A comprehensive guide to parametric estimating (with tips)

By Indeed Editorial Team

Published 6 June 2022

The Indeed Editorial Team comprises a diverse and talented team of writers, researchers and subject matter experts equipped with Indeed's data and insights to deliver useful tips to help guide your career journey.

Project managers often rely on statistical methods to estimate the costs of solutions and projects in terms of time, money and resources. Since parametric estimation is a strategy that can reliably create estimates using existing data, it can be a useful tool for project managers to make predictions. If you want to learn more about project management strategies, understanding how parametric estimation works can be beneficial. In this article, we define parametric estimation, provide a step-by-step guide on how to use it, explore its advantages and limitations and offer tips for improving its accuracy.

What is parametric estimating?

Parametric estimating is a project management strategy for estimating time, resources and costs using historic data. It uses a statistical relationship between the historic data and other variables to reliably estimate a result. It usually produces a more reliable estimate than projections from some hypothetical standard. This quantitative approach is often a standard in project management, and many businesses rely upon it to inform resource allocation, investment decisions and production deadlines.

Related: Project management skills and how to improve them

When do professionals use parametric estimation?

Professionals typically use parametric estimation when they want to reliably forecast a numerical statistic in the future, such as the time it might take to complete a project or its monetary cost. They can only use parametric estimation if they have available historic data from previous projects, such as the time it took to complete them and the resource costs they required. If you lack the historic data to perform this estimation, using similar case studies from other businesses of parallel character may be useful. Alternatively, other estimation strategies may offer realistic estimates.

What are the two types of estimations?

Parametric estimating can yield either deterministic or probabilistic estimates. A deterministic estimate is a single number value for the parameter, which the professional can adjust to account for differences between historic and current projects. A probabilistic estimate usually produces a range of estimations, varying between optimism, likelihood and pessimism. Professionals often plot these possibilities as a curve to find the most likely outcome. This works similarly to interpreting a triangular or PERT distribution.

Related: How to calculate probability (with real-life applications)

How to use parametric estimating

A parametric estimation's complexity can depend on the objectives and scale of the project, but it generally follows a similar process you can replicate. Here is a step-by-step guide to using this estimation:

1. Identify what parameters to estimate

You may start with identifying which aspects of the project you can parametrically estimate and what sort of figure you would like to receive as a result. There are several criteria that may affect how well you can estimate a parameter, such as its level of accuracy, the availability of historical data and how well it correlates with cost and time. Here is a simple example of parametric estimation:

Example: A project manager wants to estimate how long it takes their team to deal with 125 enquiries in the present. In this case, they want to measure time, and the parameter they want to measure it against is a record of how many enquiries the team can historically complete.

Related: Project management KPIs: Definition and examples

2. Find historic data to inform the estimates

You can use historic data from company records or estimates from similar projects to inform your estimations. Internal databases of costs, time and resources that record values from previous projects can be the most useful resources. In the absence of these, public data on statistics and industry-standard benchmark values may be helpful in a manufacturing context. Using assumptions and other data typically creates a different kind of estimation. You can input historic data into the previous example:

Example: The project manager finds a database tracking how many enquiries the team has managed to deal with daily over the past year. They calculate that, on average, the team can resolve 50 enquiries during an 8-hour working day.

Related: 13 milestones in project management (with definitions)

3. Identify parameters that correlate with cost and time

Once you've built a data set, you can select concrete parameters that correlate with cost and time requirements. Measurements such as produced units and quantities like size and weight are commonly correlative with cost and time. If necessary, you can test these parameters for correlation and regressions using statistical software and models. Here is what this step may look like in the case of our example:

Example: The project manager establishes that there's a clear relationship between the time it takes to complete enquiries and the number of enquiries the team can complete in a day. They can break this down further to number of enquiries per team member in a day, or enquiries in an hour, but the total in a day is usually enough for the estimation requirements.

Related: What is Monte Carlo analysis? A definition with examples

4. Calculate the parametric estimate

Once you've selected the parameters, you can calculate parametric estimates. In its simplest form, each estimation consists of one parameter in a linear relationship with either cost or time. Here's a simple formula you can use for this calculation:

(historic cost or time / historic parameter value) x current parameter value = parametric estimate

You typically take the cost of a previous project, divide that by whatever metric it used and multiply it by the current metric. This simple concept can yield reliable estimations in the present if your projects have measurable milestones. Here's how you can apply this calculation to our example:

Example: The project manager performs the following calculation: (8 hours / 50 enquiries) x 125 enquiries = 20 hours. Since it took the team 8 hours to complete 50 enquiries in the past, the manager can estimate that it would take them 0.16 hours per enquiry, which would amount to 20 hours to complete 125 enquiries.

Advantages of parametric estimation

Parametric estimation can be very useful for making estimates relating to measurable performance in historic data. Here are some of its advantages:

  • More reliable than guessing: A parametric estimation uses historic data of a team's actual performance in the past, so it can be a very reliable way to forecast a team's future performance. Since it uses a statistical relationship, it's also less subjective than a more descriptive assessment of performance.

  • Easily scaleable: Professionals can upscale and downscale the model that parametric estimation provides to any quantity or required units of work. In this way, the figure it yields can help predict figures for both larger and smaller projects easily.

  • Highly flexible units: Since parametric estimation relies on a simple statistical approach, professionals can simply swap the units of measurement to suit the kind of parameter they're testing. It can be suitable for calculating time, money, physical measurements, weights and anything with a numerical quantity.

Limitations of parametric estimation

Parametric estimation's reliance on historic data and measurable statistics can mean that it's optimal for some projects more than for others. Here are some examples of its limitations:

  • Relies on historic data: Without historic data, it may not be realistically possible for a project manager to conduct parametric estimation since available data would come from another team. As a result, this strategy is usually more effective with experienced teams that have already completed projects in the past.

  • Depends on statistics: Parametric estimation typically depends on quantitative milestones and concrete measurements of performance, such as production speed over time. It's, therefore, more useful for projects that have concrete milestones that those without.

  • Can't predict errors: The data-based approach of parametric estimation makes it a reliable measure of performance or costs relating to the past. When it comes to understanding the context of a new brief, such as increased risk or a more complicated process, an experienced professional might forecast delays more successfully.

Tips for improving the accuracy of predictions

There are some strategies you can use to render the predictions you've made through parametric estimations more accurate. Here are some tips for improving accuracy:

  • Calculate on a smaller scale: Breaking down a parameter into smaller increments of time, cost or performance measurement can give you more accurate estimations. It's usually better to make broad predictions from adding the cost or time it took for smaller measurable milestones than from whole projects at once.

  • Use probabilistic estimations: Plotting estimations as a range of results along a probability curve can give you more options for predictions. Using optimistic, realistic and pessimistic results can also better inform investment and development decisions.

  • Use statistical models: If they're available, you can use statistical models in conjunction with the estimations to take advantage of research. This typically requires a lot of experience in statistics and handling data.

  • Use more historic data: You may consult more historic data about performance and calculate averages from it to make a more accurate prediction. It's usually better to look at the average performance of a team than to optimistically consider only its best past performance.

Related:

  • How to write a cost estimation: a step-by-step guide


Explore more articles