Estimation plays an important role in project planning by making it possible to set realistic objectives, allocate the necessary resources and define appropriate deadlines. Accurate estimation provides a solid basis for drawing up detailed plans, identifying critical milestones and anticipating the effort required at each phase of the project. This helps to create a clear and structured roadmap.
Estimates also influence decision-making by providing essential data to assess the feasibility of a project and manage potential risks. They allow decision-makers to compare the expected costs and benefits, thereby helping to determine whether a project is worth undertaking.
In traditional project management, by identifying uncertainties and incorporating safety margins, project managers can plan appropriate mitigation measures and ensure that the project stays on track, even when the unexpected occurs. Furthermore, estimates serve as a benchmark for measuring progress and adjusting plans according to actual results, which improves the accuracy of future estimates and strengthens the organisation's planning capabilities.
What does estimating mean?
Estimation attempts to predict a quantity (value, cost, etc.) as accurately as possible, using the information available at the time of the estimate. An estimate is therefore an assessment based on the available data, past experience and the assumptions made to compensate for uncertainties.
The difference between estimation and forecasting
It is important to distinguish estimation from forecasting. While estimation consists of assessing future needs based on limited and often uncertain information, forecasting involves a more precise prediction, generally based on historical data and proven models. The unique nature of projects prevents having a sufficiently broad base of experience to create convergent and reliable estimation models. Each project has distinct characteristics, specific contexts and unforeseen variables that make it difficult to use past data to predict future needs.
A guessing game
Most people associate the term estimation with "guessing". This perception stems from the fact that estimates are frequently made with incomplete information and a limited understanding of how events will unfold in the future.
What should be estimated in a project?
We focus on estimating four key aspects of a project: cost, value, durations and workload.
Cost estimation
Cost estimation is essential for establishing a project's budget and for ensuring that the necessary financial resources are available. It includes assessing external expenses such as consultants and the services and materials required to deliver the project's deliverable, as well as internal costs such as the cost of employees working on the project and the infrastructure used by the project. Estimating external costs makes it possible to anticipate funding needs.
Value estimation
Value estimation focuses on the benefits expected from the project, whether financial or non-financial. It involves assessing the advantages the project will bring to stakeholders, such as increased revenue, improved operational efficiency or greater customer satisfaction. Understanding a project's potential value helps to justify investments and to prioritise initiatives according to their contribution to the organisation's strategic objectives.
Duration estimation
Duration estimation aims to determine the time required to complete each phase and activity of the project. This includes assessing the deadlines for completing tasks, coordinating the dependencies between activities and managing any delays. Accurate duration estimation makes it possible to draw up a schedule and to manage stakeholders' expectations regarding delivery deadlines.
Workload estimation
Workload estimation concerns assessing the effort required from human resources to carry out the various tasks of the project. This includes identifying the skills required, distributing tasks among team members and planning work periods. Good workload estimation makes it possible to optimise the use of resources, avoid work overload and ensure that the project has the capacity needed to achieve its objectives within the allotted timeframe.
The challenges of estimation
Factors influencing the accuracy of estimates
The accuracy of estimates in project management is influenced by several factors. The quality of the information available at the time of the estimate plays a crucial role. Incomplete or inaccurate data can lead to erroneous estimates. The experience and skills of the people involved in the estimation process are also decisive. The estimation techniques used likewise affect the accuracy of the results. Finally, external conditions, such as changes in the project environment, can influence estimates, making certain aspects of the project even more difficult to predict.
Common problems in estimation
Several common problems can affect the quality of estimates. Bias is a frequent factor that can distort the results. Optimism bias, where estimates are systematically too optimistic, and pessimism bias, where they are too cautious, are two examples of biases that can influence estimates. Uncertainty is another major problem. Even with high-quality data, there are always unknowns that can affect the accuracy of estimates. A lack of data is also a recurring issue, particularly in innovative or unique projects where few precedents exist. Finally, pressure from stakeholders to provide favourable estimates can lead to compromises and to estimates that do not faithfully reflect reality.
The problem of backward planning
The problem of backward planning (reverse scheduling) arises when deadlines are set before accurate estimates have been made. In such situations, project teams are often forced to artificially adjust their estimates so that they match the imposed dates. This practice can compromise the accuracy of estimates and lead to unrealistic plans. Backward planning increases the risk of overrunning deadlines and costs, because the estimates do not faithfully reflect the reality of the tasks and resources required.
However, this approach is legitimate from the customer's point of view, as it makes it possible to set clear expectations and deadlines for delivery. It is crucial to make clear from the outset that, if the date is fixed, not all features or deliverables may be included. By establishing this guideline from the beginning, the estimates will be less biased and will be able to better reflect the realities of the project, while making it possible to manage the customer's expectations realistically.
