Lesson 11: Sequence Activities and Estimate Activity Durations

Budgetary Estimating Techniques

-Most closely associated with detailed planning -Higher-quality estimate to support your schedule and cost baselines -This is your commitment of key stakeholder expectations -The accuracy range improves as you take more time to develop the estimates. The best way to get higher quality estimates is with bottom-up estimating methods. In other words, you start at the activity level and roll up to the work package level and up to the next-higher WBS levels. We'll discuss two budgetary estimating techniques today.

Analogous Estimating

-Most common type of estimating - A form of top-down estimating Likely begins with usable information from a prior project and supplements this information with knowledge from team members and subject matter experts. You use the actual durations from a previous project as the starting estimate for the current project. Just be sure to make adjustments for differences (in geography, scope, and timing) between your current project and previous ones. If historical information is different from what you expect, there is still value to it because it may generate productive discussions that lead to usable estimates for your current project.

Preliminary Estimating Techniques

-Quick and easy -Not very accurate -Good enough to develop the business case and obtain project approval -Establishes expectations and allows for decision-making -Also called a top-down method, rough order of magnitude (ROM), or order of magnitude (OOM) estimate

Things to Remember for the Exam

1. Best practices recommend using mandatory finish-to-start dependencies exclusively for the first iteration of your schedule. Add discretionary dependencies in later iterations. 2. The finish-to-start is the most common and lowest-risk dependency type. 3. The start-to-finish is the least common dependency type. 4. Dependency types that create overlap between predecessor and successor (SS and FF) increase schedule risk. 5. Activities may have more than one dependency, and in that case one will be the higher priority. This may lead to eliminating the secondary dependencies, so you have a "clean" picture. 6. Every activity (except for the first one) will have a defined predecessor. Every activity (except for the last one) will have a defined successor.

Activity Dependency Types

1. Finish-to-Start (FS) 2. Start-to-Start (SS) 3. Finish-to-Finish (FF) 4. Start-to-Finish (SF)

Lags

A lag delays the start of the successor as it relates to the predecessor. This moves work later in the schedule and shifts work to the right. A lag is the opposite of a lead. Instead of starting the successor earlier, you're pushing it out later in relationship to its predecessor. For example, suppose you're restoring a number of vintage office buildings as part of a downtown revitalization project. After you remove the old paint and prepare the interior walls, you'll need to apply two coats of paint. The type of paint used and the humidity in the air will slow down drying time. You'll need to allow additional drying time (about two days). A lag is usually depicted as a positive number (+2d to add two days). You can enter a lead or lag in any unit of duration, including weeks, days, hours, or minutes.

Leads

A lead changes the amount of time the successor overlaps the predecessor activity. This moves work earlier in the schedule and shifts work to the left. Suppose you need to have a city manager review your modification of the lobby in the office buildings. After talking with this official, you learn that she needs to review your work three days before you finish. To achieve this, you can model a finish-to-start dependency and apply a three-day lead between the two activities. A lead is usually depicted as a negative number (-3d for subtracting three days). The three-day lead compresses the overall duration by allowing overlap. Nothing about either activity has changed. There are the same duration estimates, resources, and work. However, the combined duration between them is compressed. A lead is only used with a finish-to-start dependency. It is a discretionary dependency based on the team's best judgement.

Activity on Node (AON)

Another lesser-used term for the precedence diagraming method (PDM) is activity‑on‑node (AON). You may encounter this term on the exam. An older method called activity-on-arrow (AOA) is no longer used because it did not easily model multiple dependencies. Instead of naming the nodes (placing the letters inside boxes), AOA places activity names on the arrows. The circles (nodes) are there as identifiers and connectors for the activities. So, in contrast to PDM, the nodes for AOA hold no significance other than to define sequential relationships. It's not likely you'll see an exam question on this method. However, it may pop up as one of your answer choices. If you don't know about it, you can't eliminate it as a possible answer.

Discretionary Dependencies

Discretionary is also known as soft dependencies. Unlike mandatory, discretionary dependencies are based on choices or best practices. For example, a choice could be between two resources or approaches. A best practice may indicate a preferred method or technique. Your project team is in a good position to define these. Discretionary dependencies often involve following the experience of other successful projects.

Estimating Durations Overview

Estimate Activity Durations provides the framework to develop an approximation (an estimate) for how long each activity will take. The accuracy range varies significantly by industry and is not testable. However, for the purpose of an example, I'll use the following: If your preliminary estimate was a range of plus or minus 75%, then a budgetary estimating range could be plus or minus 25% and a definitive estimating range is plus or minus 10%.

External Dependencies

External dependencies are those that come from outside your project or company. These are inputs supplied by anyone who doesn't work for your organization. That means you have little or no influence or control over external dependencies. For example, obtaining a permit or inspection from a government entity would be external. It would also be mandatory.

