overstated. Many examples abound of projects that experienced substantial cost escalation which can (at least partly) be attributed to poor estimating.
Prior to the start of a project, the tendency is to underestimate costs, due to bidding to secure the project or to sell the project to upper management. Then once the project is underway, the pressures are to overestimate the costs to ensure the project comes in under budget.
The Project Management Body of Knowledge’s (PMBOK) Estimate Costs process contains the policies and procedures for cost estimation. It is the second process out of four within the Project Cost Management knowledge area.
PMBOK, Section 7.2, “Estimate Costs”
Estimate Costs is the process of developing an approximation of the cost of resources needed to complete project work. The key benefit of this process is that it determines the monetary resources required for the project. This process is performed periodically throughout the project as needed.
Three Task Estimating Techniques
There are three primary techniques used in estimating:
- Analogous Estimating
- Parametric Estimating
- Three Point Estimating
In this estimating technique, you use the data from previous, similar projects to determine an estimate for the new one. For example,
- The previous house cost us $175,000 to build, therefore this one is likely to cost $200,000.
Usually there are adjustments to the previous cost data, which can be based on many different factors:
- Time value of money
- Quality level
- Material types
- Environmental conditions
- Labor availability
Analogous estimates should be the first method to try. They are usually the cheapest and fastest to make, but have a correspondingly low level of accuracy.
This method uses an established unit rate to determine an estimate. For example,
- The cost to build a house is $100/square foot.
- Flooring costs $8/square foot to install
- Plumbers cost $25/hour
Often these parametric values are published in trade publications or tracked internally over many projects. They are determined by historical data.
Clearly, for the new project to be estimated correctly, you have to decide if the project will be higher or lower than the parametric value, and by how much. Therefore, you must once again consult the list of factors that contribute to adjustments (above), and decide how much of an adjustment to make for each factor.
The accuracy of the parametric estimate depends on the accuracy of the parametric value being used to develop the estimate. In addition, the parametric value is often scaled up significantly, that is, there are alot of units. Although the relative accuracy of the estimate stays the same, the absolute accuracy increases significantly.
- Estimate #1: $2/inch x 1,000,000 inches.
- Estimate #2: $200,000/inch x 10 inches
The estimated values are the same, but in situation #1 a small increase to the accuracy of the parametric value will result in huge gains in accuracy of the overall estimate (or reduction in losses).
For this reason the scaling factor should be given a significant amount of attention in parametric estimates. Maybe the relationship is not linear and a scaling formula should be developed.
Three Point Estimating
Sometimes it is more intuitive to determine an Optimistic and Pessimistic estimate. It is easy to understand that a task is not likely to exceed a certain value, on the high as well as the low end. If you choose those values first, you can pick the average (technically called the “triangular distribution”).
E = (a + m + b) / 3
- E = The estimated value
- a = The “optimistic” estimate
- m = The “most likely” estimate
- b = The “pessimistic” estimate
If you want to tighten up the estimate to the Most Likely value, you can use the beta distribution:
E = (a + 4m + b) / 6
Often a unique circumstance causes the task to be more likely finish in one direction than the other. You can see that the three point estimating technique can take that into account.
Producing the Overall Project Estimate
There are two ways to roll up the task estimates into overall project estimates:
- Bottom Up Estimating
- Top Down Estimating
Bottom Up Estimating
Bottom Up Estimating is not a task estimation technique. Rather, it refers to the creation of individual task estimates and then rolling them up into an overall project estimate.
The accuracy of the overall estimate is dependent on the accuracy of the individual estimates. That is, two tasks have the same value, the accuracy of the overall estimate is the their average.
Note: This works only if the value of the two tasks is the same. If they are different, the larger one contributes a disproportionate share of the accuracy.
Top Down Estimating
Sometimes you have better information on the project level than the individual task level. Thus, you determine the overall estimate first and then apportion it among individual tasks. This is called Top Down Estimating.
In this scenario, you must still complete individual task estimates to ensure you have enough budget to perform each task. Thus, it is important to apportion with the right balance between tasks.
Top down estimating is rarely as accurate as bottom up estimating. The accuracy of the estimate is only as good as the top level estimate that is determined first. In fact, if the estimate is improperly proportioned, some tasks might have a greater margin of error than the original, larger estimate.
Components of Estimates
Estimating is an art as well as a science. Each task of the project should contain all items of expense, including supply and delivery, as well as an allowance for internal costs.
- Labor. Often, but not always, the largest item of expense, labor can be measured in hours, days, or years, but generally hours is used for small and medium sized projects. Not all labor is the same. A task might require an electrician, and staffing it with the plumber might require additional hours (if possible at all). More realistically, a task might require a higher skill level, such as a journeyman electrician, and if only an apprentice is available, the budget will have to be increased to account for the additional hours required by the lower skill level. Labor is multiplied by its hourly rate to determine the estimate. For internal labor, a realistic estimate of the hourly rate is necessary to account for the fact that the labor is not available to work on other projects during that time.