Consequences of inaccurate estimates
Inaccurate estimates can have significant consequences for a project. They can lead to cost overruns, where the initial budget is insufficient to cover the project's actual expenses. Deadlines can also be affected, with projects taking longer than expected, which can compromise the delivery of the expected products or services. Moreover, erroneous estimates can lead to an inefficient allocation of resources, with teams underused or overloaded with work. This can also affect stakeholder satisfaction, as expectations are not met, leading to a loss of trust and credibility for the project management team. Ultimately, inaccurate estimates can compromise the overall success of the project, jeopardising its objectives and expected results.
The problem of estimation uncertainty
Barry Boehm's cone of uncertainty[1]
The cone of uncertainty[2], introduced by Barry Boehm, is a key concept for understanding the uncertainty inherent in project estimates. It illustrates how uncertainty decreases as the project progresses and knowledge accumulates. At the start of a project, uncertainties are at their highest because the information is limited and approximate. As the project advances, the information becomes more precise and the estimates can be adjusted accordingly.
400% margin of error - initiation
At the very start of a project, information is very limited and uncertainties are high. At this stage, the knowledge required is limited to a basic understanding of the project's objectives, the preliminary requirements and the major constraints. Key activities include identifying stakeholders, the initial definition of scope and drawing up the project charter. The estimates of costs, durations and workload are largely hypothetical and can vary enormously, by up to four times more or less than the final reality.
200% margin of error - definition of needs and requirements
At this stage, an initial draft of the project's requirements and specifications is produced. The knowledge required includes more detailed requirements, the identification of the main deliverables and the first estimates of the resources needed. The key activities at this stage involve analysing stakeholder needs and developing the first functional and technical specifications. Although the estimates become more precise at this point, they still remain broad, with a significant margin of error that can reach 200%.
100% margin of error - planning
When the preliminary planning of the project is carried out, a clearer overall picture emerges. The knowledge required at this stage includes an in-depth understanding of the requirements, the technical specifications and a detailed identification of the resources. Key activities include drawing up a detailed project plan, defining the main milestones and estimating durations and costs more rigorously. At this stage, the margin of error of the estimates is reduced to around 100%, reflecting a more precise but still improvable understanding of the project's needs.
50% margin of error - design, defined solution
In the design phase, the knowledge required includes complete documentation of the requirements and the final technical specifications. Key activities at this stage include drawing up detailed plans, defining system architectures and creating prototypes. At this stage, the margin of error of the estimates is reduced to around 50%, because the information is now more complete and the initial assumptions are validated or adjusted. This step makes it possible to ensure that the design elements are aligned with the project's objectives and that they can be delivered within the estimated time and budget.
25% margin of error - specifications complete and user interface defined
When the specifications are complete and the user interface (UI) defined, the project enters a phase where the estimates become even more precise. The knowledge required includes a detailed understanding of the final features, the performance requirements and the user interfaces. The key activities at this stage involve finalising the detailed specifications, validating the UI mock-ups and refining the estimates of the time and resources required. At this stage, the margin of error of the estimates is reduced to around 25%, because the precise details of the project are now well understood and can be planned with greater certainty.
10% margin of error - delivery
Finally, as the project progresses into its delivery phase, the knowledge required includes a complete understanding of the expected results and a clear command of the delivery approach. At this stage, the margin of error of the estimates is reduced to around 10%, reflecting a high level of accuracy based on complete data and a retrospective analysis. It is important to note, however, that a variability of 10% remains present even in the estimation of small tasks, owing to unforeseen factors and the dynamic nature of projects. This variability is impossible to eliminate completely.
Comparison with PMBOK
The PMBOK (Project Management Body of Knowledge) also indicates that initial estimates are often based on limited information and that their accuracy improves as the project's details are progressively elaborated. It encourages the use of techniques such as reserve analysis to manage uncertainty and to include margins for contingencies in estimates. This method makes it possible to anticipate contingencies and to mitigate the impacts of potential risks on the project. By incorporating reserves, project managers can better manage the unforeseen variations that may arise.
The PMBOK also emphasises the importance of continuous reassessment and of updating project plans to manage uncertainty and risks, thereby ensuring that plans remain relevant and adapted to new information and changes.
Comparison with HERMES
HERMES emphasises a clear and detailed structure for the planning and execution of projects. The method stresses the importance of precisely defining objectives and deliverables right from the start of the project. However, unlike the cone of uncertainty, which uses specific margins of error to illustrate the decrease in uncertainty, HERMES does not adopt such an explicit scale of variability. The absence of this scale can make it more difficult to visualise the progressive reduction of uncertainty as the project advances.
