Integrated Master Schedule (IMS)

Understanding the As Late As Possible (ALAP) Scheduling Option in Practical Terms

, , , , , , , , , ,
Understanding the As Late As Possible Scheduling Option in Practical Terms

Many project professionals have spent entire careers without ever using the As Late As Possible (ALAP) scheduling option, although the underlying idea feels familiar. Why? Because it’s very similar to the “just-in-time” concept widely used in manufacturing and logistics.

In materials management, just-in-time means having what you need arrive exactly when you need it, minimizing storage costs and reducing inventory. The same principle can apply to project labor, but with some important cautions.

The “Right Time” for Project Work

On development or design projects, doing work too early can be counterproductive. If designs change, early work may become obsolete, forcing costly rework. The “right time” to perform a task is often determined by schedule logic. In some cases, however, it can also be guided by the ALAP constraint.

Before we explore when ALAP makes sense, let’s quickly review the two primary constraint options in Microsoft Project (and most other scheduling tools).

ASAP – As Soon As Possible

As soon as possible:

  • Is the default setting for forward-scheduled projects (when you set a project start date).
  • Means tasks are pushed as early as possible, immediately after their predecessors finish.
  • Is ideal when you want the earliest possible completion and clear visibility into float/slack.

In an ASAP chain, every task begins at the earliest opportunity, pushing resources as far to the left as possible on the Gantt chart as illustrated in Figure 1.

Figure 1: As Soon As Possible Scheduling Option
Figure 1: As Soon As Possible Scheduling Option

ALAP – As Late As Possible

As late as possible:

  • Means tasks are scheduled as late as possible without delaying the successor or project finish date.
  • Is used in backward-scheduled projects (those planned from a fixed finish date) or when you want to defer work until the last responsible moment.
  • Microsoft Project automatically places each task at the latest feasible start date that still satisfies all constraints.

Switching a chain of tasks to 100% ALAP dramatically shifts all work to the right on the timeline as illustrated in Figure 2. The impact on management is significant: Every task now has zero total slack, which means any delay, even one day, directly delays the project finish. Multiple paths can appear “critical,” making control and reporting more complex.

Figure 2: As Late As Possible Scheduling Option
Figure 2: As Late As Possible Scheduling Option

When ALAP Makes Sense

There are legitimate reasons to use ALAP selectively. For example:

  • When a task consumes resources you don’t want engaged early (e.g., expensive equipment rental or specialized consultants).
  • For just-in-time deliveries or procurements where early completion has no benefit.
  • When modeling backward scheduling. For instance, working from a fixed delivery date toward today.
  • A mixed schedule. Mostly ASAP but with a few ALAP tasks can balance flexibility, cost control, and realism as illustrated in Figure 3.
Figure 3: A Schedule Using ALAP and ASAP
Figure 3: A Schedule Using ALAP and ASAP

A Real-World Example

One of H&A’s senior scheduling consultants once faced this exact dilemma while helping to prepare a multi-year, multi-billion-dollar defense proposal for a project with strict annual funding limits. 

With less than two weeks before the submission deadline, the Proposal Director was exasperated: “I keep asking the engineers what can be delayed! Why does everything have to happen up front? The front-loaded schedule is blowing our funding cap!”

A quick inspection revealed the problem: every task was set to ASAP. The entire effort was jammed toward the beginning of the timeline, creating a massive early demand for resources. After several failed attempts to persuade the engineers to move work later, the consultant proposed something unconventional: “Let’s flip the question. Instead of asking what can we delay, let’s ask what must be done now.”

The H&A scheduling consultant converted the entire schedule to ALAP, instantly shifting all work to the far right of the timeline. The resulting view inverted the problem, from overspending early to under-spending, and gave the team a new way to discuss priorities.

In meetings, engineers were asked to move tasks from ALAP to ASAP one at a time, stopping when the annual funding limit was reached. The discussion changed from “Why can’t we do this now?” to “What can we afford to do this year?”

The result wasn’t elegant, but it solved the immediate problem: the funding limits were clearly observed, the resource profile became manageable, and the trade-offs were visible to everyone.

How ALAP Affects Critical Path and Risk

Because ALAP tasks consume all available float, they appear critical even when they may not truly drive the project finish. This can obscure the actual critical path, making it difficult for project managers to distinguish between genuine schedule risks and artificial ones. In Earned Value Management (EVM) environments, this matters. Earned value metrics depend on knowing which tasks drive completion. Excessive use of ALAP can lead to misleading forecasts and distort DCMA data quality metrics such as the Total Float test and the Critical Path test. For this reason, auditors often recommend using ALAP sparingly and documenting the rationale wherever it’s applied. 