Mandatory Dependencies

If something must happen, then it is considered required or mandatory. Mandatory dependencies are also known as hard dependencies. For example, you cannot buy a piece of equipment until you first define which piece of equipment you need. The first iteration or two of schedule development is primarily based on mandatory dependencies.

Lags #2

Lags are also used when you must account for time passing but no work is being done. The classic example is pouring concrete. You can't start the successor (put up walls) immediately after pouring concrete because it will take three days for the concrete to cure enough for the team to stand on it. No work is being done during the curing time, but we have to account for those days on the schedule. Another example is sending out a contract for signature. Even if you send it overnight and they immediately overnight the signed contract back to you, that's two days you must account for in the schedule. A lag can be used with all four dependency types (FS, SS, FF, or SF). It is a discretionary dependency, and no resources are associated with it (although schedule shifts may impact resource availability).

Delphi Estimating Technique

Like PERT, Delphi also requires multiple estimates. The difference is that with PERT you obtain three estimates (O, P and ML) from one person. With Delphi, you obtain one estimate (a "most likely") from multiple people. There are two main benefits of using Delphi. First, using multiple estimates will mitigate estimator bias. Second, the final estimate is consensus based. You can learn more about how to use Delphi and more specifically wideband Delphi in the Resource section. It will take a few rounds to reach consensus. With each round, you typically throw out the highest and lowest estimate and then question each estimator on what they took into consideration when they arrived at their estimate. The feedback helps every other estimator refine their own estimates and eventually reach consensus.

Finish-to-Start (FS)

P = Predecessor S = Successor The FS dependency is the most common and provides the lowest risk because the successor has 100% of its predecessor complete. There are no overlapping activities with the finish-to-start dependency. The downside is it also creates the longest schedule, which isn't always an option. It is a best practice to create your first iteration or two of the schedule using only FS dependencies.

Start-to-Finish (SF)

P = Predecessor S = Successor The start-to-finish dependency is rarely used. Look at the example and the arrow direction. The predecessor must start before the successor can finish. How can this be? Many people struggle with envisioning this dependency. This is used when the scheduling of the successor determines the schedule position of the predecessor. It's the opposite underlying assumption of the other three dependency types. There are a couple of situations where this type could be used. Examples: 1. Just-in-time (JIT) procurements: Let's say you are working on a large construction project and you want to mitigate the risk of equipment and materials being stolen from the job site. To do this, you want to schedule deliveries as close to the time that they will be needed on the job site. So first, you must build the schedule so that you know when those work efforts will occur (successor). Second you place the procurement activities (the predecessor) on the schedule in relationship to the successor to make sure you order them at the right time. 2. Perishable materials: The SF can be used when you are concerned about expiration dates or spoilage of perishable materials. Like the just-in-time example above, perishable items can't be purchased too far in advance, so you need to know when you'll need it (successor) so that you can determine the earliest time to order them (predecessor).

Start-to-Start

P = Predecessor S = Successor This dependency states that you can start the successor after you start the predecessor. There is overlap between the two activities because you don't require 100% of the predecessor to be complete before starting the successor. Creating overlap will shorten the schedule, but it also increases risk. If something happens in the predecessor that impacts what you've already built in the successor, you'll end up redoing the successor work. The assumption with the SS dependency is moving both activities to start as soon as possible.

Finish-to-Finish (FF)

P = Predecessor S = Successor When you use the finish-to-finish dependency, the successor cannot finish until the predecessor has finished. Like the SS, there is overlap between activities. But unlike the SS, you use the FF when you want things to finish as late as possible. If my project is to create a website for a government agency, I can't finish the site until the government provides me with all the inputs I need.

PERT Formula

PERT = (P + (4*ML) + O) / 6 ML = Most likely outcome P = Pessimistic estimate O = Optimistic estimate You will need to memorize this formula for the exam. What makes the PERT formula unique is that the most likely is always four times more likely to occur than either the optimistic or pessimistic estimate. In a three-point estimate, the weighting for most likely can be whatever you believe is most appropriate. For example, if you believe the most likely is as likely to occur as either the optimistic or pessimistic estimate, then the formula would be (O+ML+P)/3. Note that the denominator changed because the number of elements in the numerator has been reduced to three. PERT calculations represent a 50% confidence level. Imagine a normal distribution of data or bell curve. PERT would be the exact midpoint where 50% of the responses is to the left and 50% of the responses is to the right of that point. On the exam, you may be asked to calculate the expected value. The word expected refers to 50% statistical confidence.

Activity Dependency Modifiers

Project managers have a significant amount of flexibility when scheduling activities utilizing discretionary dependencies. Not only do they want to define dependencies, but they want to fine-tune them by defining exactly when an activity will start. You can modify activity dependencies using leads and lags.

Predecessors and Successors

The activity that you are currently looking at is called the *target activity*. The placement on the schedule for any target activity is dependent on another activity. These dependent activities are called *predecessors*. Likewise, other activities are dependent on the target activity and must occur after it. These are called *successors*.