HERMES offers well-defined project stages and clear roles and responsibilities for each phase, which helps to reduce uncertainty through better organisation and a clarification of expectations. This method also emphasises the importance of communication and documentation throughout the project. By ensuring constant communication and accurate documentation, HERMES helps to minimise misunderstandings and to ensure that all stakeholders share a common understanding of the project's objectives and progress.
Estimation techniques
Estimates must be made by people who are competent and experienced in the area being estimated. When individuals without adequate expertise in the subject attempt to provide estimates, they risk simply injecting random values into the system, which seriously compromises the accuracy and reliability of the forecasts. Clearly stating that one does not know is very valuable information for risk analysis and for assessing the quality of estimates. It makes it possible to recognise areas of uncertainty and to identify the areas requiring particular attention or additional expertise.
Analogous estimation (top-down)
Analogous estimation, also known as "top-down" estimation, consists of using data from previous similar projects to estimate the costs, durations and resources required for the current project. This method relies on past experience and makes it possible to compare the current project with earlier projects in order to obtain faster and often more accurate estimates. Although this technique is effective, it depends heavily on the similarity of the projects compared and on the quality of the historical data available.
Bottom-up estimation
Bottom-up estimation consists of breaking the project down into smaller tasks, estimating each of these tasks individually, and then aggregating these estimates to obtain an overall total. This method offers a high level of accuracy because it examines each component of the project in detail. However, it can be very time-consuming and requires an in-depth understanding of each task of the project. This technique is particularly useful for complex projects where a detailed estimate is needed for precise planning and resource allocation.
Three-point estimation
Three-point estimation uses three scenarios for each estimate: optimistic (best case), pessimistic (worst case) and most likely. Using these three data points, a simple average can be calculated
This method provides an overview of the various possibilities, but may lack precision because it treats each scenario equally.
The PERT[3] method (Program Evaluation Review Technique) refines this approach by calculating a weighted average. The formula for the PERT average is
This method gives more weight to the most likely scenario, which makes it possible to obtain a more realistic estimate.
In addition to the average, PERT uses variance to measure the uncertainty associated with the estimate. The variance is calculated as follows:
The variance helps to understand the extent of the uncertainties and to quantify the risk associated with the estimate. The greater the variance, the higher the uncertainty.
It is important not to communicate only a single-point estimate, but always to provide a range or an average with a variance. This makes it possible to signal the uncertainty and to manage stakeholders' expectations. By presenting a range or an average with a variance, project managers can better prepare contingency plans and anticipate deviations from the initial forecasts.
Parametric estimation
Parametric estimation uses algorithms and statistical models based on known parameters to calculate estimates. For example, the cost per unit of production or the duration per unit of work. This method relies on historical data and mathematical relationships to produce estimates. It is particularly effective for projects where precise data and well-defined models are available.
Using use case points[4]
A specific method of parametric estimation uses use case points to estimate the development effort. This technique assesses the complexity of the use cases and of the actors involved in the system, then applies adjustments to reflect the technical and environmental complexity of the project.
The elements that go into the calculation are: the technical complexity factor (TCF) and the environmental complexity factor (EF). These coefficients require calibration based on previous projects to ensure their accuracy and relevance in estimates. Calibration makes it possible to adapt the models to the specific features of each organisation and each project, thereby guaranteeing more reliable forecasts that are adapted to the realities on the ground.
[1] Barry Boehm was a computer scientist and software engineer renowned for his major contributions to systems engineering and project management. He obtained his doctorate at the University of California, Los Angeles (UCLA). Boehm worked for prestigious companies and organisations such as TRW Inc. and DARPA. He is notably famous for having introduced the spiral model of software development and the concept of the cone of uncertainty. His work has had a lasting influence on estimation and risk management methods in projects.
[2] IEEE Transactions on Software Engineering, SE-10-1, 1984, DOI :10.1109/TSE.1984.5010193🔗
[3] PERT (Program Evaluation and Review Technique) is a project management method that helps to plan and coordinate the various tasks. It is based on the graphical representation of the stages of a project, called a PERT diagram, making it possible to identify the critical tasks and to estimate the time required for their completion. Developed in the 1950s by the US Navy for complex projects, this technique aims to improve project planning and control by taking into account uncertainties and variations in time.
[4] Applying Use Cases: A Practical Guide, Geri Schneider, Jason P. Winters, Addison-Wesley, 2001🔗