Total Floats: with or without project deadlines.

Total Floats: with or without project deadlines.

One of the most important outputs of the Critical Path Method (CPM) is a calculation of activities Total Floats (TF). However, there is no commonly agreed approach to calculate TF when the project has deadlines. Different project management tools have different features to manage deadlines and the ways they calculate TF.
Should the deadlines be a part of the calculation or not, and why?

Spider Project provides both perspectives simultaneously. Primavera (P6) and Microsoft Projects users have to choose which type of calculation they prefer. For effective project optimisation and control, it is very important to be aware of both perspectives. This post explains how scheduling tools calculate TF with deadlines and how additional Critical Path  Metric: “TFD: Total Float Deadline” could be developed in P6 and MSP.

Critical paths without project deadlines

Critical Path calculates the sequence of project activities by taking into account the following project constraints:

  • Technical (business process) dependencies;
  • Activity and resource calendars;
  • External incoming dependencies;

Other types of constraints: resource dependencies, material supply and financial limits could also be added but related dependencies have to be identified manually or with a scheduling tool that supports Resource Critical Path. 

Additionally, projects may have situations when all prerequisite conditions have already been achieved, but the activity is still not able to commence. For example, training is scheduled to start next month. In theory, such a constraint could be added as an activity calendar constraint but practically it is not convenient to create a new calendar for each ‘Start Not Earlier Than’ (SNET) constraint presented.

It is much easier to add the SNET constraint to the related activity.

SNET constraints are required as sometimes activity could not commence earlier even though all predecessors completed.

All constraints in the classical critical path are from the Left side of activities or milestones. As a result, the Negative Total Float is impossible as nothing constrains activities from the right.

Total Float with project deadlines

A project deadline represents a desired target date. This target is represented by the ‘Finish No Later Than’ (FNLT) type of constraint, which is always from the RIGHT side of an activity or a milestone.

  • The deadline could be put to the end of the project or an intermediate milestone, for example, the ‘Go-live’ date
  • A project may have one or many deadlines, for example: ‘Output N1 implemented’, ‘Output N2 implemented’, etc.

Schedules with deadlines may have activities with Negative Total Float if the deadlines are taken into analysis.

Which type of Total Float is more important?

Both types have advantages and disadvantages, and it is very important to analyse both perspectives!

Total Float

Total float is the total amount of time that a schedule activity may be delayed from its early start date without delaying the project finish date, or intermediary milestone.

A Project Critical Path allows seeing the true Total Float (Total Slack in Microsoft Project):

How many days (or hours) could an activity be delayed before it imposes a delay on the project END date or at least one of the DEADLINEs is impacted?

Based on the above definition, TF could not be negative. Only zero or positive. However, it is useful to know how an activity is progressing against project targets as well. If an activity has negative TF, acceleration is required. If the TF is positive, it has a contingency.

On the other hand, if we allow TF to be negative, it would be much harder to understand which activities are on the project’s critical path.  A critical path with imposed deadlines may have all project activities with positive or negative Total Float.

Critical Path Example

Let’s review the next project. It has 3 streams of works: ABF, ACDF and AEF. All dependencies are finish-to-start.

Without calculating a critical path, it is not easy to understand which activities have a contingency (aka Total Float or Total Slack).

The Project Critical Path gives us answers to the next questions:

  • Which activities drive the project completion date?
    It is very easy to understand that: activities with TF=0 are critical: ABF
  • When the project is going to be completed?
    Answer: in 18 days, 24th of the month.
  • How much contingency non-critical activities have?
    Activities with TF > 0 have next TF: C = 3 days, D = 3 days, E = 5 days

Now assume that this project has a deadlinea and activity F have to be completed by 22nd.

How is the CP calculation going to change if the deadline is added to the schedule?
The Critical Path will be the same: ABF

Would the Total Float now be different? Different scheduling tools give different answers to question!

Critical Path in the context of the (22nd) deadline:

Now, when the schedule has positive and negative Total Floats, it is not easy to understand which activities drive the critical (ABF) path and what is the REAL size contingency non-critical activities have.

What if activities C or D are delayed for 2-3 days? It delays the project end date ONLY if any of the critical activities A, B or C could be accelerated! Otherwise, this delay will not impact the CURRENT delivery date.

It is very important to know both perspectives: Project Critical path and Dealine Critical Path! 

Unfortunately, Primavera and Microsoft Project users have to decide which perspective is more important for them and stick with the approach.

Once, while performing schedule analysis, I’ve asked myself, is there a way to see both perspectives in these tools? With some tricks I have found workarounds that is explained below.

Deadlines in scheduling tools

Let’s review how popular scheduling tools calculate Total Float with Deadline dates.

