Project Managers are typically under great pressure to complete the project in as little time as possible. It is therefore imperative that project managers become familiar with the basic techniques to shorten the critical path and the many different ways to do this.
One of the primary benefits of Gantt charts is that they tell you all the activities that are along the critical path (shown in read). These are activities whose duration cannot be delayed without the negative effect of delaying the entire project. There are situations where a critical activity, due perhaps to a constraint, is not on the longest path, and, therefore, will not delay the entire project. Refer to this blog for a discussion on the “The Longest Path” and a more in-depth discussion on these unique critical activities.
The critical path therefore determines the total length of the project. If you can find a way to shorten the critical path your project will end sooner and your project stakeholders will be happier. Hence the effort to shorten the critical path cannot be over emphasized.
When we commence our effort to shorten the critical path we need to keep in focus the impact on the quality, scope, cost and duration of the project. The ideal is for the quality and scope to get better or increase and for the time to decrease. Life, however, in the scheduling world is usually not so ideal. Your critical path shortening will come with trade-offs and you will have to determine both the positive and negative effects on the project. Come along as we discuss critical path shortening techniques and sharpen our skills in this important project management endeavor.
This article discusses several ways to shorten the critical path of a project. It is not, however, an exhaustive treatise on critical path reduction.
The effects of the critical path shortening methods discussed in this article are displayed in Table 1.
Table 1 – Effects of Critical Path Shorten Methods
Note the legend. All these methods are for critical activities. Also, the ‘?’ in the legend means the effect is unknown, and you will need to examine your particular situation to determine the impact. These methods are not necessarily listed in order of preference. Although one, two, and six are definitely favored approaches.
Create Parallel Paths From Sequential Paths (Fast Tracking)
Many times the biggest bang for the buck comes from changing activities in series to activities in parallel. So look for sequential relationships that are not necessary. Many times schedulers will include two tasks in a finish-to-start (FS) sequential relationship because they assume that the successor activity is performed by the same crew working on the predecessor activity. This may in fact not be so, in which case you may have a soft dependency at most. And you can schedule the two activities concurrently, so they are performed in parallel.
Performing soft dependency activities in parallel instead of sequentially can be a huge time saver. This is known as fast tracking the schedule. A simple example, Figure 1, is that install windows and install doors do not need to be sequential, performing them in a parallel start-start (SS) relationship makes sense, Figure 2.
Particularly, if there are, again, separate resource crews for the two tasks and there are no issues with resource crews competing for work space.
Here, quality and scope will not be impacted (possibly cost as well), but the possible time reduction is significant.
Change Sequential Dependencies Into Partial Dependences (Fast Tracking)
Changing a sequential dependency into a partial dependency is only possible if you do not have a hard (“set in stone”) FS relationship between the two activities. In a hard FS relationship you are saying that the successor activity cannot start until the full completion of a predecessor activity. A classic example is your predecessor activity is lay foundation and your successor activity is install wall frames, Figure 3.
You absolutely must be finished with laying the foundation before you can begin installing walls. But what about the activities dig trench and lay pipe for an underground pipe installation. Dig trench is the obvious predecessor and lay pipe is the successor. Now we know that you must complete dig trench before you can complete lay pipe. But must your dig trench activity be completely done before the commencement of your lay pipe activity? The answer is no. So you can change your sequential dependencies into partial dependencies. This, again, is referred to as fast tracking the schedule.
Our recommendation is that you have a start-to-start relationship between the predecessor dig trench and successor lay pipe activities with a positive lag on the successor lay pipe activity, Figure 4.
Here the lag is waiting time (3D – three days) between the start of a predecessor activity and the start of the successor. Using the lag allows you to delay laying the pipe activity several days until there is enough room in the trench to proceed with installing the pipe. Note there is an open network issue with the SS partial dependency, which is discussed here.
Others recommend changing the FS relationship by adding a negative lag. (Another name for negative lag is lead.) So you adjust your FS relationship by adding a negative lag or lead, Figure 5.
