Analog and parametric estimation is what we will discuss today, partly to continue the discussion started in the last article about how to estimate the resources of a project.
Project estimation is a key aspect of project management and analogical and parametric estimation techniques are the most commonly used methods.
These kind of estimates are universally applicable to any kind of project of any kind of organization.
They are used to perform any type of estimate, whether it concerns time/duration, effort, resources, or costs.
Analog estimation is a technique used to estimate the duration or cost of an activity or project using historical data from a similar activity or project.
Analog estimates are made based on the time or cost employed by similar previous projects.
Therefore, these estimates are based on team experience or project history.
The only disadvantage of this method is that the estimate may not always be accurate.
By applying this method, historical data from similar previous work can be used to estimate the current work, however, be careful when applying this method.
This should only be used when reliable data from similar works is available. Otherwise, this method may be counterproductive.
Here is an example:
Let’s suppose you want to estimate the timing required to paint a house. Let’s also assume that you have access to data – quality data – of the actual duration of another similar project in the past. This data could come from the same house or one with a similar structure and size, painted in the same location and during the same season – there may be a difference between painting in summer or winter.
The parametric estimation is executed on a unitary basis and employs the ratio between variables to arrive at the cost or duration of an activity or project.
Compared to the analog estimation, the parametric one is more accurate, but the measurement must be scalable to confirm accuracy.
You can use this method only after you have identified one or more parameters and devised an algorithm or formula to perform the specific calculations.
These calculations will then be performed on the historical data obtained and, unlike the previous technique, it is not necessary for the historical data to derive from a similar work or project.
The algorithm, or formula, should be good and effective enough to produce predictable results. Otherwise, this method cannot be used.
Here is an example:
Let’s suppose you want to estimate the duration to paint a house and you know the size of its walls. Let’s also assume that you have access to detailed data on the actual duration of another project, where another building – this time with different sizes – was painted some time ago in another location.
In this case you can use two parameters to get to the parametric estimation: These are:
- first parameter = size of the overall surface area to be painted
- second parameter = average duration for painting one square meter of surface area
Our formula for the estimation will therefore be:
Estimated duration = (size of the overall surface area to be painted) * (average duration for painting one square meter of surface area)
Thanks to this formula it is therefore possible to estimate the duration required to paint the house of the project.
Let us now use specific data to illustrate the example more clearly:
- It is necessary to paint 1000 square meters of walls of a house.
- On average, it takes 6 minutes to paint one square meter of the area.
- So you will need 1000 * 6 minutes = 100 hours to paint the whole house.
Difference between analog and parametric estimation
Many professionals misunderstand the difference between analog and parametric estimation because they believe that parametric estimation does not need historical data. This, as seen above, is not true.
Historical data are in fact used in both estimation techniques of the project, only in the analogical one data similar to the project in question are required, while in the parametric one more general data are required.
Here are the main differences between these two techniques in the following table.
|Uses historical data from a previous similar project
|Uses a formula based on historical data from a previous project
|When is it used?
|Usually in the early stages of the project when only high quality data can be used
|Usually when comprehensive data is available
|Only when quality data from a similar project is available
|Only when it is possible to devise a formula, an algorithm or a statistical model
|Usually less accurate. Depends on the expertise and experience of the individual making the estimate
|Usually more accurate. It depends on the accuracy of the available data and the refinement of the model used for the calculation
|Usually less expensive in terms of effort and time
|Usually more expensive in terms of effort and time
Generally speaking, the estimation of a project’s activities is not always easy.
How come? Usually because the only time you know exactly how long it takes to complete a project is when it comes to completion.
Up to the delivery point, the project teams led by the project manager, they employ “guesswork” to predict the future.
And the bigger and more complex a project is, the more confused the future will be.
Wrong estimates mean missing deadlines and over-budget spending, two of the main causes of a project’s failure.
Being an experienced estimator is therefore a key skill for a project manager and using proper project management software can be a great advantage in this case.