Primavera, Microsoft Project and Spider Project have completely different features to manage project deadlines and calcualte Total Floats.

Spider Project

In Spider Project, deadlines could be added on a project level or an activity level.

Project Deadline:

Spider Project (SP) allows adding SNET and FNLT constraints for any activity or milestone.

Based on the TF definition it could not be negative! This is exactly how Spider Project calculates it. The SP CP algorithm takes all constraints into the consideration, and if an activity doesn’t have an available contingency, TF will be zero.

If a user wants to know TF in the context of deadlines, there is another metric- Float Negative, that shows how many days (or hours) the activity is ahead or behind schedule.

Both metrics could be shown in Gantt Chart Report.

This is what we need but how we can get similar metrics in P6 and MSP? 

Microsoft Project

Microsoft Project has two options to set deadlines: by using constraints or the ‘Deadline’ feature.

The first option would be to use a ‘Finish No Later Than’ (FNLT) constraint.

However, such date enforcement may cause schedule conflict, and the project completion date will be incorrectly shown as 22/05/2023.

All constraints in Microsoft Project (excl. ‘Start No Earlier Than’) must be avoided. 

ASAP and ALAP are not constraints but activity priorities. It is unclear why Microsoft mixed up these two concepts.

So, if constraints are not recommended, how to set up Deadlines then? Microsoft has an important advantage over Primavera: the Deadline feature!

Instead of adding a deadline as a constrain date, MS Project has a separate Deadline field for that:

In this case, the project completion date will be driven by logic, and Total Slack is calculated in the context of project deadlines.

The Microsoft Deadlines don’t impact the calculation of the critical path but the impact calculation of Total Slack only.

We have found a good way of calculating the critical path correctly and understanding the Total Slack but lost the opportunity to understand the true Total Floats.

As a workaround, project deadlines could be stored in a custom field (Date), and Deadline Total Slack is stored in a separate custom field (Number).

TotalSlack Deadline Metric:

We store the deadline date in the DeadlineDate field instead directly in Deadline:

When we want to know Total Slack in the context of deadlines, we copy the deadline date to the Deadline field, recalculate TF and store the result in the custom ‘TotalSlack Deadline’ field.

Then, deadline dates are removed from the Deadline field, and the user can see both types of Total Slack: with and without deadlines!

By creating Microsoft Project Macro, this process could be fully automated:

Sub TotalSlackDeadline()

‘ Macro Created 7/01/2023 by Alex Lyaschenko.

‘ Custom Field ‘Date1 (DeadlineDate)’ stores Project Deadline Dates

‘ Custom Field ‘Number1 (TS Deadline)’ store Deadline Total Slack

    SelectRow Row:=0, RowRelative:=False

‘ Copy Dates from DeadlineDates to Deadline

    SelectTaskField Row:=0, Column:=”Date1″, RowRelative:=False, Height:=7

    EditCopy

    SelectTaskField Row:=0, Column:=”Deadline”, RowRelative:=False

    EditPaste

 ‘Switch On formula Number1 = [Total Slack]/480 (8h per day)  

    SelectTaskColumn Column:=”Number1″

    CustomFieldPropertiesEx FieldID:=pjCustomTaskNumber1, Attribute:=pjFieldAttributeFormula, SummaryCalc:=pjCalcNone, GraphicalIndicators:=False, AutomaticallyRolldownToAssn:=False

‘ MSP automatically calculate Deadline Total Slack

‘ Switch Off the formula

    CustomFieldPropertiesEx FieldID:=pjCustomTaskNumber1, Attribute:=pjFieldAttributeNone, SummaryCalc:=pjCalcNone, GraphicalIndicators:=False, AutomaticallyRolldownToAssn:=False

‘ Delete Dates from Deadline fields

    SelectTaskField Row:=0, Column:=”Deadline”, RowRelative:=False, Height:=7

    EditClear Contents:=True

End Sub

Primavera

Primavera doesn’t have a similar to Microsoft Project the deadline feature, and activity deadlines could only be organised via constraints.

Primavera has 9 (!) types of constraints, and it is not intuitively clear what is the key difference between them. Often planners don’t apply them incorrectly. It is probably one of the biggest factors that impact the quality of Primavera schedules.

As discussed above, there are only two genuine types of constraints: NET (Start on or After) & NLT (Finish On or Before). The rest is a ‘scheduling noise’ that impacts the quality of schedules.

Primavera has four ‘Finish’ constraint types. Constraints impact the calculation of the critical path and the TF differently.

