Advertisement

Buffer

  • Mario Vanhoucke
Chapter

Abstract

The weekly Monday morning meeting with Jacob, Mark, and Emily was destined to be exciting but inherently different than the previous meetings about the tennis stadium construction project, not only because of Jacob’s unusual late arrival, but certainly because of the presence of Joanna Barnes. In his weekly GlobalConstruct CEO’s letter to all members of the company, Jacob had already stressed to the team members the importance of the timely delivery of the project, and he had introduced Joanna as a new member of the core team of this challenging project. No one really knew what to expect, but a new challenge was just around the corner. Some things never change.

The weekly Monday morning meeting with Jacob, Mark, and Emily was destined to be exciting but inherently different than the previous meetings about the tennis stadium construction project, not only because of Jacob’s unusual late arrival, but certainly because of the presence of Joanna Barnes. In his weekly GlobalConstruct CEO’s letter to all members of the company, Jacob had already stressed to the team members the importance of the timely delivery of the project, and he had introduced Joanna as a new member of the core team of this challenging project. No one really knew what to expect, but a new challenge was just around the corner. Some things never change.

“The project risk analysis has been unanimously approved by all members of the Australian government team,” is how Jacob opened the meeting, with a shivering in his voice that balanced between excitement and despair. Mark and Emily knew that good news was always followed by decisions that should be made, and so they listened carefully to how Jacob continued his introduction. “While I previously said that manpower to execute the various activities was not an issue for the Australian government, thanks to the almost endless supply of workers, the situation is quite different for our own senior project management consultants, who will have to supervise all the work.”

Jacob introduced Joanna as the head of GlobalConstruct’s Human Resource Management Department. Joanna was not unknown to the other team members, but the fact that Jacob had invited her to their otherwise closed and intimate meetings was unprecedented and raised some suspicion. Joanna was well aware of this—so she immediately took over from Jacob and continued: “During the difficult negotiations with the Australian government, we agreed to deliver nine senior project management consultants from our company for the daily supervision of the project work.” Joanna proceeded to say that GlobalConstruct had to find a way to have access to nine project management consultants with enough expertise for this challenging project, and she argued that everything had to be done to guarantee their availability during the life of the project.

Jacob nodded affirmatively. “It might sound easy, but this resource-management task should be taken seriously—so it’s worthwhile discussing in today’s meeting,” he said in concluding his introduction of the meeting.

Resource Management

Mark hadn’t heard about the project management consultants’ availability problem prior to this morning’s meeting, and so he had prepared an initial project plan that displayed the required number of project management consultants for each individual activity without taking Joanna’s restriction into account. Unaware, until this morning, of the limited availability of nine consultants, he presented the proposed plan to the members of the team (as shown in Figure 4-1) and assured everyone that no significant actions were needed in order to cope with this unexpected limited availability. Although Joanna had little knowledge of project management topics, she knew how to read a Gantt chart—she immediately responded that these proposed plans were unrealistic and not good for the company’s people. As head of the Human Resource Management Department, she immediately saw that the limited availability of the nine senior project management consultants was not always respected, and she drew the team’s attention to the resource over-allocation in the resource profile of Mark’s proposal.

With a voice that revealed the strength of a woman who would fight for the well-being of her people, she started a watertight monologue. “Plans should be realistic!” she said emphatically, and then immediately continued: “and an over-allocation of resources occurs when project managers have been encouraged to meet unreasonable expectations.” She looked at Mark accusingly. “I believe we are right in the middle of such a situation. I do understand that you are pushing the senior project management consultants’ allocation beyond obtainable limits in order to meet constrained schedules and budgets. But I hope you are well aware that an over-allocation puts unreasonable pressure on people and can be costly—not only in overtime money but also in decreasing performance, and possibly in loss of interest and burnout.”

Jacob loved the way Joanna talked about the business and her people, and he appreciated the way she brought excitement, passion, energy, authentic meaning, and joy to her work. Nevertheless, this time he felt it was necessary to interrupt her fear of bad resource management. “Relax,” he said. “I believe that we can easily cope with the resource pressure and solve the over-allocation without shifting the deadline further, which the client wouldn’t accept anyway.” He pointed at the resource profile of Figure 4-1 and looked at Mark. “Can you do something without changing the plan dramatically?” he continued. “I mean . . . just a small update of the plan in order to cope with the limited project management consultants’ availability problem?” An affirmative short nod was the only reply Jacob got. Mark had always been a person of few words—he believed his experience was on his side and that he would solve this issue easily without much disruption to the original plan.
Figure 4-1.

The approved plan and the assigned resources (the senior project management consultants) for the tennis stadium construction project (note the over-allocation of the resources in the resource profile)

Buffer Management

While, in Mark’s view, the resource over-allocation problem could be easily solved by a few minor changes to the plan, Emily kindly interrupted in her own unique, familiar, uncomplicated way: “I have a brilliant idea!” she exclaimed, “Superb and totally new!” Her never-ending search for new challenges, her distaste for the conventional path, and her ability to translate novel concepts into practical experience had often resulted in surprisingly unexpected outcomes, not always leading to better business results . . . but often considered important eye-openers to change direction and explore new paths. The audience could do nothing but listen.