Here the negative lag or lead says you can proceed with the successor activity a period of time, say 3 days, before the completion of the predecessor activity. Negative lags are not recommended. And for further discussion on why negative lags are not recommended refer to “The Negatives Of Negative Lags“.
If your absolute preference is for the use of negative lags then use them sparingly. And be forewarned that many government agencies forbid the use of negative lags and, therefore, will not approve your schedule if you include them.
It is generally thought that making sequential dependencies partial dependencies will not impact the quality or scope, but will reduce time. Great! This is what you want.
Adding Resources (Crashing)
Adding additional resources to a task, which is known as crashing the schedule, is a way to cut down on the task’s time without sacrificing scope. This will most likely increase cost, because the time savings of a shortened activity usually does not result in a comparable resource cost savings to offset the added cost of additional resources. So with crashing your time supposedly goes down, but your cost, most likely, will increase.
Others argue that you might not even get the time savings you hoped for. They say this because your additional resources may not be as skilled, requiring training from your more skilled workers. Therefore, there is an initial ramp up time for new and/or unfamiliar workers to become productive. This results in an initial early on slowdown in the hopes of gaining time later on, and that may not happen.
Another important consideration before crashing an activity is work space. Will you have enough work space for all your additional workers to be productive and actually increase crew productivity? If your construction work is on a submarine, the answer may be, obviously, no.
Also consider carefully what is known as the law of diminishing returns when adding resources. This is where adding more and more resources results in less and less impact on the schedule time reduction. There also comes a point where adding more resources may not improve productivity it may actually hinder it. This could be caused by increased communication issues with larger crews or other factors, as working space described above.
Here the time savings is questionable. The effect of using untrained resources may reduce costs but makes the quality questionable. The scope, however, should have no impact. Many times a project manager’s gut reaction is to add resources to compress the schedule, but this just is not the best first option as described above.
What Else To Look For?
Rather than add resources look for these additional time saving possibilities:
Review the estimated durations of critical activities. Perhaps, the scheduler “padded” the estimated activity duration. Padding is where the estimator adds extra time to an activity because they do not have enough information to make an accurate estimate. Decreasing duration estimates benefits the schedule time, but may impact quality and, perhaps, even scope.
Review the estimated durations of lags on the critical path. Again, lags are waiting times between activities, such as the time to let a concrete foundation cure before installing the walls on the foundation. Perhaps, you really do not need one week for your concrete to cure before commencing your successor activity. The benefit is reduced time, but, perhaps, at the expense of quality. Scope is not affected.
Look for extremely long duration critical tasks. These may be split into smaller tasks and some of these resulting smaller tasks may be optimized by performing them in parallel. Ideally, this should not affect quality and scope, but possibly reduce time.
Look at critical path project scope. Are all the deliverables necessary or can deliverables be removed and, therefore, activities eliminated from the critical path? Of course, this reduces scope, and, therefore quality, but the time savings may be significant and of primary importance. Its tough business deciding to cut off tree branches, but if they are dead weight they have to go. And it just might save the tree and/or project.
Shortening the schedule time, one of three variables of the triple constraint of Time-Cost-Scope/Quality, is a major responsibility of the project manager. Yes, time is one of the major variables from which we measure project success. A project manager who is skilled at schedule compression will earn their keep.
Generally, the best way to shorten the schedule is to adjust activities performed in series so that they are performed in parallel. If this cannot be done in whole then it may be good to do this in part.
Making sequential dependencies partial dependencies is therefore also a favorable schedule compression technique. With this goal in mind you may want to look to split particularly long duration activities into several activities. You just might find that some of these shorter activities can be performed in parallel, in whole or in part.
Although crashing the schedule by adding resources is a popular method of schedule compression it just is not one of the best approaches. Costs, most likely, will increase. Quality may deteriorate and you may not even get any time savings. Therefore, we recommend you familiarize yourself with some of the other schedule compression approaches.
For more ideas on “Critical Path Shortening” refer to “Forecast Scheduling With Microsoft Project 2010” by Eric Uyttewaal.