  • Finish On or After
    This constraint type should be represented by the ‘Start On or After’ constraint, plus the activity duration or activity type changed to ‘Level of Effort’.
  • Finish On or Before
    FOOB constraint is the same as ‘Finish No Later Than’ constraint in Microsoft Project and the ‘NLT’ constraint in Spider Project.
  • Finish On
    It shouldn’t be applied as it may cause a schedule conflict.
  • Mandatory Finish

    It shouldn’t be applied as it may cause a schedule conflict.

A deadline, ‘Finish On or Before’ constraint doesn’t cause schedule conflict but impacts TF  calculation.

This constraint type could be used for calculating TF in the context of project deadlines. However, if such constraints are applied, it is impossible to understand the true TF!

As a workaround, project deadlines could be stored in the Deadline field, and an additional Critical Path Metric: ‘TF Deadline’, is calculated only when detailed Critical Path analysis is required.

  • When a Primavera user wants to see the true Total Float, all ‘Finish On or Before’ (FOOB) constraints should be deleted. It is easier to do by Global Change. 

When the user additionally wants to know TF in the context of deadlines, they can automatically copy the deadline dates from the Deadline Field, recalculate the critical path, copy Total Float to the TF Deadline field and Remove all FOOB constraints.

Add Deadlines as FOOB Constraints:

Primavera’s Deadline field, compared to Microsoft Project, only displays the Deadlines in Gannt Chart and doesn’t impact TF.   

Copy TF to TF Deadline field:

The result of the calculation is stored in the separate TF Deadline field. So, it is possible to see both types for Total Float, with and without deadlines.

Summary

One of the greatest advantages of the critical path method is the ability to understand the available contingency for each activity in the project. However, the contingencies could be calculated with and without project deadlines. For effective project planning and control, it is important to know both perspectives.

Popular scheduling tools have different features to manage project deadlines that impact the calculation of critical paths and Total Floats.
Spider Project provides both metrics simultaneously.
Primavera and Microsoft Project users have to decide which perspective is more important for them.

With proposed workarounds, users can have an additional Critical Path metric: Total Float Deadline and understand both perspectives simultaneously!
The additional metric is important for effective project control and forensic analysis.

Alex Lyaschenko

PMO | Portfolio Planning & Delivery | PMP | P3O Practitioner | AgilePM Practitioner | Six Sigma

Schedule Quality Metrics. Out-of-sequence Activities

Schedule Quality Metrics. Out-of-sequence Activities

One of the important schedule quality metrics is ‘out-of-sequence activities’.

“Out of Sequence Activities” (OOS) – a result of executing the works without following the logic as planned.

There are many very good posts that explain what is “Out-of-sequence activity”, “How to identify them” and even “How to remove them from the schedule”. I am not going to repeat that. You can find links to these posts below.

The questions that I wanted to discuss in this post are:

  • Why do projects have out-of-sequence activities?
  • Is it good or bad to have them?

Why do projects have ‘out of sequence’ activities?

The short answer to this question: the plan was wrong. However, there are GOOD and BAD reasons why the plan was wrong:

GOOD reason: The project team is agile.

A project may have OOS activities if teams working on site find a better way of doing things and the project manager/planner didn’t think of it first. Planning guru Rafael Davila from Puerto Rico once said: “If you have out-of-sequence working it does not necessarily mean it is a planning error, it is common when you plan to avoid too many people working at the same time in the same place and then during the course of the job you find out it is convenient to reduce the waiting time. Life and construction works are complicated, pretending to model such with only FS relationships is a utopia.”
Any project has stable technical (business process) and not stable non-technical dependencies to address resource, material, financial and space constraints. If a project delivery team is agile (yes, even construction projects could be agile!) they are constantly looking for an opportunity to remove non-technical dependencies and commence work earlier.

BAD reason: Schedules don’t apply correct logic.

Most project managers and schedulers are aware that there are 4 primary types of schedule dependencies: Finish-to-Start (FS), Start-to-Start (SS), Finish-to-finish (FF) and Start-to-finish (SF). Recognised project management and scheduling standards (PMBoK, APM Body of Knowledge, Praxis Framework, etc) explain each type but somehow miss one critical point:

Four types of project dependencies are mutually exclusive. It is not possible to create a simulation of one type by applying another type correctly.

If a business process requires the application of, for example, SF dependency the scheduling fragment could not be correctly simulated by replacing the SF dependency with FF, SS or SF types.

Next time when you notice that project (or portfolio) scheduling guidelines recommend to minimise the application of SS and FF dependencies and prohibiting the application of the SF type be sure:

  • A Schedule is not going to be used to drive project delivery. Maybe only for reporting.
  • It is going to be many ‘out of sequence’ activities when the progress is collected.

The quantity of SS, FF and SF dependencies depends on the nature of the project. Some projects are going to have more parallel and overlapping activities than others.