On a recent business trip with her husband and son, Emily had attended a workshop where she had heard about a new upcoming project management technique, heavily inspired by the novel Critical Chain written by Eli Goldratt. Her experimental mentality was notorious, and everybody was aware that one day she would want to implement this often-heard-of technique on one of the company’s projects. Everybody knew that that day had now arrived, and Mark realized that the minor changes to the original plan were surely the easy way to cope with Joanna’s limited resource availability problem . . . but not Emily’s way.

“I’m sure that everybody in this room agrees that the timely delivery of this tennis stadium construction project is essential for the company, and for the well-being of our people, for that matter,” Emily said rhetorically. She didn’t provide any room for a reply, continuing immediately to explain her idea. She explained that a critical chain project plan should provide the timing only for the start of activity chains, rather than for the individual activities. She showed that this meant that, for many activities, approximate start times and estimated durations were used, rather than the strict milestones that typified traditional project plans. Emily argued that the focus should not be on the timely completion of the activities, but rather on the project as a whole. “You shouldn’t forget that the only thing that matters is the overall project performance, and not the performance of the individual activities,” she added as a last undeniable fact, “so that’s why you should give that novel approach a try!”

Having recruited Emily only three years ago, Jacob realized again why he never had any doubts about hiring this young but highly talented woman, and said: “So, why don’t you explain the basic idea to us?”

Emily smiled with a mischievous innocence and said, “It’s simple as tea!” while sipping her coffee and writing only three words on the whiteboard:
  • CUT - PLAN - BUFFER

CUT. Activity durations are often set as low-risk estimates to guarantee timely completion but should ideally be set more strictly to avoid unnecessary delays in activities and lack of focus. “Therefore,” Emily argued, “activity durations should be cut to remove this unnecessary protection. I suggest we cut the activity durations in half (and round up to the nearest integer in case of fractional values).”

PLAN. A project plan must be constructed, similar to the one proposed by Mark, but activities should be scheduled as late as possible within the desired project deadline to postpone important cash outflows. “Obviously, when an over-allocation occurs, we must try everything to resolve it by shifting activities.” She winked at Joanna, and continued, “We don’t want our people to be overworked, do we?”

“That’s not realistic!” Mark argued. “You first cut all activity durations in half, and then, in addition, you schedule them close to the deadline, leaving no slack at all. In my humble opinion, that simply guarantees that you’ll miss the promised project deadline from the moment the slightest thing goes wrong.” Emily smiled with that wiseacre face of hers, which people could only appreciate once they knew her better. “Wait a minute,” she whispered, “because this is where the beauty of the critical chain concept comes in. You should buffer!”

BUFFER. The safety time removed from the activities should be aggregated and moved to strategic points in the project plan in the form of buffers to manage the impact of variation and uncertainty around projects. Emily explained that the sizing and placing of the buffers should be done in a unique way according to the critical chain concept, and she mentioned two types of buffers—project and feeding—as the mechanisms for protecting the project deadline from unexpected delays.

Despite her limited knowledge of project management, Joanna saw Jacob’s approving look, and so she quickly agreed to consider this novel concept as a worthy alternative before further criticizing both Mark’s and Emily’s inability to respect the resource-availability limits. Rumors had been going around that Jacob and Joanna had had a relationship when they were young, but the truth was that they both had been, and still were, passionately engaged with their jobs, and were open to any idea that might satisfy their never-ending hunger for improvement. Backed by Jacob’s affirmative reaction, Joanna said that she was open to analyzing the critical chain planning method as long as not more than nine senior project management consultants would be allocated to the project at all times.

Jacob suggested they explore the usefulness of Emily’s method and compare it with the quick-and-easy solution proposed by Mark. “Convince me of the pros and cons of both approaches,” he said to both Mark and Emily. “Keep me in the loop, and let’s discuss the results in tomorrow’s meeting. Same place. Same time,” he added before he left the room with Joanna, leaving Mark and Emily alone with yet another exciting challenge.

“You would think there’s something more going on between those two,” Mark said. But Emily was already crunching numbers and changing plans. Her never-ending search for improvement made her unpleasant and aloof at times. Still, she knew that there was no time to waste. Tomorrow’s meeting was less than 24 hours away.

Assignment

Given the two different approaches for allocating and planning resources, Jacob suggested that a comparison between Mark’s traditional leveling approach and Emily’s novel buffering approach should enable the company to make a final decision on how to respect the limited availability of the senior project management consultants. Before he left the meeting room, he had told the team that he would forward his action list summary to every member in the room, as he usually did. Much of the work for tomorrow’s meeting now rested in the hands of Mark and Emily, and no intervention was currently required by the other team members, apart from familiarizing themselves with the basic concepts of project resource management and coming back to tomorrow’s follow-up meeting with a critical mindset.

