Scheduling and estimation techniques. How do you predict the future? How do you promise a finish date for a project that has never been done before? These are the tools that turn a hopeful guess into a credible, defensible plan. For the PMP exam, understanding these techniques is vital because the exam relentlessly tests your ability to pick the right tool for the right situation, not just memorise definitions.
Jump to a section
A simple way to picture Scheduling and Estimation Techniques
The head contractor of a home renovation project. They are the single person responsible for turning a vague dream of a new kitchen into a checklist of actual tasks, figuring out the order those tasks must happen, and guessing how long each one will take.
First, the contractor breaks the renovation down: demolition, new plumbing, electrical rewiring, installing cabinets, fitting countertops, painting, and floor tiling. This is the Work Breakdown Structure (WBS) — the complete list of everything that must be done. Next, they figure out sequence: you cannot tile the floor before the plumbing is in, and you cannot paint before the cabinets are hung. This is the schedule network diagram, showing which tasks depend on which others.
Then the contractor estimates durations. The plumber says the new pipes take three days. The electrician says two days. But the contractor knows the plumber and electrician need the same wall space, so they cannot work at the same time. The contractor must add buffer time for when materials arrive late or when a wall hides rotten wood. This is analogous to a project manager adding 'contingency reserve' to cope with known unknowns.
Finally, the contractor creates a calendar that shows the whole project will take six weeks, not two. They have a critical path — the longest chain of dependent tasks that, if any one slips, delays the entire project. That is the fundamental structure of scheduling and estimation: decompose, sequence, estimate, and manage the critical path.
Scheduling and estimation are the twin engines of project planning. Without them, a project is just a wish. The goal is to answer two hard questions: 'How long will this take?' and 'When will it be done?'. The PMBOK Guide describes several distinct processes and tools to answer these questions.
First, let us define the key terms. An 'activity' is a single piece of work that has a duration, like 'write login page' or 'paint bedroom'. A 'milestone' is a zero-duration marker that indicates a significant event, like 'design approved' or 'code complete'. The 'Work Breakdown Structure' (WBS) is the hierarchical breakdown of all work into manageable chunks. Think of it as a numbered outline of every deliverable.
Once you have the WBS, you sequence the activities. This creates a 'schedule network diagram', which looks like a flow chart of boxes (activities) connected by arrows (dependencies). There are four dependency types:
Finish-to-Start (FS): Task B cannot start until Task A finishes. This is the most common. For example, you cannot paint a wall until the drywall is installed.
Start-to-Start (SS): Task B cannot start until Task A starts. For example, you might start writing documentation as soon as you start coding, but not before.
Finish-to-Finish (FF): Task B cannot finish until Task A finishes. For example, testing cannot finish until all coding finishes.
Start-to-Finish (SF): Task B cannot finish until Task A starts. This is rare but exists.
After sequencing, you estimate the duration of each activity. Three primary techniques exist for estimation:
'Analogous Estimating': You look at a similar past project and say, 'That login feature took two weeks, so this one will also take two weeks.' It is fast but inaccurate if the projects are not truly similar.
'Parametric Estimating': You use a statistical relationship. For example, you know a painter paints 10 square metres per hour. If the wall is 100 square metres, the estimate is 10 hours. This is more accurate if the parameters are reliable.
'Three-Point Estimating' (Program Evaluation and Review Technique — PERT): You provide three estimates: optimistic (best case), pessimistic (worst case), and most likely. The formula is (Optimistic + 4 * Most Likely + Pessimistic) / 6. This gives a weighted average that accounts for risk.
Now you have durations and sequences. You input these into the network diagram and calculate the 'Critical Path'. The critical path is the longest sequence of activities from start to finish. Any delay on the critical path directly delays the project end date. Activities not on the critical path have 'float' or 'slack' — time they can be delayed without affecting the final date.
A key technique used here is 'Critical Path Method' (CPM). It calculates the earliest and latest each activity can start and finish without delaying the project. Activities with zero float are on the critical path. If you have two critical paths, the project is high risk.
Another technique is 'Resource Leveling'. Sometimes your plan says two activities happen at the same time, but you only have one person to do both. Resource leveling shifts the schedule to avoid overloading resources, but it often extends the project end date.
There is also 'Fast Tracking': doing sequential activities in parallel. For example, starting testing before all coding is finished. This adds risk of rework but can compress the schedule. 'Crashing' is adding more resources (like hiring extra staff) to shorten a task. This usually increases cost.