It is not possible to define strict percentage targets and claim that the schedule doesn’t have the decent quality only because there are more than 1X% of other than FS dependencies.

DCMA-14 points Schedule Health Assessment

DCMA-14 points metric N4: ‘Finish-to-Start relations’ often get misinterpreted. This is one of many examples (schedulereader.com):
“The finish-to-start relationship provides the most explicit presentation of the project schedule activities. The other types of relationships, which can be identified in a schedule, are finish-to-finish (FF), start-to-start (SS), and start-to-finish (SF). However, it is not recommendable for these to be used since they are harder to monitor and control.” 

What this metric is actually showing is that a schedule with a high percentage of other FS dependencies requires higher attention. It doesn’t mean that schedules don’t have low quality. There could be valid reasons why there are more than usual SS, FF and SF dependencies or maybe the project is applying SS instead of FS to hide project progress. The reasons must be discovered and accepted.
Also, if a schedule has a very high percentage (some have 100%) of FS dependencies, it is a great candidate for a deep schedule review. It is likely that the project doesn’t have an experienced planner, or the planner is forced to follow low-maturity portfolio standards.
If a schedule has 100% of FS dependencies ‘DCMA 14-points’ incorrectly shows that this metric is ‘Green’, when it should be flashing red.

The quantity of SS, FF and SF dependencies depends on the nature of the project. Some projects are going to have more parallel and overlapping activities than others.

It is a misleading metric and should not be applied blindly.

BAD reason: level of details

Some schedules don’t have the right level of detail. Planned sequenced activities are actually overlapped. In this case, the project may consider completing decomposition or adding lags and leads.

Fast-tracking

Another reason why projects may have out-of-sequence activities is fast-tracking.

Fast Tracking [Technique]. A specific project schedule compression technique that changes network logic to overlap phases that would normally be done in a sequence, such as the design phase and construction phase, or to perform schedule activities in parallel. (Practice Standards for Scheduling, PMI)

Fast-tracking is always associated with risk. If a project team took this risk and started the successor activity earlier by breaking the technological process, the question that needs to be asked is: “Has this risk been analysed, and if yes, why is the schedule not updated?”. If the analysis was completed without schedule-based what-if analysis, it indicates that the project doesn’t have schedule and risk management integrated.

Scheduling Tools

Projects managed in Primavera and Microsoft Project are likely to have more out-of-sequence activities for the bad reasons. There is still no front-end access to the dependency table, only via the back-end and it is not possible to differentiate a type of dependency in these tools. So, for planners, it is much harder to change logic as they don’t know which dependencies are technical dependencies and should not be touched (unless it is approved fast-tracking), or if it is a dynamic resource dependency.

Primavera recently added a ‘dependency note’ feature, but for large organisations, it would take years to update for the planners to have access to this feature.
Also in Primavera, it’s not easy for P6 planners to change future logic, they focus more on progress updates.

Schedules managed in Spider Project typically have fewer OOS activities as broken dependencies could be easily identified within the tool. Spider Project has front-end access to the dependency table and a powerful filtering mechanism. 

OOS Types

As there are four types of dependencies, there are also four basic examples of how the activity could be OOS. However, as an activity may have lags and leads there are other, more complex scenarios of OOS logic.

It is important to identify all OOS cases, not only broken FS dependencies.

Summary

  • Out-of-sequence activities represent broken logic.
  • There are good and bad reasons why projects may have OOS activities.
  • OOS logic has to be identified and fixed including other than FS broken dependencies.
  • It is very important not just to fix already broken logic, but also to analyse why the project has broken activities and, if required, apply changes to the outstanding activities and mature schedule management processes.

Alex Lyaschenko

PMO | Portfolio Planning & Delivery | PMP | P3O Practitioner | AgilePM Practitioner | Six Sigma

Monte Carlo Simulation Challenges. Critical Path Volatility Challenge.

Monte Carlo Simulation Challenges. Critical Path Volatility Challenge.

The ‘3 points estimation’ known as a PERT methodology is relatively simple to use for cost, but for the schedule, it’s more complicated. The fact that some activities are on the critical path means that the volatility of those specific activities, as opposed to the set of all project’s activities, can make a big difference for Monte Carlo Simulation.

Risks and Critical Path

It is well accepted that risks are measured based on the probability of the event (or condition) occurring and the consequences.

Risk Exposure = Probability (Event) * Consequence

The same risk may have different consequences for Time, Cost, Safety, Reputation and other project objectives.