Action List Summary

  • Compare the two planning approaches to solve the resource over-allocation issue:
    • Mark’s leveling approach: Analyze the easy solution to solve the resource over-allocation problem by shifting activities in time.

    • Emily’s buffering approach: Analyze the more advanced critical chain solution using the cut - plan - buffer methodology.

  • Analyze both plans as follows:
    • Identify the major flaws in each approach, and suggest improvements.

    • Calculate the total safety time in the plan for both approaches. Safety is expressed as the total amount of slack in the plan for Mark’s approach, and as the total size of all buffers in Emily’s approach.

    • Make a final decision that is acceptable to all team members and that can possibly be approved by the company’s client.

Uncomfortable Entrance

When Mark entered the meeting room the next morning, he met Jacob and Joanna in a close conversation. Jacob leaned back and laughed loudly, while Joanna sat on the other side of the square table seemingly enjoying their early-morning intimacy. As Mark was not certain he had been noticed, he interrupted clumsily to announce that his colleague Emily had just called him to say that she would be five minutes late.

“No problem,” Jacob replied, “we were just joking about her.”

“Joking about what?” Mark asked, feeling uncomfortable by his unexpected interruption of what seemed to be an intimate conversation between his two colleagues.

“Well, my heart is already in my mouth knowing that Emily undoubtedly has prepared her novel resource planning approach down to the smallest detail.” He laughed again, even more loudly than before. But Joanna noticed the uncomfortable hesitation in Mark’s voice and so, in a friendly voice, asked him whether he had been able to alter the plan to resolve the resource over-allocation.

“I have indeed an alternative plan to discuss, but it might not be as exciting and novel as what Emily will propose,” he answered. “I don’t think it will be an easy meeting today.”

Jacob replied that making difficult decisions was part of GlobalConstruct’s DNA, and he told Joanna that the main reasons why he had chosen the job as CEO were to satisfy his desire to act as a facilitator and problem-solver and to feed his ambition to find solutions for complex problems in collaboration with a small but excellent team.

“No worries, a problem is just a solution in disguise,” Jacob said enthusiastically. “We’ll figure it out.”

“That’s right,” Mark said, now feeling somewhat more relaxed.

“Ha, here you are, Emily!” Jacob said. “Let’s start our meeting—there are tough decisions ahead,” he joked.

Leveling (Mark’s approach)

Mark quickly forgot his uncomfortable entrance and immediately put two plans on the table, printed on large sheets to reveal all the tiny little details. The first sheet was a copy of the approved plan presented yesterday (Figure 4-1), which, despite its approved status, clearly suffered from an over-allocation for the senior project management consultants. Based on the leveling approach he had proposed the day before to solve the over-allocation issue, he now brought an alternative plan that closely resembled the approved plan, but with some activities shifted further in time (to the right in the Gantt chart) to get rid of the resource problems (Figure 4-2).

“The basic idea of my new plan is easy,” he said. “What can’t be done in parallel should be done in sequence.”

Mark reminded the team that the initial approved plan showed an over-allocation for the senior project management consultants, and pointed at the resource profile graph of Figure 4-1. He showed the team that the over-allocation started at week 17 with two extra project management consultants required for four weeks (above the limit of the nine consultants available), and a second period with an over-allocation occurred starting at week 29, where up to 13 senior project management consultants would be required (i.e., four more than currently available) for another four weeks.
Figure 4-2.

Mark’s leveling approach: An alternative project proposal without any resource over-allocation

“We could simply accept the over-allocation, and let the project management consultants work overtime,” Mark said, “but I know that Joanna is not very keen on accepting overtime during the planning phase.”

“I don’t have anything against overtime work,” she replied, “but I don’t think we should build it into our plans.” Joanna told the team that planned overtime often had a negative impact on a project’s health and on how project management consultant teams worked, and she reminded everyone that the project might simply stall, come to a halt, or even fail completely. “Working overtime is occasionally acceptable to solve acute project problems, but it should be avoided during planning as a work-around to solve a scarcity of resources.”

“That’s why I have the alternative plan ready for approval,” Mark said. Mark showed that delaying activities 4 (“filling playing field/track”), 14 (“painting”), and 15 (“lights”) would lead to a more leveled use of the project management consultants, thereby resolving all over-allocations and issues raised by Joanna. He pointed to the arcs in the Gantt chart and to the resource profile of the alternative plan and showed that, thanks to these activity shifts, the availability of nine senior project management consultants would never be exceeded during the project’s life.

“An easy solution,” Jacob responded, “but since this requires a change in the plan, it should be approved by our client.”

“Exactly,” Mark said, “but since the project deadline of 48 weeks remains intact, I don’t see any reason why our client would object to this changed plan.”

