Project Forecasting consists of taking the project status information and extrapolating the current project performance to the end of the project. Forecasts can be made with respect to project duration, overall project cost, performance/quality level of project deliverables, or any combination of these.
A key element in forecasting is to review the risk events that occurred and the remaining risk triggers. A caution when doing forecasting, ensure you have adequate information to realistically forecast performance. A general rule of thumb is to wait until an activity, phase, or deliverable is at least 25% - 40% complete before trying to forecast. Prior to that point one should stick with the original estimate, modified by any appropriate risk mitigation activities that have occurred.
Forecasting arms the project manager with valuable knowledge enabling proactive project and resource management. It creates confidence in meeting client demands and allows thoughtful consideration for outsourcing. Embracing a more proactive philosophy may just inspire others to follow the lead.
When forecasting project duration, the key is to understand the schedule performance and schedule risk of the activities on the critical path. Those activities will be the ones that drive the project completion date. On a resource constrained project, or a project with unpredictable resource availability, this can be very difficult because the lack of resources causes the critical path to vary. It generally comes down to expert judgment and gut feel. When it is vital for the project to be completed by a certain date, a suggestion is to convert the schedule tracking to a countdown mode where everything is measured in terms of how many days before project completion. Also, holding meetings more frequently in order to quickly assess when it is falling behind is a good option.
When forecasting total project cost, it is good to rely on forecasting methods that are embedded in the Earned Value Management system. Unfortunately, many organizations do not have the financial systems in place that enable earned value management. When that is the case, the project manager should rely on trend forecasting - which is sometimes called "straight-line" forecasting. Trend forecasting takes the current project spending and extrapolates that rate of spending until the end of the project. This provides a rough forecast, but it does not take into account the effect that different activities may require resources that spend at different levels. The resources that perform the remaining activities may be higher or lower cost than the preceding resources. Also, it does not take into account that the project may be ahead or behind schedule. If the project is ahead of schedule, the spending done to achieve that condition inflates the extrapolated value of the project final cost. If the project is behind schedule, the lack of spending creates an extrapolated value of total project cost that is too low.
When forecasting the performance or quality of project deliverables, the project manager could rely on prototypes and preliminary analysis. When the project does not have these, the risk that the project will not achieve the desired performance or quality established at the time of project planning is higher. If performance is the most important attribute of the project deliverables, then the risk of missing the forecast project duration or cost is much higher. The principle involved is the "Rule of Ten's." According to this principle, the cost to correct a technical issue goes up by a factor of 10 as the project moves from one phase to the next. Therefore it is imperative that performance issues be identified as early as possible.
The following table lists predefined spread curves provided by Oracle Fusion Projects.
Ensures a linear distribution of financial or project plan values across periods.
Use case include preliminary and resource cost, e.g. Engineer, administrator, vehicle, office rental, etc. These costs typically will be incurred from the beginning until end of project at almost same rate.
10-10-10-10-10-10-10-10-10-10
Ensures a back-loaded distribution of financial or project plan values across periods. Assigned amounts increase over succeeding periods.
Use case include demobilization works, documentation, inspection/testing works, where the cost may incur very little at beginning but will incur more cost at end of the duration.
0-5-10-15-20-25-30-35-40-45
Ensures a back-loaded distribution of financial or project plan values across periods. Assigned amounts decrease over succeeding periods.
Use case include mobilization works, temporary works, project setup, which will incur more cost at beginning of the project duration. These can also apply to cost item where you need to make massive payment before construction of the work e.g. procurement of Mechanical equipment.
45-40-35-30-25-20-15-10-5-0
Ensures an S-shaped distribution of financial or project plan values across periods.
18-10-8-10-15-17-18-17-15-8
Ensures a bell-shaped distribution of financial or project plan values across periods. Assignment of plan values is highest in the middle periods.
Typically use for construction works which will gradually increase in cost and reach to peak at the middle of construction period.
0-4-10-12-14-12-10-4-0-0
Manual and custom curve are very subjective, if there are some fixed costs which the company know exactly when it will incur, then Manual and custom curve can be applied. Manual curve will require the user to enter the monthly value in the time-phase spreadsheet (see screenshot below). Whereas Custom curve is predefined by user in configuration and value will be spread by system based on the predefined formula.
Reference
Forecasting Methods and Applications 3rd Edition, Steven C Wheelwright. WSE Publishing, 2008.
https://www.soa.org/sections/pred-analytics-futurism/prof-forecasting-methods.pdf
https://docs.oracle.com/cd/E25054_01/fusionapps.1111/e20384/F552278AN652C7.htm