  • The time consequence is highly dependent on the criticality of impacted activities. If impacted activities are on the critical path# [including delay], the project is going to be delayed. Otherwise, it will impact some parts of the project but not the overall delivery date(s).
  • Risk consequences are not always a stable parameter. Some of them depend on project conditions.
  • Risk consequence has to be measured based on future conditions when the risk is likely to occur, not the current situation.
  • When the risk is likely to occur project critical path may be different from the current critical path. It could be due to the impact of project uncertainties, other risks, scope changes or any other reasons.
  • At that time, there will be only two possibilities: the critical path impacted or not. In other words, the probability (CP) that the risk has a time impact will be 0% or 100%. Do not mix with the Probability of the event, which is not dependent on the critical path.

However, when the risk is assessed, it is not always possible to predict if the impacted activities are going to be on the critical path or not, so the probability (CP) could be different from 0% and 100%.

Probabilistic vs Deterministic Delivery Model

If the Monte Carlo Simulation is performed based on a dynamic PROBOBALISTIC model (aka schedule) that includes data uncertainties and “conditional branches ”, the critical path will be recalculated during each calculation, and only the event probabilities need to be assessed prior to the simulation.

PROBABILISTIC project delivery systems, like Spider Project, could additionally calculate the probability of each activity being on the critical path, which is known as ‘activity criticality index’.

However, if the Monte Carlo Simulation is performed based on a DETERMINISTIC schedule (developed in Primavera, Microsoft Project or another CPM-based scheduling tool) and risk data is collected in a separate risk register, the probability that impacted activities are going to be on the critical path (CP) has to be assessed manually. The accuracy of such an evaluation would be significantly different from calculated probabilities and it will have following effect the results of the Monte Carlo Simulation.

Some risk management tools provide users with an option to define if the critical path is going to be impacted or not with a binary option (Yes/No). Users potentially could assess it’s based on the current critical path, but, with some rare exceptions, it is impossible to say whether the future critical path is going to be impacted or not as it depends on other materialised risks.

Some project Risk consultants just assume that the Critical Path is going to be the same

(Probability (CP) =100% or 0%) and only assess event probability even if the previous similar projects had critical path volatility.

Risk Exposure

If the (time) consequence depends on the Project Critical Path, or to be precisely the probability that impacted activities are going to be on the critical path when risk is materialised, Risk Exposure also depends on this parameter:

Consequence (Net) = Probability (CP) * Consequence (Gross)

So:

Risk Exposure = Probability (Event) * Probability (CP) * Consequence (Gross)

or

Risk Exposure = [ Probability (Event) * Probability (CP)] * Consequence (Gross)

Cost Impact

So far, we have discussed only time exposure. What about the cost exposure? Is it also dependent on the critical path? Yes, but in a different way.

Risk cost consequence has two parts: direct cost impact and cost of delay.

Direct cost impact only depends on the Probability (Event) and the cost of delay only depends on Probability (CP).

Summary

The volatility of the Critical path means that risk consequence is not a stable parameter and depends on future project conditions when the risk is likely to occur. It is hard to accuratelly predict which activities are going to be on the Criticla Path at that time.

The results of Time and Cost probability distributions received based on deterministic the CPM-based schedule and predefined risk data could be misleading and should be used for project decisions with caution.

Risk consequences have to be calculated in each simulation. Dynamic probabilistic model with fully integrated risk data is required for reliable Monte Carlo Simulation.

Alex Lyaschenko

PMO | Portfolio Planning & Delivery | PMP | P3O Practitioner | AgilePM Practitioner | Six Sigma

Risk vs Uncertainty

Risk vs Uncertainty

At project management conferences, LinkedIn discussions and even in one of the most recognised Project Risk Management book, I’ve found that uncertainty is just a risk with 100% probability. Not sure what is the origin of this statement, but a risk with 100% probability is an issue, not an uncertainty.

Managing project risks and uncertainties is essential for mature risk management and successful project delivery. Lack of understanding differences between these two concepts creates serious project planning and control issues and often results in project delays and cost overruns. In this post, we will review what is common between these two concepts and where the key differences are.

Project Risk Management Standards

Risk and Uncertainties are interrelated. Risk arises out of uncertainty and generates uncertainty.

Project Management Institute (PMI) and Association for Project Management (APM) used these terms widely in their project management and risk management standards. Both organisations suggest dividing the term Risk into two terms: Risk Event and Overall Project Risk. PMI and APM don’t provide definitions of Uncertainty but mention that the source of uncertainties is complexity, ambiguity, and volatility.

