How to Setup your Gantt Charts with Realistic Time Estimates

How to Setup your Gantt Charts with Realistic Time Estimates

Written by

Monday, 20 February 2017 03:58

Chances are that at some point in your career you worked on a Project that extended considerably past its deadline... or it was doomed to end that way if a massive amount of extra resources was not added at the last moment.
Keeping deadlines is of fundamental importance and yet, we seem to be so good at missing them. Why does that happen so often?


Estimated Time Required vs Actual Time Taken

While working on improving construction productivity in the 1980s, Glenn Ballard and Greg Howell observed the important limitation of traditional construction work planning and scheduling methods - typically, only around half of the tasks in a weekly plan were completed as planned (Koskela et al. 2014).
A similar study was conducted in Tallinn, Estonia on one of the largest Estonian general contractor’s construction site. The conclusion was that over eleven weeks, on average only 38% of construction tasks were completed on time. This simply means that our project planning process is highly unreliable.
We tend to underestimate the time necessary to perform Tasks.

Perhaps the reason for this behavior can be found in our wish to adapt to the ever growing needs to be faster and more efficient. In order to conform to this need (and please our Bosses and Clients) we do everything we can to squeeze more work out of our resources and we come to a hard stop only when we realize we cannot squeeze Time.

A more advisable approach would be to take time to consider Time properly and make informed decisions so that we can have better chances to finish the works within the deadline we are about to set... after all this is what Gantt charts are meant to do: provide realistic estimates of implementation time of a complex project. 


Slice it first

Humans are really bad at estimating how much time it takes to get tasks done. This is partially due to the fact that most tasks are a complex sequence of smaller activities, We should then refrain from guessing the time it takes to complete the "big picture" task and always start by slicing the work into bite-sized chunks.
In fact, estimating the time it takes for performing just ONE activity, is something we can deal with reasonable precision. Consequently, slicing a task as much as possible offers a better chance for estimating time more accurately. 

The first step to time estimate should then be to split a task into actionable sub-tasks.


Rough Time Estimate

This is a very quick way to estimate a safe time budget: assign a time to each actionable task and then multiply that time by 1.5 times.

This brutally simple approach is highly effective and in many cases will deliver very positive results.
However, in some case you just cannot take all that time and you are forced to squeeze it down further. 
Basically you are asked to take time off your estimate and this increased the risk of failure.

To minimize the risk of underestimating time, you must take more time upfront for planning. You must go deeper into the details of each task and adopt a better way to estimate the time for its completion.


Careful Time Estimate

At this point, the best approach you can adopt is to involve who actually has to perform the task and ask them how much it would take in their opinion.
This has the double advantage of delegating some of your planning work (freeing time for you to do something else) while at the same time getting a second opinion on what you have done so far.

Gathering more opinions about execution times allows using some simple math to come up with better estimates. The theory behind this is the Beta Probability Distribution.

There is no need to go in deep analysis of this kind of model and we can jump right into the estimate by collecting only three numbers:

  • O = optimistic time estimate;
  • M = normal time estimate;
  • P = pessimistic time estimate.

Say that, for each task, you ask three persons to estimate the execution time. This leaves you with four numbers: your first estimate + the new three opinions.
It is easy to see that O is the shortest time proposed and P is the longest.
M can be defined as the average (reasonable) time. A good practice for calculating M is to discard O and P and averaging the remaining values.

Once you have O, M, P for a given task, all you have to do is to put them together and get a new total time:

T = (O + 4M + P) / 6

Here are some numeric examples:

  T1 T2 T3 T4 O P M T
Task 1 3 5 6 8 3 8 5.5 5.5
Task 2 12 18 15 24 12 24 16.5 17
Task 3 16 8 9 12 8 16 10.5 11


T is the time you can confidently use in your Gantt chart.

As you can see the final time T can be quite different from the individual estimates.
This should not pose a problem. On the contrary, it should serve as a practical demonstration that estimating correct times for execution of tasks is a hard endeavor and it is indeed worth to use this kind of statistical approach in order to save our Construction Project from troubles in later stages.  

Sign up

Leansite Mobile App

for Construction Management

is available on

Come get your

free Leansite trial


Published in Masterplan/

Read 952 times Last modified on Tuesday, 21 February 2017 08:01

Andrea Bronzini

Brilliant and committed professional.
Andrea has been working in construction industry since 2004 and he was certified as Passivhaus Consultant in 2010.
Among other skills, Andrea specialises in Sales and Marketing and collaborates with Leansite in quality of Marketing Consultant.