A personal blog.

Making Gantt charts work

  • Tue 16 June 2020
  • Blog

Using Gantt charts as a form of project schedule for fixed-price IT projects, I’ve encountered the following problems:

  1. To build a schedule, you have to put all the tasks in a specific order. It’s typically a finish-to-start dependency. However, this is not what happens sometimes. More often than not, it’s possible to start working on the next task even if you are not done with its predecessor (say, it’s blocked). As such, the schedule doesn’t accurately represent what is going on in reality.

  2. When there are many tasks on the bar chart, it becomes challenging to maintain and update the schedule. To add a new task, for example, you will need to weave it into the existing structure of dependencies. It’s a pain in the neck most of the time.

  3. Some tasks are iterative by nature. Where the chart says the task is done once, it is actually iterated several times before we can mark it as completed. As with point #1, we have a discrepancy between the plan and reality.

To solve the challenges listed above, I’ve decoupled the work packages from the tasks/activities. I’ve put on the bar chart only the work packages and built dependencies among them, which is easy to maintain. At the same time, all the tasks have been moved to a separate tracker, where I can add/remove tasks without hassle. In a tool like Smartsheet, the activities from a separate spreadsheet can be linked to corresponding work packages, so I can still monitor the percentage of completion in the main project schedule.

Using the task tracker on regular status calls with engineers is also very handy to track progress. Smartsheet has a Kanban view to visualize the task flow, which is quite cool.

I’ve been experimenting with this approach for some time, and it has drastically simplified project schedule maintenance for me. But I appreciate that in different contexts, with different types of projects, the same approach won’t always work out well. However, I believe building complex network structures and predicting each activity to the day is what drives people crazy using these bar charts and leaves them unsatisfied with the tool. Therefore, keeping the schedule at a high level while handling low-level tasks in a separate tracker should make life easier for many project managers.