Project Event

Risk Event Definition by PMI:

Risk is an uncertain event or condition that, if it occurs, has a positive or negative impact on at least one project objective. (The Standard for Risk Management in Portfolio, Program, and Projects)

Risk Event Definition by APM:

A risk event is an uncertain event or set of circumstances that, should it occur, will have an effect on achievement of one or more of the project’s objectives. (Project Risk Analysis and Management Guide)

Overall Project Risk

Overall Project Risk Definition by PMI:

Overall project risk is the effect of uncertainty on the project as a whole, arising from all sources of uncertainty. This includes individual risks and the exposure to the implications of variation in project outcome, both positive and negative. Overall risk is often a function of complexity, ambiguity, and volatility. (The Standard for Risk Management in Portfolio, Program, and Projects)

Overall Project Risk Definition by APM:

Overall project risk is joint effect of risk events and other sources of uncertainty. Project risk results largely from the accumulation of a number of individual risk events, together with other sources of uncertainty to the project as a whole, such as variability and ambiguity. (Project Risk Analysis and Management Guide) 

Unfortunately, apart from the definitions, both standards don’t provide any methods and techniques to measure and control Overall Project Risk. There is no common acceptance among project and risk managers on what it is the “beast” and how to deal with it.

Success Driven Project Management (SDPM) suggest dividing risks into two major categories: uncertainties and events, where uncertainty is a property of project parameters. Based on SDPM, changes caused by risk are discreet events, while changes caused by uncertainties are continuous.

Let’s look at risks and uncertainties more closely.

Risk

Less formally, risk is defined as the situation of winning or losing something worthy.

In projects, stakeholders are interested in understanding how such situations may impact project objectives, usually represented as the delivery of outputs (or critical deliverables) to the committed dates with the agreed budget. Depending on the type of project, objectives may also include quality, benefits, reputation and other important aspects of the business.

“Worthy” means that while risk events may impact other aspects of the project, these are the only important ones in the context of the overall delivery.

For example, a project team may raise a risk that a particular activity might be running late or next month’s forecast could not be met if a risk event materialised.
While this information still may be necessary for some stakeholders, a project risk usually is raised in a broader context:

  • Does it mean that if the activity is late, the project will also be late?
  • Does it mean that if the next month’s actuals are more than was forecasted, the project will be over budget?

‘Risk impact’ is a dynamic characteristic, not a stable parameter. A risk may have no impact on the project objective now, but the situation may be different tomorrow, and the risk would become the most critical risk in the project. Without Quantitative Risk Analysis methods, it is not easy to identify such changes timely.

There are many books and risk management standards with recommendations for risk control. While uncertainties are also considered in these books and standards, I think this area of project risk management is less developed and definitely less controlled.

Primary Uncertainty

Uncertainty is a situation where multiple alternatives result in a specific outcome, but the probability of the outcome is not certain.

Any project has hundreds, if not thousands, of situations when the outcome is uncertain. However, any project has quite limited project parameters with uncertainties.
Examples of parameters that may have uncertainties:

    • Volume of Work
    • Resource Productivity Rate
    • Resource Quantity
    • Resource Calendar
    • Activity Calendar
    • Material Quantity
    • Activity Duration
    • Workload
    • Lag Duration
    • Material Cost
    • Unit Cost
    • Currency rate
    • Resource Rate
    • Inflation Rate

There are examples of bottom-level uncertainties, also known as base, or primary uncertainties.

  • Most, but not all project parameters, have primary uncertainties.

Example:
Office calendars, compared to construction calendars, don’t depend on weather conditions.

  • Data collection challenges

It is easy to collect primary uncertainty data as each refers to only one project parameter. Three-point estimation is a common practical approach. Usually, the critical constraint for data collection is not the information itself but the lack of opportunity to store it in a project management system. Not many of them are designed to store and apply primary uncertainties to calculate project delivery plans. Spider Project is an excellent example of such a system when primary uncertainties are stored as corporate norms and used to calculate project delivery plans with probabilities to achieve committed project objectives: dates, costs, benefits, etc. 

I believe there is a lack of recognition of the benefits of capturing and applying uncertainty data for the reasons explained below. Many organisations with such experience are in non-English-speaking countries and usually don’t publish results in recognised PM Journals.

Secondary Uncertainties

The outcome of primary uncertainties may result in uncertainties of other project parameters, especially activity duration and cost uncertainties, which are secondary or middle-level uncertainties.

  • The secondary uncertainty is the result of one or multiple primary uncertainties.

Example:

An activity has uncertainty in the volume of work and productivity of assigned resources. Uncertainty in the duration and cost of the activity is the result of both uncertainties.

  • Activity duration and cost uncertainty may or may not depend on other primary uncertainties.

So, it could be a primary or secondary uncertainty. It is very confusing, but who said that project management has to be simple?

Example:

An external project review takes 24-40 hours. It means that the ‘volume of work’ of this task is measured in hours, and in the worst-case scenario, the report is going to be completed when 40 hours of effort, even if the quality of the report needs to be compromised.