Jacob told the team that the client was probably well aware that they should accept some degree of change when working on any project, but he was not so sure it was wise to propose a change request so early in the project phase. Although he agreed with both Mark and Joanna that an over-allocation should be avoided during the planning phase, he nevertheless clearly stressed that a change in plan was not a desirable move in this early phase of the project. He looked at Mark and thanked him for the easy solution to cope with the over-allocation of the senior project management consultants.

“Since your alternative plan does not differ too much from the initial approved plan, I prefer that we stick to the approved version of the plan,” Jacob said.

He agreed that building in overtime during planning was indeed not a very desirable approach, as Joanna had previously said, but Mark’s alternative proposal clearly illustrated that the allocated overtime could be resolved easily. Therefore, he preferred to stick to the initial plan, despite its overtime, and to decide on some activity shifts to resolve the over-allocation during the project’s progress.

“When we actually detect a real over-allocation during the progress of the project, we now know that simple activity shifts will solve these resource problems very easily,” he said.

The team agreed with Jacob that making these small changes to the initial approved plan was probably not worth the effort, as it would lead to unnecessary discussions with the company client. However, before Jacob would make a final decision, he proposed taking a look at Emily’s alternative approach, which could provide a completely new insight into how to cope with the over-allocation of resources.

“Food for thought,” Jacob summarized to end the discussion. “Emily, you had an alternative view on this planning process, so why don’t you tell us about it?”

Buffering (Emily’s approach)

Emily hadn’t said a word during the meeting, since she was eager to present the novel approach to the team. Although she realized that Mark’s quick and easy solution was elegant in its simplicity, she preferred to present an alternative approach—a more complex one and hence, in her opinion, more promising. Emily reminded the team that she had explained the cut - plan - buffer approach in the previous meeting, and she now wanted to illustrate this three-step approach using the tennis stadium construction project data. She had prepared some calculations, and with a pen and cloth in her hands, she was ready for some drawings on the whiteboard. Everybody realized that, once again, this would be the starting shot for a teaching session beyond the normal practice of project management for GlobalConstruct.

“Are you ready for a new approach?” she asked to her team mates.

“We certainly are!” Everybody responded in choir. Was there any other choice?

Emily asked the team to take a quick look at Mark’s alternative proposal and reminded them that this resource-leveling approach simply required delaying activities until resources were available. Although she praised this approach for its simplicity, she warned the team that it didn’t leave much room for activity slack in the schedule. Moreover, she referred to Mark’s new plan as a milestone plan to stress that the planned activity times typically marked specific points along a project timeline. She explained that a milestone planning approach gave little to no incentive to finish activities earlier than the planned finishing times and should therefore be handled with care.

“The problem with milestones,” Emily continued, “is that they quickly become self-fulfilling prophecies.”

Jacob, known for his passion for history and a notorious bookworm, quickly reacted by reminding the team that a former British naval historian named Cyril Northcote Parkinson had written a book about this phenomenon, which since then had been used in lectures on efficient time management. Emily smiled affirmatively and immediately wrote the following management law on the whiteboard:
  • Parkinson’s law!

  • “Work expands to fill the time available for its completion.”

  • “What’s the problem?” Joanna reacted. “As long as correct time estimates for the project activities have been made?”

  • “That’s exactly the problem,” Emily replied. “Management laws always come in pairs.”

  • Just below the first law, she wrote a second famous law on the whiteboard:

    Murphy’s law!

  • “Anything that can go wrong, will go wrong.”

She explained that Parkinson’s law, in combination with Murphy’s law, would cause delays in the activity milestones, which undoubtedly would have an impact on other activities and the team working on the project, and eventually on the total project duration.

Emily explained that, when the activity finish times act as milestone targets, most people will not begin to work on the activities immediately, but will instead start at the last possible moment. This eliminates any potential safety margins and puts people under stress and pressure. She said that this phenomenon is known as the student syndrome —but she wondered whether this was only applicable to students. Everybody smiled, because they recognized this practice in their daily life. She told the team that Parkinson’s law often occurred due to this student syndrome, and that the law consumed all safety time for each project activity, which eventually led to delays due to the inevitable intervention of Mr. Murphy.

“That’s why the slightest problem will cause a delay and disrupt the milestone planning,” Emily said.

“You should take that knowledge into account by allocating extra time for activities,” Joanna replied.

“Your reaction goes right to the heart of the matter!” Emily replied thankfully. “You should not!”

Everybody saw that Emily was delighted by the surprised looks on her team member’s faces, as she was well-known for her habit of explaining new matters that often went straight against normal business practice. She told the team that her amazingly counter-intuitive reaction lay at the core of the cut - plan - buffer approach, which she was now about to explain using the tennis stadium construction project data.

CUT. Emily introduced the concept of aggressive time estimates , which relied on the crucial insight that projects as a whole—rather than the individual project activities—should be protected against unexpected delays. Since activity estimates usually contained some degree of protection to anticipate the effect of the student syndrome and Parkinson’s and Murphy’s laws, people often recommended that you remove this safety time from these activity estimates. She looked at Joanna and told her that this was exactly why the allocation of extra time to activities was not the way to avoid these potential activity delays, while she wrote the following words on the whiteboard:
  • NEW! USE AGGRESSIVE TIME ESTIMATES

“The protection of activities against delays should not be a goal,” she said. “We should protect the project!”

