Most guidelines allow lag in scheduling. The caveat is the lag must be properly applied as its incorrect use may overstate risk in the schedule.
Negative lag is considered never appropriate. Limited use of positive lag is agreeable to most scheduling guidelines, but, again, it must be properly applied. A common misuse of lag is to target a specific date in the schedule. This error causes issues, in particular, for Monte Carlo simulation, which attempts to account for a schedule’s probability distribution. Lags often overstate the risk in a Monte Carlo simulation.
This article demonstrates improper use of lag and how to correct this lag scheduling problem.
We start our discussion with a review of the precedence diagram. The cause and effect relationship between activities is delineated in the precedence diagram, Figure 1.
Figure 1
The precedence diagram defines the predecessor as the cause and the successor as the effect. In this way the scheduler can model the interface between activities. Acceptable precedence relationships include finish-to-start (FS), start-to-start (SS), and finish-to-finish (FF). These relationships help the scheduler model the true cause and effect relationships between activities.
It may also be helpful to define a waiting time between activities. This waiting time is referred to as lag. A FS relationship modified by a positive lag, Figure 2, says that in addition to predecessor A completing you must wait a defined period of time before implementing successor B.
Figure 2
Lags are strictly for a fixed period of time that takes no resources and will not change. It’s the unchanging nature of lag that can make it problematic in scheduling.
Lag works well to model wait time for processes, such as the curing of concrete or the drying of paint. Yes, lag has it place in the echelon of scheduling tools. But many times schedulers use lags in ways that are not appropriate.
The most common misstep is to use long lags to put a successor activity on a specific date. This is referred to as target date scheduling. The scheduler unrealistically inserts a lag to achieve a preferred start date for a successor activity. This lag usage may be indicative of poor or incomplete logic. Improper lag insertions have a way of making the so-called schedule little more than a glorified calendar.
Figure 3 displays an improper insertion of lag.
Figure 3
In this schedule we want the pre-construction kick-off meeting to take place on 12-June-2019. To achieve this date we insert a lag between start project and preconstruction kick-off meeting.
What takes place during this waiting period? Our schedule is not very transparent, which affects its overall quality. Worse still if the start project milestone delays for any reason the lag between start project and preconstruction kick-off meeting will not adjust, accordingly. This is a problem. Any updates in the predecessor activity will make preconstruction kick-off meeting miss our planned 12-June-2019 date.
Our jerry-rigged schedule lag makes the schedule both less transparent and static. Activity updates will require manual adjustments to the lag, which is not practical. You want the successor start dates to automatically adjust according to changes in the predecessor activities.
In Figure 4 we have a corrected schedule that is more transparent and dynamic.
Figure 4
In this schedule we replace the lag with activities and logic. Three activities take place in this apparent void between the start project milestone and preconstruction kick-off meeting. You have the review and approval of the safety plan in parallel with the review and approval of the environmental protection plan. Approvals granted, the project moves on to ordering material and equipment. The durations of the approvals and ordering material and equipment combine to make preconstruction kick-off meeting take place on 12-June-2019 as desired.
Our improved schedule tells the true story of the schedule. Any delays to approval safety plan or approval environmental protection plan, and we know we must shorten the duration of ordering material and equipment to place our kick-off meeting on the preferred 12-June-2019 date. With the lag included in the schedule, any delay in start project causes the preconstruction kick-off meeting to over shoot its planned date, which is not good. This also has a negative effect on the accuracy of Monte Carlo simulation.
Summary
Finding or defining predecessors is preferable to insertion of long lags to place activities on specific dates. Lags are less transparent and inhibit the desired dynamic nature of a schedule. They also overstate risk in the Monte Carlo simulation. In this simulation activity durations change many, possible thousands of, times and the successor date will be ‘incorrect’ in most of these iterations.
For these reasons consider replacing lag in the schedule with defined activities that have estimated durations, and good logic. The introduction of activities allows the examination of risk, and the Monte Carlo simulation will be correct.