  • Not all primary uncertainties impact duration and cost of activities.

Example:

‘Resource Quantity’ uncertainty may impact the order of planned activities but not in their durations and cost.

  • Result of Monte Carlo Simulation (MCS) highly depends on the type of uncertainties used as a base for analysis.

Example:

Results of MCS analysis performed based on secondary uncertainties may significantly vary from the analysis performed based on primary types, as important correlations between project parameters are missing. This is another big topic to discuss, and I’m planning to cover it in a separate post under the “Monte Carlo Simulation challenges” posts.

Project Level Uncertainties

Multiple primary and secondary level uncertainties and related risks result in probability distributions of project commitments. These distributions could be considered as ‘top-level’ or tertiary uncertainties.

Examples of top-level uncertainties include:

• Delivery date of committed output
• End of the design phase
• Cost of project

• Probability distributions could be considered in the context of individual project objective or with the relationship to other objectives.

Risk Uncertainties

Projects usually estimate risks with probability, as a percentage of the likelihood of the event or condition to occur, and the impact estimated in days and $. However, this information is often subjective as different project stakeholders have different opinions about the probability and impact.

As a result, a risk may have uncertainties in the probability, the impact(s) or both at the same time.

Cost of Knowledge

Both, uncertainties and risks depend on the knowledge of estimators. Sometimes it may be possible to gain this knowledge and reduce uncertainties but usually, it is associated with extra effort and additional cost.

Sometimes it is worth spending extra effort to gain missing information that could help with project decisions, sometimes not. In other situations, the required information may not exist and any effort to clarify the information is a waste of time and money. It is another separate big but interesting topic to discuss.

Portfolios with advanced project management collect historical data and apply more accurate estimations to outstanding delivery and future projects. It is important to mention that, if possible, historical data has to be collected with primary types of uncertainties. If only secondary (activity level) uncertainties are collected, the critical context would be missing, and the historical data could not be applied correctly.

Example:

An activity has uncertainty in duration due to the uncertainty in the productivity of the planned resource: new bulldozer. The uncertainty is clarified during project execution and, in theory, next projects could use this information to reduce volatility. However, if only duration uncertainty was documented and another bulldozer with a different productivity rate is assigned to perform the same type of work, the historical data would be misleading as the critical context is missing to understand that this type of work has a different duration now.

Impacts of Risk and Uncertainties

  • Not all uncertainties have an impact to project objectives.
    Sometimes, under current conditions, regardless of the result of the uncertainty, it would be no impact to project commitments.

Example:

Next month, a project must pay for materials in a different currency. There is uncertainty about next month’s currency rate. This uncertainty could be an opportunity, treat or has no impact to project cost at all. It depends on:

  • Which way does the currency rate go? Maybe it is an opportunity, not a threat?
  • Currency rate the project forecasted. Maybe it was forecasted based on the worst rate?
  • Conditions in the contract. Maybe the project has a fixed currency rate?
  • A risk is always a threat or opportunity from one perspective (time, cost, etc). Uncertainties could be a threat and an opportunity to the same perspective simultaneously.

Example:

There is uncertainty in the volume of work that will impact the duration and cost of the activity. The uncertainty was estimated as a range (100 – 150 m2). If the project plan has 125 m2 as a base, it would be an opportunity to complete the activity ahead of schedule or a threat of delaying the completion.

  • The same risk could be an opportunity for one perspective and a threat for another.

Example:

The bulldozer with certain productivity is planned to be assigned to work. There is a 20% chance that this bulldozer is not going to be available, and the project needs to hire another more expensive but also more productive bulldozer. If the risk is materialised, it will be a threat to the project cost, but also an opportunity to project delivery date (if planned work is on the critical path). So, this risk is a threat to cost and an opportunity to time. When a project uses a risk register and classifies all risks as a threat or opportunity, the mitigation actions may solve some potential issues but generate others.

Creeping Normality

Primary uncertainties impact projects via very small changes, but commutatively, they may have a large effect. Such an effect could be even more severe than the impact caused by project risks.

Project control teams usually have less control over activities with sufficient contingency. Positive and negative changes caused by uncertainties slowly and in a non-linear way “eat” available contingency, and when suddenly activities start to be critical and impact project delivery dates, it is already too late to apply mitigation actions. The contingency already has been spent. It is even more applicable to project cost when a lot of minor cost increases slowly “eat” cost contingency.

I think this could be due to “Creeping Normality” cognitive bias, also known as “Landscape amnesia”.