Joanna nodded silently, recognizing that Emily was a master at constructing logic from scratch—often starting with the wrong assumptions her team members had made and gradually guiding them into new, unexplored directions. Still, she was not entirely convinced that removing protection from activity time allocations was the right thing to do. Nevertheless, she decided to continue listening to Emily’s reasoning before raising any objection.

“As an example, let’s cut the activity duration in half to get rid of the unnecessary safety time, and round the duration to the nearest integer for simplicity’s sake,” Emily suggested.

Before giving her team the chance to react, Emily continued: “I know that this is a very aggressive (no pun intended) approach—having activities with much shorter durations and no protection at all—but let’s look at the second phase of the new approach.”

PLAN. Emily continued in a stepwise fashion, explaining that in a second phase the project should be planned with these aggressive time estimates while avoiding any over-allocation of resources similar to Mark’s approach. Much against everyone’s expectations, she advised the team to plan the work backward from the expected project completion date, so that each activity would start as late as possible. Emily carefully drew a Gantt chart on the whiteboard (Figure 4-3) and reminded the team that the activity time estimates had now been cut in half to obtain the aggressive estimates. She especially drew their attention to the fact that this plan, despite a much shorter activity duration, closely resembled the new plan for approval proposed in Figure 4-2 by Mark.

“Just as Mark did for his new alternative plan proposal, I have scheduled some activities in a sequence to avoid any resource over-allocation. Moreover, I have also left a gap between activities 3 and 4 for the very same reason.”

“As far as I understand, the only difference with my approach is that the activity durations are much shorter, and that these activities are now scheduled as late as possible,” Mark added.

“That’s correct,” Emily replied.

“Then I don’t see any added value compared to my simple planning approach,” Mark concluded, although he realized that Emily had not yet finished her explanation.
Figure 4-3.

Latest start schedule with aggressive time estimates after leveling

Before she gave a reasonable answer to Mark’s reservation, Emily started to add some notes on her drawing of Figure 4-3. She explained that the longest path in this resource-leveled network had to be equal to the total planned project duration, and she noted that this longest path was very much similar to the critical path, although there was one major difference. “While the critical path consists of a sequence of activities linked only by precedence relations between them, the longest path on the resource-leveled schedule can also consist of links due to limited resource availability,” she said. She referred to this sequence of activities as chain 1 - 2 - 6 - 7 - 8 - 9 - 10 - 16 and called this path the critical chain . She told the team that this was a new name for the old “critical path” concept, but now extended to resolve resource over-allocations.

“In this case, the critical chain is identical to the critical path,” Emily explained, “but they could have been totally different due to the fact that some activities have been shifted to resolve resource over-allocations.”

Next to her Gantt chart, she wrote the following words on the whiteboard:
  • CRITICAL CHAIN: CRITICAL PATH AFTER RESOURCE LEVELING

“Since this critical chain is now the longest path in the project, it determines the planned project duration of 25 weeks,” Emily said. “But obviously, since we rely on aggressive time estimates, this new plan contains no safety time whatsoever,” she added as a comment.

Before going more deeply into this lack of safety time, she immediately continued by pointing to the other activities in the Gantt chart that were not part of this so-called critical chain. She told her team members that these activities could be considered as non-critical activities, and since they didn’t lie on the critical chain, they should be part of another chain of activities that merged somewhere in the critical chain at one point or another. Then, without any hesitation, she wrote the following words on the whiteboard:
  • FEEDING CHAIN: ANY PATH FEEDING THE CRITICAL CHAIN

Apparently, the non-critical activities were part of so-called feeding chains, which entered (or fed) the critical chain at some point in the project. Emily explained that a first feeding chain consisted of activities 11 - 12 - 13 - 14 - 15, which connected activity 15 (end-activity of this feeding chain) with activity 16 (an activity on the critical chain). She drew special attention to a second feeding chain consisting of activities 3 - 4 - 5, which technically did not merge into the critical chain (since activity 5 was a project end activity), but merged into the final project completion date. Then, she returned to Mark, who had just claimed that this approach was very similar to his approach, apart from some tiny differences.

She admitted that her novel approach was not fundamentally different from Mark’s approach, as it also aimed at resolving resource over-allocations by shifting activities in time—although now the shifts were made to the left in the Gantt chart (backward), while Mark’s approach invoked activity shifts to the right. She reminded the team that the resulting schedule had a critical path, which was now called the critical chain, since it contained a sequence of activities with shifts in some of these activities to resolve the unavailability of resources. However, Emily immediately recognized that the major drawback of her approach so far was that the activity durations were set to very aggressive values—and therefore, the total project duration of 25 weeks could not be considered a realistic estimate, and could therefore not be compared with Mark’s much more realistic project duration estimate of 48 weeks. Despite these weaknesses, she ended her explanation by stating that the critical/feeding chains concept was crucial for understanding the next phase of the cut - plan - buffer approach. She then suddenly fell silent to stimulate reaction from the team.

