For some reason, controlling schedules has traditionally focused on controlling start dates instead of controlling the rate of production. As a scheduling tool, CPM promotes this behavior. You will always be better off in the CPM world if you start tasks as soon as possible. I have seen several CPM schedules which were completely misleading because work had started on all the floors at once. The schedule update was showing an early finish. In reality, the same crew was working on small part of the scope of each activity on each floor. CPM used the original planned durations and the series of activities on ten floors which were planned to be sequential and take five days (total of fifty days), were suddenly going to be all finished in five days. This is an extreme example but illustrates how CPM reporting rewards early start even if it does not actually bring any benefits to the project.
In Location-Based Management, start dates are delayed until the crew can work continuously (for tasks which have been decided to be continuous). This maximizes crew productivity. Start dates are also delayed by buffers. It is a good practice in planning to leave a few days or a week of time between two continuous tasks so that delays impacting predecessor do not automatically cause a cascading delay to all the successors. In a schedule without buffers any problem of the predecessor causes a problem to the successor making the project very difficult to control and causing a lot of firefighting and anxiety in projects. While this is easy to plan in theory, it is very difficult to implement during actual production.
Superintendents are not used to making mobilization decisions on the basis of long-term productivity of each operation. Their planning horizon tends to be a few weeks. The focus is on finding locations where work can start and figuring out what can be done NOW to improve progress on site. This also seems to be the natural way to make decisions. I have designed a production control game which all Vico Production Controlling training attendees can play. The game is based on a simple six story building with just a few tasks but realistic situations. Almost all the players remember to add buffers and plan for continuous flow. However, almost all of them “push control” the subcontractors and force them to start according to the schedule even if previous tasks are delayed. Both in reality and in the game they get penalized with slower production, location congestion, too many trades active at the same time, and slow starts by subcontractors.
Caption: The Production Control Game is a virtual workshop for your team to learn schedule optimization techniques and their consequences in a simulation. Played over email, participants react to turns of events at the virtual jobsite. The goal, of course, is to respond to various disruptions and challenges by enacting control actions. The best score is determined by the actual finish date, subcontractor manhours, and number of control actions taken. During the game summary conducted over GoToMeeting, the team finds out what other players did differently and how that affected their scores. The best control actions win the day - and the game!
Early starts have limited or no benefit when the work cannot be carried out continuously. It is only possible to control the production of tasks with long, continuous production runs. Any control action will take time to implement – you need to discuss with the subs, they need to mobilize additional resources, and so on. If you are focusing on 5-day activities instead of a 50-day continuous task your only option is to focus on starts because you are unable to control production rates. This is one of the most important fact of life which causes projects to face continuous firefighting today. It is better to wait before starting the next trade and fix the production problems of the predecessor. You will be able to make better time with this approach – focus on the problem task, not the successor tasks.
When is the right time to start the next task in its first location? All the constraints should have been satisfied. The design related to the first few locations should be available, all RFIs answered, materials ready to be delivered, and equipment available. Predecessors should have completed in the location and cleaned up the location. When all of this has been completed, wait until the predecessor has advanced far enough ahead of the successor. The best way to see this visually is to calculate the number of free locations by looking at the vertical distance between forecast lines. Then it is time to start and with enough work ahead of the subcontractor you will be able to convince them to work fast without being afraid of running out of work.
The figure below shows the status of week 6 of one player in the production control game (note: each group will have different actuals so your game will most likely be completely different). The player made the choice to start Studs on week 7 even though Overhead MEP was not completed. If the production rate of studs is as high as planned this would not make any sense because they would run out of work almost immediately. In any case it would not make sense to start before locations 1 and 2 are finished. This would leave two locations as “space buffer” and give enough time to decrease crew size before conflict if production rate was too high. Figure 2 shows what happened by starting too early – slow down of both MEP and Studs contractor. Starting early can result in finishing late…
Figure 1: The player made the decision to start the following week. Based on forecasts start date should have been delayed by at least 1.5 more weeks.
Figure 2: Too early a start resulted in loss of productivity, a few production alarms, and slowdown of the predecessor
Would you like your team to learn more about Production Control with this on-site game? Just let me know!
And be sure to catch up with the other best practice articles in this series:Part 12: Deliveries and Lay-Down Areas