Creeping normality is a process by which a major change can be accepted as normal and acceptable if it happens slowly through small, often unnoticeable, increments of change. The change could otherwise be regarded as remarkable and objectionable if it took place in a single step or short period.

It creates an illusion that ‘risk events’ is the biggest factor in project delay and cost overrun when, in reality, changes caused by uncertainties have a bigger effect caused by project risks.

Projects that document uncertainties and compare them with actuals can notice “patterns” and apply mitigation actions before it is too late.

Relationship Between Risks and Uncertainties

While it is obvious that Risks and Uncertainties are interrelated, the relationship between them is not always obvious.

  • Uncertainties in one part of a project may have a negative (or positive) impact on risks in another part of the project.

A risk associated with activities on the critical path is likely to severely impact project delivery dates. Changes caused by uncertainties in another part of the project may have an indirect impact on the risk activities total float and the risk mitigation actions should be changed as well. However, as projects usually don’t control uncertainties, it is hard to notice when the impact of risk is changed. In the future post, I am going to demonstrate this effect with a more detailed example.

Summary

Risk and uncertainties are important elements of Project Risk management. They are interrelated but have critical differences that are not always understood correctly. While Project risk management standards recognise the importance of management of uncertainties and risk, projects across the Globe often don’t have mature control over uncertainties.
For project success, it is not sufficient to only control risks. Mature Risk Management also has to include the management of uncertainties.

Alex Lyaschenko

PMO | Portfolio Planning & Delivery | PMP | P3O Practitioner | AgilePM Practitioner | Six Sigma

Monte Carlo Simulation Challenges. Distribution Shapes Challenge.

Monte Carlo Simulation Challenges. Distribution Shapes Challenge.

Risk simulation systems offer a choice of user-selected distribution shapes that can be implemented for each and every activity. Some tools offer only primary types, others have a large menu of over 50 distributions.
However, selecting a specific shape for each of 1,000+ activities (on top of generating three estimates for each item) requires many hours invested. The majority of users run the system using one of the distributions applied to all activities with uncertainties. It is mostly Triangular, Beta or Log-normal distributions.

What they may not be aware of is that different activity distribution shapes can substantially alter the result of the analysis. The result of the simulation performed with beta (PERT) and Triangular distributions at the P80 mark is likely to have a noticeable 8-15% difference. At the same time application of the same distributions at the different P-point may have a less noticeable difference.

Some construction and consulting companies may use it to get extra funds (or time).
All they need to do is just to run the risk analysis with another distribution type!
Receivers of project risk simulation reports need to review which distribution shape was applied and they need to understand how it may alter the presented result. As a minimum, there must be consistency between reports produced over time.

The Triangular Distribution

When projects don’t have subjective historical data and only a subjective range (2-point estimation) is available, it makes sense to calculate the Most Likely estimation as an average between the optimistic and pessimistic estimations: ML=(O+P)/2 and apply the triangular distribution.

Example:

A team needs to perform 10 activities. Each activity may take anything between 8 and 12 days. In this case, the most likely duration would be 10 days and symmetrical triangular distribution should be applied.

However, if there is a reason to believe that the most likely estimation is different, the asymmetrical triangular distributions would be the choice.

Another advantage of the triangular distributions is that result of the Monte Carlo simulation would be the same regardless of the applied risk simulation tool, with an assumption that both tools don’t have other issues which we are going to discuss in future posts.

The disadvantage of the triangular distribution is that often the pessimistic and optimistic estimations represent rare extremes and it has to be reflected in the shape. In this case, projects need alternative shapes that could reflect this anomaly. The most known are the beta and Log-normal distributions.

Beta Distribution

The famous PERT method also based on Three-point estimations is able to address “extreme ends” anomalies. The PERT addresses it by taking More likely estimation 4 time often than Optimistic and Pessimistic estimations: T = (O + (4* ML) + P) / 6.
A beta distribution is often considered the best choice for project simulations. While it sounds very scientific, each risk management tool uses its own formula for beta-distribution! In fact, it is actually a family of distributions known as PERT, beta-PERT, modified PERT, etc.

This means that if a beta distribution was used as a base, an analysis is performed with the same initial data and the project delivery model but different risk simulation tools will produce different results.

Summary

The application of different distribution shapes and the way they are applied to the project delivery model could significantly alter the final result of generated distributions calculated with the Monte Carlo Simulation method. However, this challenge should not be considered as a stand-alone problem. Other challenges, like data collection, quality of project delivery model, and others, may have even more significant effects. In this case, the variability of results caused by incorrect application of distribution shapes should be considered a comparatively smaller issue.

Alex Lyaschenko

PMO | Portfolio Planning & Delivery | PMP | P3O Practitioner | AgilePM Practitioner | Six Sigma