Note: in a sophisticated scheduling environment, it is possible to make a copy of the integrated master schedule (IMS) and revert to ASAP to look for critical paths in the normal sense.  

Combining ALAP with Other Constraints

In practice, project managers often use a blend of constraint types. For example, you can combine ALAP with “Must Finish On” or “Start No Earlier Than” dates to simulate external dependencies such as contract milestones, funding release dates, or material delivery windows. This hybrid approach allows the schedule to model reality while maintaining logical control. However, it’s important to track these constraints carefully. Too many “hard” constraints of any type can reduce the schedule’s dynamic nature and make automated forecasting less accurate.

Guidance from Industry and Agencies

Industry and government scheduling guides consistently advise restraint when using ALAP. The DCMA data quality tests consider the presence of ALAP tasks as a potential red flag because they can mask schedule float and obscure the true drivers of program completion. Similarly, the GAO’s Schedule Assessment Guide recommends minimizing artificial constraints and using logic-driven sequencing whenever possible. ALAP may be appropriate for modeling constrained resources or fixed delivery milestones, but it should always be justified and documented. Within DoD and NASA programs, reviewers often require clear evidence that ALAP usage is intentional, controlled, and limited to well-understood modeling cases. It should never be used as a workaround for poor sequencing.

Key Takeaways

  • ASAP emphasizes early starts, clear float visibility, and traditional forward scheduling.
  • ALAP emphasizes delayed starts, tighter resource control, and is useful in backward or funding-constrained planning.
  • Use ALAP sparingly and intentionally as it can obscure float and create multiple critical paths.
  • In creative problem-solving, toggling between ASAP and ALAP can reveal insights about timing, funding, and necessity that might otherwise remain hidden.

Final Thoughts

The ALAP constraint is a powerful but double-edged tool. It can simplify discussions about funding limits, resource phasing, and timing priorities, but it also carries risk if used indiscriminately. Like most features in commercial off the shelf (COTS) scheduling tools, its value depends on the user’s intent and discipline. The best project schedules blend logic, transparency, and flexibility. Understanding when to use ALAP (and when not to) can make the difference between a reactive plan and a truly managed one.

Interested in Learning How to Use More Advanced Scheduling Techniques?

Master schedulers skilled at asking the right questions to solve project management challenges hone their craft based on years of experience and working with other scheduling experts. There are always opportunities to learn more. H&A routinely offers basic, advanced, and tailored scheduling workshops taught by senior master schedulers with decades of experience in all types of project environments using common scheduling tools such as Microsoft Project and Oracle Primavera P6. Give us a call today to get started. 

Humphreys and Associates also offers basic and advanced EVMS training as well as tailored EVMS training that aligns with a client’s EVM System Description. 

Understanding the As Late As Possible (ALAP) Scheduling Option in Practical Terms Read Post »

Improving Integrated Master Schedule (IMS) Task Duration Estimates

, , , ,
Improving Integrated Master Schedule (IMS) Task Duration Estimates

One of the top reasons projects fail is because of poor task duration estimating for an integrated master schedule (IMS). Without accurate and consistent estimates, project outcomes can become unpredictable, leading to missed deadlines, budget overruns, and overall project failure. A realistic schedule is required to place the necessary resources in the correct timeframe to adequately budget the work as well as to produce credible estimates to complete and to forecast completion dates. While missed deadlines and budget overruns are detrimental for any project, there can be additional business ramifications when producing schedules in an Earned Value Management System (EVMS) contractual environment.

While there are effective methods available to improve task duration estimates, they are often underutilized. A common reason for this oversight is the lack of time allocated to developing the project schedule and determining task durations.

During the proposal phase, initial durations are typically estimated at a more summary level than the detailed execution phase. The proposed work is often defined at a level one to two steps higher than where the actual tasks will be performed. After project initiation, the team’s initial effort is to break the work down into more manageable tasks. This decomposition is crucial for achieving more accurate estimates. It’s no surprise, then, that the initial breakdown efforts often result in duration estimates that don’t align with the proposed durations.

Parkinson’s Law tells us that work expands to fill the time available. If task durations are excessively long, costs will inevitably rise. To counter this, it’s important to require estimators to provide both the estimated effort and the duration needed to accomplish the task. This approach helps to gain a better understanding of the scope of the task and to avoid unrealistic estimates. If you see a task that requires 10,000 hours with a duration of 2 weeks, then you immediately would suspect something is wrong with the estimates.

Techniques for Developing More Accurate Task Duration Estimates

What are your options? H&A earned value consultants and senior master schedulers often employ the following techniques to help a client produce a more realistic IMS.