- Tools and Equipment. Each task must be assigned the necessary tools and equipment to carry it out. This includes everything from software to heavy equipment, but small items like pens and paper can be lumped into a catch-all item (other, miscellaneous, etc.). Many tools are purchased once and used on many projects. In this case they can be omitted if the organization has truly accounted for the cost in another area, however for major items (for example, heavy equipment) the cost should be spread over many projects to ensure an accurate project cost to the organization.
- Materials. These are things that are consumed by the project. Materials can be a major item (such as in construction) or a small item (such as in software development). Unlike the other items, these should be estimated from the specifications, not bottom up from the WBS. For example, a house building project should estimate the amount of building material from the plans.
- Services. This includes the cost of services such as consultants. Usually these are provided on an hourly rate basis which can be included directly in an estimate. However, escalation is a significant concern here, as consultants tend to ask for more money if the required number of hours is greater than expected.
- Facilities. This includes the cost of offices, warehouse space, or any facility necessary to carry out the project. If the office space is already accounted for within the organization, this could be omitted, but any facility whose prime tenant belongs to the project should be included in project costs.
- Sub-contractors. Many projects need external resources to complete parts of the project work. A contract will need to be drawn up, and appropriate contract management needs to be in place to ensure the work is done according to the specifications and payment terms are followed.
- Financing Costs. Often there are additional costs associated with financing of equipment or other products the project needs to purchase. The cost of leasing, renting, insurance, and interest on loans should be accounted for within the project estimates, as well as the time value of money and return on investment metrics inherited from the performing organization.
Estimates always contain trade offs. Since resources are finite and limited, we must balance the use of one resource with the scarcity of the others.
- Time vs. Cost
When you compress the time to carry out a task, you increase the cost, via the addition of more resources
- Quality vs. Cost
If you want to increase the quality level of the final product, you increase the cost.
- Quality vs. Time
If you want to increase the quality level of the final product, you increase the time required to build it.
These three variables, Time, Cost, and Quality, are referred to as the Project Management Triple Constraint. The project manager must decide where on each spectrum they want to be, and tampering with any of the three variables requires the entire project to be re-balanced.
If you add resources to the project (known as “crashing” the schedule), the completion date would be sooner but there is a corresponding associated cost. Additionally, you can add higher skilled resources, which are more expensive but decrease the required time. However, adding more skilled resources results in a diminishing return as decisions tend to be made slower with more skilled resources working on the same task.
Also, labor resources can be “worn out” if their schedule is compressed, and suffer a corresponding decrease in motivation and productivity throughout the project. Finally, workers who are dedicated to one project are more productive than ones that are spread out among several projects.
Levels of Accuracy
There are generally three levels of accuracy in a project estimate:
- Ballpark. Plus or Minus 100%. Ballpark estimates are analogous estimates taken from previous projects. They should not be used as a reason to justify a project. Rather, their only purpose is to ascertain whether a more accurate estimate is necessary.
- Rough Order of Magnitude (ROM). Plus or Minus 50%. ROM estimates utilize analogous or parametric estimates to narrow the accuracy even further. Published parametric values such as cost per square foot for a building project, or cost per line of code for a software project, ensure that a level of accuracy is achieved that can be used to justify a project or make a decision to move forward.
- Definitive Estimate. Plus or Minus 10%. This is the most detailed estimate. It means the estimator has itemized each unit of expense (checklist above) and estimated it using available data. Each item of expense is attributed to a task in the WBS, which is rolled up using bottom up estimating to obtain an overall project estimate.
Contingencies and Management Reserves
A contingency is a small addition to an estimate to account for known uncertainties (known unknowns). Contingencies are added at the task level, that is, final task estimates include all items of expense plus a contingency.
These estimates are then rolled up via Bottom Up Estimating to produce an overall project estimate. Management Reserves are applied to the overall project estimate to account for additional costs that are unknown and could pop up anywhere (unknown unknowns).
Both contingencies and management reserves are derived from the project risk management process, whereby risks to the project’s success factors are identified, prioritized, and ranked. This Reserve Analysis identifies the most important risks to the project which are quantified into contingencies.
The term “estimate” generally does not include profit. If the purpose of the estimate is to bid on a project, profit is added to the estimate and it is called a bid.
Once task and project estimates have been produced, the project Cost Management Plan should contain the background information and assumptions used to derive the estimates. That way, if anything changes, the estimates can be revisited if necessary. This information includes:
- Basis, i.e. how it was developed
- Range and confidence level
- Known constraints
- Known/defined risk