The two broad approaches of cost estimation are the top down approach and the bottom-up approach. The top-down approach entails the project manager charting out overall project size, functionality, process, environment, people, and required quality level, estimate macro level costs, and partition the estimate into a top-down work breakdown structure. The bottom-up estimation models entail aggregating individual estimates for each task in the Work Breakdown Structure (WBS).
The top-down approach allows managers to consider all cost subsystems whereas the bottom up approach allows for a more detailed assessment
The cost of manufacturing any product or undertaking any project is the sum of various input factors such as labor, material, and equipment. A basic cost estimation model identifies the unit costs of each input factor, determines the number of units of inputs required to generate the desired output, and multiplies unit cost with units required to obtain total input cost for the variable. Total cost for the products is the sum of total input value for each variable. Such estimation may at times require application of statistical techniques such as regression analysis to relate the cost of obtaining the required outputs to important characteristics or attributes of the system, or to allocate the extent of joint costs to the specific activity.
This is a simple and straightforward, but laborious method. Very often, a lack of clear-cut information on input costs impede this approach.
An analogy costing model is a non-algorithmic costing model that estimates the cost of a project by relating it to another similar completed project.
Take a completed project, similar in scope, size, structure, environment, constraints, and functions to the current project as its benchmark. Use reasoning and analogy to relate the actual costs incurred for elements of the completed project to similar elements in the current project. Such estimation may be conducted either top-down or bottom-up.
This method has the advantage of assessing costs based on actual project experience rather than theoritical assumptions. Success, however, depends on the extent to which the previous project bears resemblance to the current project, and such relation remains subjective.
A related method is the parametric estimation model, or taking historical information as the base, making assumptions regarding changes, and extrapolating the information to the preset project
The expert judgment costing model makes the assessment of costs by leveraging the experience of one or more subject matter experts (SME).
A common expert judgment model is the Delphi technique, which involves constituting an expert panel, conducting a survey where each expert states their opinion, followed by discussion, and repeating the exercise multiple rounds until all the experts develop a consensus to identify a common cost estimate.
Agile projects rely on the experience of people who actually undertake the work and remain familiar with actual costs, rather than theoretical or boardroom experts detached from the actual operations.
Expert judgment methods are grounded in practical realities but remain highly subjective. Success depends wholly on the skills and competence of the experts chosen.
Parkinson’s Law holds that “work expands to fill the available volume or time.” Using this percept, cost estimation is based on identifying available resources rather than any objective assessment or estimates. For instance, the cost estimate for a software project with a timeframe of 12 months and personnel requirement of five entails direct calculation of the wages of five people for 12 months. This method, though simple and straightforward, may create unrealistic estimates and may not cover all bases.
The price-to-win model estimates cost based on the customer’s budget rather than internal resources or capabilities, and by extension depends on the product pricing, which in turn depends on market forces.
This is a practical approach, for ultimately the net cost is what the customer pays, minus profits, and any excess scope or size invariably gets toned down to match customer budgets. For instance, if a software project was estimated at 500 person-months, it would invariably be toned down to 300 person-months if the customer could afford only that much. On the flip side, this approach does not consider the possibility of projects incurring loss owing to scope creep or any other reason.
The most popular algorithmic cost estimation model for software projects is the Constructive Cost Model (COCOMO II), developed by Barry Boehm and Ellis Harrowitz in 2000.
The basic COCOMO'81 model is a simple static model that considers the software development cost as a function of a program’s size expressed in estimated lines of code.
The intermediate COCOMO'81 model computes software development cost as a function of program size and a set of four subjective cost drivers: product factors, computer factors, constraints, and personnel factors. Product factors include aspects such as required reliability, complexity, usability, size of database, and more. Computer factors include constraints such as execution time, storage, turnaround, and platform volatility. Personnel factors include capability of the analyst, application, language, and tool experience, personnel continuity, and more. Project factors include multi-site development, software tools and more.
The detailed COCOMO'81 model incorporates all characteristics of the intermediate version and also incorporates an assessment of the cost driver’s impact on each step of the software engineering process.
Regardless of the cost estimation model chosen, success depends on a good understanding of how the model works and how to apply it to determine project costs. A project manager is better off adopting a familiar yet limited albeit suitable model rather than another model, which may enjoy better industry recognition but which the project manager remains incompetent to handle.
- Carnegie Million University. “Cost Estimation.” https://pmbook.ce.cmu.edu/05_Cost_Estimation.html. Retrieved July 08, 2011.
- Leung, Hareton & Fan, Zhang. “Software Cost Estimation.” Department of Computing, The Hong Kong Polytechnic University. https://www.st.cs.uni-saarland.de/edu/empirical-se/2006/PDFs/leung.pdf retrieved July 08, 2011.
- “Cost Estimation Models.” https://yunus.hacettepe.edu.tr/~sencer/cocomo.html. Retrieved July 08, 2011.
Image Credit: freedigitalimages.net/renjith krishnan