  1. Establish a Probability Goal. It is essential to set clear expectations for the estimating team. Without guidance, teams may default to estimates with a 50/50 probability of success, which is a recipe for failure. Instead, directing the team to aim for estimates within a 75% to 80% probability range can lead to better outcomes.
  2. Break Down Tasks. Decompose tasks into smaller, more manageable components. The further out the task’s horizon, the greater the variability in estimates. For example, asking someone to estimate the drive time from Washington, DC, to Boston without specifying the vehicle, route, limitations, or conditions introduces unnecessary uncertainty.
  3. Use Professional Judgment. Engage someone with experience in the specific type of work required for the task. A seasoned expert will provide more accurate duration estimates based on their knowledge and experience. Often, we ask the potential task manager to do the estimate, but that person may not be the one with the most related experience or knowledge about the work.
  4. Leverage Historical Data. If the task or a similar one has been done before, use that historical data to inform the estimate. This approach provides a realistic benchmark for future estimates.
  5. Use generative AI. If you have access to an AI capability along with access to historical data, that could be an option to leverage the source data using specific prompts to glean relevant information. As with all AI tools, always verify the generated results to ensure it is a useful basis to substantiate the estimate.
  6. Apply Parametric Estimating. When possible, use parametric analysis to estimate the durations. For example, if it took a specific number of days to clean up a certain amount of toxic waste under similar conditions, this data can be used to estimate the duration of a new but comparable task.
  7. Engage Multiple Estimators. Gathering estimates from more than one person helps to reduce individual biases and provides a more rounded estimate.
  8. Apply the Delphi Method. This technique involves three knowledgeable individuals providing estimates or three-point estimates. The initial estimates are analyzed, and the results are shared with the estimators without attributing specific values to any individual. After discussing the findings, the estimators revise their estimates based on the collective insights, leading to a more refined and accurate duration estimate.
  9. Use Three-Point Estimates. Ask estimators to provide best-case (BC), most likely (ML), and worst-case (WC) durations, along with their reasoning. Applying a formula like the Program Evaluation and Review Technique (PERT) duration formula (1BC+4ML+1WC)/6 can yield an adjusted and realistic estimate. You can vary the best and worst case estimate for risk if you have information on that.

    To see how this simple approach can work, walk through this exercise. Ask yourself how long it takes you to drive to work most of the time. Let’s say the answer is 45 minutes. Then ask yourself how long it would take on a Sunday morning in the summer when the roads were dry (the best case). Let’s say your answer is 25 minutes. Then ask yourself how long it would take on a Monday morning in the winter during a moderate snow event (the worst case). You tell yourself 90 minutes. Now you have enough information to calculate the PERT duration.

    Best Case = 25 minutes
    Most likely = 45 minutes
    Worst Case = 90 minutes
    PERT Duration = (25 + 180 + 90)/6 = 49 minutes

    Finally, let’s say you ask yourself how likely it is that you end up on the high side instead of the low side. If your answer is it is much more likely to encounter conditions that slow you down, you would modify the formula to use one and a half times the worst case (25 + 180 + 135)/6 = 57 minutes. That longer duration shows the impact of your impromptu risk analysis and provides a duration that has a much higher probability of being achievable.

    Now think about the same scenario but conducted by you interviewing three people who drive the same route to work. That would approximate the Delphi method.
  1. All or something less. It may not be necessary to analyze every task to the degree suggested. Even if you could do the analysis along the top several critical paths that would be an improvement. If you were to apply numerical factors to the tasks in related portions of the project that would be impactful. For example, all mechanical design tasks or all software development tasks.

What is the best approach?

You will need to analyze your project and determine which approach or approaches would yield useful information at a reasonable cost. If you apply your own thinking on how to improve your duration estimates, you will undoubtedly find a method most suitable for your situation. Depending on a project’s complexity and risk factors, you may also find it useful to take a more formal approach. Conducting a schedule risk assessment (SRA), a probabilistic assessment of a project’s outcome, can help you gain a better understanding of where the duration risk exists in the schedule.

H&A earned value consultants and scheduling subject matter experts often assist clients to establish basic guidance to help scheduling personnel to get into the habit of adequately defining tasks and using techniques to improve duration estimates. This is critical to be able to produce well-constructed and executable schedules to improve the likelihood of achieving project technical, schedule, and cost objectives.

H&A offers a range of project scheduling training workshops that can help schedulers to implement industry best practices. These workshops also cover how to take the next step to implement advanced scheduling techniques such as schedule risk assessments to ensure the schedule is realistic and achievable. H&A earned value consultants and master schedulers often provide one-on-one mentoring using the scheduling tool of choice to help scheduling personnel work through the learning curve of using advanced network scheduling techniques to produce executable schedules.  