Precedence Diagramming Method (PDM) #2

The most important thing is to sequence activities using correct logic and to create schedule paths or series of activities. A path is a continuous series of activities that lead from the project start to the project end. You may have multiple paths. The number of paths will be determined primarily by two factors: 1. The number of project resources. If you only have one resource, there can only be one activity scheduled at any given time. If you have five resources, theoretically you could have five activities occur at the same time. 2. Mandatory dependencies will specify when two activities must occur in sequence. There are some rules when creating a PDM schedule network. First, each PDM should have a start and finish milestone. This ensures all activities are connected. Every activity must have a predecessor and successor defined. Arrows are used to indicate not only that two activities have a relationship but also which dependency type is modeled (SS, FS, FF, or SF). When one activity has two successors, there is a path divergence. Activity A is the predecessor to both activity B and C. When one activity has two predecessors, there is a path convergence. Activities D and E are both predecessors to activity F. In this simple diagram, there are only two paths. The first path contains activities A-B-D and F. The second path contains activities A-C-E and F. There is no relationship between activity B and C. They are on two different paths and may occur in the same timeframe. These are parallel paths. Sequential work is activities found on the same path. It's helpful to use a whiteboard and involve your team members. This will get the creative juices flowing. You can then use a computer program to handle the finer details of activity sequencing.

Precedence Diagramming Method (PDM)

The next step is creating a diagram that displays all activities and their relationships to each other. The most common tool or technique is the precedence diagramming method (PDM). The flowchart format, made up of boxes and arrows, is easy to understand and use. The boxes (also called nodes) indicate the work to be done, and the arrows indicate the relationships between work efforts.

Estimate Categories

There are three broad categories of estimates: preliminary, budgetary, and definitive. We'll discuss the first two in depth, as these are most likely testable. Preliminary estimates are developed early in the project. They can be created quickly, but they are not very accurate. Budgetary estimates are developed during planning and are used when setting your baseline. They take more time but are more accurate than preliminary estimates. We'll discuss a couple of estimating methods for each category. The two most common preliminary estimating methods are analogous and parametric. Three point and Delphi are the two most common budgetary estimating methods. Note: Any of these estimating techniques may be used for duration or cost estimating. Exam Tip: On the exam, look for the words preliminary and budgetary when referring to estimating. It will clue you into the timing and expected accuracy level of the estimates.

Internal Dependencies

These dependencies are typically between activities within your project, and you have control over these. You and your team will work together to determine these dependencies as you build the schedule.

Definitive Estimates

This is a bottom-up estimating technique that is typically performed immediately preceding the scheduled work. Sometimes known as a two-week look-ahead. The intent is the revalidate estimates on near-term work and refine the estimates now that you know who will perform the work. So definitive estimates are most closely associated with the implementation phases of a project. They are the most accurate. As subject matter experts (SMEs) help define activities and estimate who will perform the work, they also assist with determining how long activities will take. This helps in making accurate estimates. Make your assumptions and basis for your estimates explicit. Management and/or a customer will likely question your numbers. They may ask you, "is that the best you can do? Can't you complete these tasks in less time?" You'll need to communicate your rationale and defend your estimates. Then, finally, you will need to update your activities list to reflect any new or changed information.

Parametric Estimating

Uses known values and multiples them by a variable. The known values could be widely used published rates, such as building costs for new housing construction ($100/square foot) or the number of gallons of paint required (300 square foot coverage per gallon of paint). There may also be published rates for the amount of time common work efforts should take. If there isn't a published rate but you have lots of history in the form of data points, you can create your own published rates. You can use these to help estimate durations. Parametric estimating works best when there's a tangible outcome for the work. It doesn't work well with intangibles. Examples of intangible activities include performing research or estimating the amount of time it will take to edit a manuscript.

Dependencies

You may have heard terms like "dependency" or "logical relationship" when referring to sequencing activities. PMI tends to use the term dependency, and personally, I prefer it too because it accurately describes what you're trying to model. Something is dependent on something else before work can move forward. There are many types of dependencies, and they can be categorized as either: mandatory or discretionary and either external or internal.

Predecessor and Successor diagram

You will draw an arrow between the predecessor and successor to indicate what type of dependency exists. The most common dependency type is the finish-to-start (FS). In this case, the predecessor must finish before the successor can start. To draw the arrow for an FS dependency, the arrow begins on the finish side (or right side) of the predecessor box and terminates on the start side (or left side) of the successor box, as shown below. The arrow placement is the standard convention, so it's important to know it. You may have to diagram something on the exam.

Three-Point Estimating

you'll get three estimates from a single estimator. These estimates are the optimistic (best case), pessimistic (worst case), and most likely (typical) estimates. You then perform a weighted calculation. PERT is the most commonly known three-point estimating technique. The formula for PERT is: Where O = optimistic, ML = most likely, and P = pessimistic.