Still searching for a sound motivation for removing the activity protection (due to the reduction of activity durations to obtain aggressive activity times), Joanna reacted quickly: “What about Murphy’s law now?” she asked with an irritated tone to her voice. “With these reduced activity times, I believe that the chance that we have activity delays is now even bigger than before,” she almost shouted.

“That’s what I told Emily in the previous meeting when she initially introduced the new idea,” Mark replied, “but we know that Emily has a solution for every problem, don’t we?”

“I understand your point, Joanna,” Emily commented. “With aggressive time estimates, a latest start plan without any degree of slack, and Mr. Murphy’s law, we’ll end up with a perfect cocktail for project delays and unhappy clients!”

“That’s why I said that we should add more safety time to the activities, instead of reducing the allocated time,” Joanna said, but she was beginning to understand where Emily was heading to.

“Indeed,” Emily replied. “We will now go to the third phase; we’ll add buffers. We will indeed add safety time to catch up delays. But we won’t add it to the activity estimates—we’ll add it to other places in the project schedule instead.”

It became clear that Emily wanted to use the previously mentioned weaknesses and flaws of her new planning approach to show the importance of buffering and to actually improve the overall project plan.

BUFFER. Obviously, Emily was well aware that the reduction of the activity time estimates to aggressive durations removed all protection and reduced the project plan to a schedule very prone to delays. Therefore, she introduced the general idea of project buffering as a way to re-introduce the removed activity safety time back into the project plan at certain predefined places in order to protect the project against delays.

“It’s a matter of giving and taking,” she said.

Mark and Joanna stared at Emily as if that last statement were Emily’s way of pleading guilty to having ignored Joanna when she’d said that the aggressive time estimates approach was too prone to project delays. But Joanna was starting to appreciate the whole idea of project buffering, and so she said to Emily: “You took the safety time from the activities, and you will now put it back at other places in the project network, right?”

“Exactly.” Emily replied. “And part of the removed activity time is added back into the schedule by using two types of buffers.”

First came the so-called project buffer concept. Emily explained that a single project buffer should be added at the end of the critical chain to protect the project completion date from delays. Any delays in critical chain activities would be absorbed by this buffer without having an effect on the project completion date, and this protection would last until the project buffer was consumed. She continued by telling her team that a project buffer should protect the project completion date against changes in the activity durations along the critical chain, so it would not serve as a protection for delays in the other activities on the feeding chains. Meanwhile, she wrote the following words on the whiteboard:
  • PROJECT BUFFER.

  • A single buffer at the end of the critical chain to protect the project completion date

Then came the so-called feeding buffer concept. Once Emily got going, nobody could stop her. She continued by saying that the project buffer was inserted in the project schedule to protect the project deadline against changes in the critical chain, but she warned the team that the critical chain should also be protected against changes in any feeding chain. She explained the important difference between the single project buffer and multiple feeding buffers, and said that this was the beauty and elegance of the new planning approach.

“If delays occur in critical chain activities, the project buffer will be consumed,” Emily argued. “However, if delays occur in activities on the feeding chains, the project buffer won’t be consumed at all, but instead the feeding buffers will absorb these delays.”

Consequently, feeding buffers at the end of any feeding chain should be added to the project baseline schedule as a way of protecting against delays in the feeding chains. She took her pen and wrote:
  • FEEDING BUFFERS.

  • Multiple buffers at the end of the feeding chains to protect the critical chain

It was clear to all team members that buffers would be used as a catch-up mechanism for activity duration variability by adding safety time to the various chains in the project plan. When activities suffered from delays, the other succeeding activities on the same chain would start later than expected, and this late start would be absorbed by part of the buffer. Emily told the team that activities no longer had strict starting times (as was the case in a traditional milestone plan), and so delays were now acceptable as long as the buffer was not completely consumed. However, to guarantee the timely completion of the project, these buffers should be sized according to the properties of the chains they protected, such that the buffers would not likely be consumed. Ideally, the size of these buffers would depend on the expected variability of the activities on the chain where the buffer was placed (which is the critical chain for the project buffer and the feeding chains for the feeding buffers). The impact that possible activity delays might have on the critical or feeding chains depended on many factors that were often well beyond the company’s knowledge and awareness.

“I believe it is a general rule of thumb that the buffers for activities with high risk should be sized appropriately, while low-risk activities do not require a big buffer. For now, I suggest that we take, just for illustrative purposes, a buffer size equal to 50 percent of the length of the chain,” Emily said, always seeking to simplify complex topics, as any good teacher should do.

She erased the text on the whiteboard and started drawing some Gantt charts to explain her idea.