Call us today at (714) 685-1730 to get started.

Improving Integrated Master Schedule (IMS) Task Duration Estimates Read Post »

EVM Consulting – Modeling & Simulation

, , , , , , , , ,

Fighter Jet Air Plain Flying in Front of Moon

Forewarned is Forearmed

Forewarned is forearmed. John Farmer, of New Hampshire, said that in a letter in 1685. But that advice is most likely biblical and very much older. No matter the source of the thought, we should take it as divine guidance if we are project managers. Maybe we should have it cut into a stone tablet, so we can share it with our team members.

Most of our work as project managers is spent in the “controlling” phase which is made up of the three steps “measure, analyze, act.” Our EVMS and IMS exist to be able to support this management function. The measuring part is done very well in our EVMS and our IMS; we know where we are and how we got there. The analyzing is equally well handled in the IMS and EVMS. Only the management task of acting is not well supported. Generally, we lack decision making support and tools.

EVM Consulting - Measure, Analyze, Act

Deterministic Path

No matter how well constructed and how healthy our IMS is, it has a deterministic path forward. The logic links between the activities are there because we expect them to be fulfilled. Indeed, if activity “B” is a finish-to-start successor to activity “A” we fully expect that at some point activity “A” will finish and will provide its output to activity “B”. That is a single path forward and it is a deterministic path. It is also a somewhat simplistic model.

EVM Consulting - Deterministic Relationships in EVMS

Multiple Outcomes

Our management system asks us to perform root cause analysis followed by corrective action. But what if there is more than one corrective action to be taken. And worse; what if the corrective actions can have multiple outcomes with each enjoying its own probability. That means multiple choices and multiple outcomes. How would we show that in our plan? How would we analyze the multiple possible futures that such a situation presents?

Happily, there are ways to model a future without a set path. And once we have the future model, there are also ways to simulate the outcomes to give the probabilities we need to decide which actions to take. We are talking about probabilistic branching, and we are saying that we can build a probabilistic map of the future to use in making decisions; especially making decisions on corrective actions.

Take a simple example of running a test on the project. The expectation is that eventually we will pass the test. We will keep trying until we do. In the IMS deterministic model the test portion of the IMS might look like this:

EVM Consulting - Run the Test then Use the Product

Simulation

We can simulate this situation with different expected durations for the test. That is helpful information, but it does not explain or even capture what is going on in those different durations. It looks like we are just taking longer to do the testing but is that really what is happening? What is going on here? We certainly don’t show that.

In the real world, this simple model might have three potential outcomes. There might be three paths we can take to get to the point where we use the product. Each path has a time and money cost. We might run the test and find that we passed. Or we might have to stop the test for issues on the item or the test setup. We might even fail the test and must correct something about the product to improve our chances of passing a rerun. Eventually we will get to a usable product. But what do we put in our estimate and our plan? What do we tell the resources we need? What do we tell the boss? The customer?

EVM Consulting - Real World Testing

Full Future Model

We now have a much better understanding of the future and can explain the situation. We also can simulate the situation to find out the most likely time and cost outcomes, so we can explain the future without any histrionics or arm waving.
If the issue is important enough we can build out the full future model and simulate it.

EVM Consulting - Full Future Model and Simulation

No matter how far we pursue the model of the future, having a valid model and being able to stand on solid ground are very valuable to us as project managers.

This is not to say that we should model out complex situations as a routine in the IMS. That would be impossible, or at least prohibitively costly. We are saying that when situations arise, we need to be able to use the IMS to help us make decisions.

This type of probabilistic modeling of the future is particularly useful in defining major decision points in our plan. When we reach a decision point the IMS may have multiple branches as successors but that implies we take every branch and that is not valid. Modeling each branch and its probabilities is valid. In the example below, where the milestone represents a decision point, we have shown three possible paths to take. If each were modeled out into the future with time and cost data, we should have the information we need to choose the path we wish to pursue. Without processes and tools like this, we would be flying blind.

Future Blog Posts

This discussion will be continued in future blogs to develop a better foundational understanding of the process and power of probabilistic modeling in our EVMS.

EVM Consulting - Decision Point

Good information sets the stage for good decisions. The IMS and the EVMS have sufficient information to help us model the pathways ahead of our critical decisions. We just need to learn to take advantage of what we have available to us.

Find out how an experienced Humphreys & Associates EVM Consultant can help you create a full future model and simulation of your most vital EVMS Systems. Contact Humphreys & Associates at (714) 685-1730 or email us.

EVM Consulting – Modeling & Simulation Read Post »

Scroll to Top