“The basic idea of the project and feeding buffering approach is easy, but the way it is done is sometimes tricky and often requires pragmatic solutions,” Emily said. Meanwhile, she continued drawing three different Gantt charts on the whiteboard as a foundation for three different proposals for sizing the buffers (as shown in Figure 4-4), which she wanted to discuss with her team members.
Figure 4-4.

Emily’s buffering approach: three alternative project proposals without any resource over-allocation

Proposal 1. 50% buffer. In her first proposal, Emily told the team that she worked backward, adding a project buffer at the end of the critical chain and feeding buffers at the end of each feeding chain, ensuring that their size was equal to 50 percent of the length of the chains (rounded up to the nearest whole week).

“It’s an easy and quick approach,” she said, “but one not free of problems.”

She had previously calculated the length of the critical chain to be 25 weeks, which meant that a buffer of 13 weeks should be enough to absorb all unexpected delays. Similarly, the feeding chain consisting of activities 3, 4, and 5 had a total length of 17 weeks, and so it received a buffer of 9 weeks. She referred to the top picture of Figure 4-4 and added the buffers to her drawings.

“The problem is the feeding chain with activities 11, 12, 13, 14, and 15,” Emily confessed.

“The buffer size should be 7 weeks, which is 50 percent of 13 weeks, rounded up,” Joanna said.

“Exactly,” Emily replied, “but I added a buffer of only two weeks.”

She told Joanna that inserting buffers forced the corresponding chain to move to the left of the Gantt chart, but since the feeding chain could not be shifted more than two weeks, the chain could not be buffered properly.

“The predecessor of activity 11 (the first activity of the feeding chain) is activity 7 of the critical chain, and therefore the shift to the left is restricted to a maximum of two weeks,” she explained.

“In that case, the second feeding buffer is not big enough to absorb the potential delays of the feeding chain,” Mark added.

“Exactly. This means that the feeding chain is not properly protected, and any delay in this feeding chain might consume the small buffer too quickly,” Emily said.

She explained that consuming this feeding buffer too quickly would lead to a delay in the activities on the critical chain, thus eating up the project buffer as well. Consequently, the project buffer would be consumed as a result of delays in the feeding chains!

“This is against the concept of buffering, since the project buffer is sized to capture delays only in the critical chain, and not in the feeding chains!”

“So, buffering is not a good idea after all?” Joanna asked, somewhat confused.

“There’s a solution,” Emily replied.

“There always is,” Jacob chimed in. Everybody was surprised that he was still listening, because it had seemed that he left the discussion a while ago to work on his laptop answering emails.

Proposal 2. Split project buffer . In a second proposal, Emily had split the project buffer into two parts, still guaranteeing that its total size of 13 weeks remained intact, as that was the right size for the expected delays in the critical chain. However, splitting the buffer into two parts allowed her to increase the size of the second feeding buffer from two weeks to the appropriate seven weeks, long enough to absorb possible feeding chain delays.

“The critical chain now includes two separate parts of the same project buffer (PB(a) and PB(b)) and is equal to the line 1 - 2 - 6 - 7 - 8 - 9 - 10 - PB(a) - 16 - PB(b) on the middle picture of Figure 4-4,” Emily said.

She explained that the first part of the buffer (PB(a)) was set to five weeks by delaying the last activity on the critical chain, and this shift released exactly enough time to increase the feeding buffer from two weeks to the desired seven weeks for feeding chain 11 - 12 - 13 - 14 - 15.

“The feeding chain now has a seven-week buffer,” Mark remarked.

“I get it; all buffers are now large enough to absorb possible delays, assuming that a 50 percent buffering approach is taken,” Joanna confirmed.

“Indeed, and as you can see, the total project duration is now equal to 38 weeks, exactly the same as in my first proposal. This is still ten weeks less than Mark’s proposal, but now with the appropriate buffer sizes.”

“Is this your final proposal?” Jacob asked.

“Let me add one tiny little improvement,” Emily said.

Proposal 3. Total buffer ≈ removed safety time. In a third proposal, Emily had made some minor adjustments to the previous Gantt chart by delaying the last project activity (16) even further, now starting at week 32 instead of 29, which enabled her to increase the buffer of the second feeding chain by three weeks, and led to an increase in the other buffers as well. This shift increased the planned project duration from 38 to 48 weeks, exactly the same as Mark’s alternative proposal, to allow a fair comparison between the two schedules.

“Recall that I removed all safety from the activity durations to get aggressive time estimates,” she said.

“Indeed, you reduced all activities by 50 percent,” Joanna recalled.

“The total amount of time that you removed from the activities equaled 52 weeks. I did the math,” Mark replied (this number is equal to the total difference between Mark’s normal activity times and Emily’s aggressive activity times).

Emily told the team that increasing the planned project completion time to 48 weeks, and the shift in only the last activity, resulted in a total (project + feeding) buffer size of exactly 52 weeks.

“You gave us back exactly what you previously took away,” Joanna replied.

“I told you it’s a matter of giving and taking!” Emily said.

It was typical of Emily to bring the discussion to a point where a fair comparison between two alternative approaches—Mark’s traditional approach and her much more radical approach—could be made effortlessly.

“Time for an evaluation,” Jacob said.

He closed his laptop and went straight to the whiteboard, asking the team to compare Emily’s buffered proposals with Mark’s proposal.

Review

Emily started the evaluation by admitting that the buffering approach was meant to be a proposal to discover other possible planning approaches different from the usual business practices, but she stressed that she had no intention of forcing a new approach on the company without spending time critically evaluating its merits and its pitfalls. She realized that implementing this approach would require a fundamental change in mind-set because it entailed convincing people to change their activity-estimating process to an aggressive approach.

“Honestly, I don’t think we are ready for this today, but I believe it’s worth exploring the idea for future projects,” she said.

“I believe that the appropriate sizing of the buffers is crucial to the success or failure of the approach,” Mark added.

“Exactly,” Emily replied. “Sizing buffers at 50 percent of the chain length is a rule of thumb without any sound reasoning.”

“We could easily take other factors into account that might cause delays when sizing the buffers,” Mark added.

“Such as?” Jacob asked.

“The project network structure, the scarce resource availability, the accuracy of the time estimates, the project’s inherent risk, and probably many more,” Mark replied.

“The general idea is that buffer sizes should be enlarged as the potential delays in the chain become bigger,” Emily summarized.

Everybody nodded.

With the promise that she wouldn’t go into the details, Emily said that the previously calculated sensitivity metrics they had proposed in their risk management meeting (the criticality index, significance index, and schedule sensitivity index) could be used to predict the size of the potential delays in the critical chain and feeding chains. “It is my best guess that these sensitivity metrics can be used to size the buffers appropriately according to our knowledge of risk,” she said.

“Food for thought,” Jacob said, “but not for this project.”

“I believe there is a second reason why we should not accept this new approach overnight,” Joanna added. “Since Emily’s proposed plan no longer relies on fixed activity milestones, but rather on flexible activity start times with extra buffers to absorb potential delays, it won’t be very easy to communicate the plan to the client and potential sub-contractors working with our project management consultants.”

“Indeed,” Emily said, “how do you tell a sub-contractor that the starting time for his work is yet to be decided?”

Everybody laughed, recognizing that this would certainly not be the best way to build up a good relationship with GlobalConstruct’s sub-contractors.

“It requires a certain degree of flexibility from everyone, or even a totally different mindset,” Emily added.

“That’s difficult, but not impossible,” Jacob once again concluded, “but certainly not appropriate now at the start of such an important project as our tennis stadium construction project.”

“If the time is not right for accepting this novel idea, despite its promising features, which of the two alternative plans should we select then as the most appropriate for moving ahead?” Joanna asked.

Everybody stared at Jacob. He was the one to make the final decision.

Decision Review

“Before we finish this meeting, let’s briefly go over what we have discussed,” Jacob said.

He praised the new buffering technique as an important eye-opener with promising features that should be further refined for future company projects, and he thanked Emily for her clear explanation of the general idea. Then, he expressed his concerns about changing the approved plan (Figure 4-1) to Mark’s new proposal (Figure 4-2).

“Mark, I am well aware that your solution solves the resource over-allocation issue,” Jacob said, “but a shortage of project management consultants is an internal issue that we should solve when the project is in progress.”

“I understand,” Mark replied. “It’s not mandatory to change the plan, as long as we are aware of the over-allocation of resources.”

“My suggestion is to keep the original plan as it is, despite some resource over-allocations,” Jacob said. “Any objections?”

“It was a very interesting meeting with relevant new concepts that will have an impact on the way I look at project management,” Joanna replied. “I wasn’t a huge fan of allowing over-allocations in the plan, but I guess now that it’s the best alternative and we can change it later as the project progresses . . . I have no objections.”

The others confirmed by nodding to Jacob.

“No doubt we’re an excellent team,” Jacob said, “and it reminds me of a quote I once heard:

There are three solutions to every problem: accept it, change it, or leave it.

If you can’t accept it, change it.

If you can’t change it, leave it.

Yes, we did consider alternatives. We even seriously thought about implementing a change in the plan, and we evaluated two different alternatives. One was easy, with small changes (Mark’s approach), the other was totally new and implied a radical change (Emily’s approach), but we didn’t accept either of them. I call that courage! We didn’t change, we left it as it was. But at least we thought about it!”

“That’s the power of innovation: think outside of the box, then think twice, but don’t rush and change things too fast,” Emily said.

“I think we all agree that we should leave the plan as it is, and start this challenging project!” Mark said proudly.

“Thank you all for coming today—and we’ll see each other in a few weeks time,” Jacob said, concluding the meeting.

Copyright information

© Mario Vanhoucke 2018

Authors and Affiliations

  • Mario Vanhoucke
    • 1
  1. 1.GentBelgium

Personalised recommendations