The PMP exam also distinguishes between 'duration' (calendar time an activity takes, including weekends and holidays) and 'effort' (the actual number of person-hours of work). A task might require 40 hours of effort but take two weeks of duration if the person works part-time.
All these techniques exist because without them, schedules are based on hope or pressure from stakeholders. The structured approach gives the project manager data to defend the plan and to make trade-off decisions when things go wrong.
Decompose Scope into Activities
Start with the Work Breakdown Structure (WBS). Break each work package into individual activities (tasks) that have a duration. For example, the work package 'User Login' becomes 'design login page', 'build login API', 'write unit tests', and 'integration test'. This step ensures every piece of work is accounted for.
Sequence Activities with Dependencies
Decide the order of activities using dependency types (FS, SS, FF, SF). Create a network diagram showing arrows between tasks. This reveals which tasks can be done in parallel and which must wait. The goal is to model the realistic sequence of work, including lead and lag where appropriate.
Estimate Activity Resources
Determine what resources (people, equipment, materials) are needed for each activity. For example, 'build login API' needs one senior developer and one test environment. This step informs duration and cost. Without knowing resources, duration estimates are meaningless.
Estimate Activity Durations
Apply estimation techniques (analogous, parametric, three-point) to assign a duration to each activity. Factor in resource availability, holidays, and productivity. This step produces the raw numbers that feed into the schedule.
Develop the Schedule
Input all activities, dependencies, resources, and durations into scheduling software or a manual chart. Apply CPM to find the critical path and float. Perform resource leveling or smoothing if needed. The output is a final schedule baseline — the committed plan against which progress will be measured.
Control the Schedule
During execution, track actual start and finish dates against the baseline. If a critical path task slips, take corrective action (crashing or fast tracking). Use earned value management (EVM) to calculate schedule variance and SPI. This step ensures the project stays on track or adjusts proactively.
A real IT project manager works with these techniques daily, even if they use different names. Consider a mid-size software company building a new customer portal. The company has a team of four developers, one designer, one tester, and one product owner. The project manager is responsible for the schedule.
The first task is breaking the work down. The PM sits with the team and creates a WBS. High-level items might be 'User Authentication', 'Dashboard', 'Payment Integration', and 'Testing'. Each of these is broken down further. 'User Authentication' becomes 'design login screen', 'build login API', 'write unit tests', and 'integration test'. Each leaf item is a work package that can be estimated.
Next, the PM creates the schedule network diagram. The team identifies dependencies:
'Design login screen' must finish before 'build login API' can start (Finish-to-Start).
'Write unit tests' can start only after 'build login API' starts (Start-to-Start), because tests are written against code that exists.
Testing in general cannot finish until all feature coding is done (Finish-to-Finish).
The PM then estimates. For 'build login API', a developer says it is similar to a feature built five months ago. The PM uses analogous estimating and says two weeks. But the developer is new to the team and the technology is slightly different. To be safe, the PM asks for a three-point estimate. The developer says optimistic 5 days, pessimistic 15 days, most likely 10 days. Using the PERT formula: (5 + 4*10 + 15)/6 = 10 days exactly. The PM records that.
Now the PM enters all tasks, dependencies, and durations into Microsoft Project or Jira. The software calculates the critical path. In this case, the critical path runs through: 'design login screen' -> 'build login API' -> 'integration test payment' -> 'user acceptance testing'. This sequence takes 45 days. The PM sees that 'write documentation' (30 days float) and 'create marketing page' (20 days float) are more flexible.
The PM then presents the schedule to stakeholders. The product owner says the launch date must be 40 days, not 45. The PM suggests options. 'Fast tracking' is risky but possible: start integration testing of payment before the login API is fully done, saving 5 days. Or 'crashing' by hiring a freelance developer to help with the login API, which costs more money but saves 3 days. The PM presents these trade-offs clearly.
During execution, a developer gets sick. The login API task slips by three days. Because it is on the critical path, the project end date slips by three days unless the PM takes action. The PM decides to crash by paying overtime to one healthy developer, getting back two of the three lost days. When the exam tests you, it will expect you to identify that only critical path delays matter for the end date, and that crashing increases cost while fast tracking increases risk.
The PMP exam tests your ability to choose the correct technique based on scenario details, not just to define them. Here is what they love to focus on.
First, the distinction between 'known unknowns' and 'unknown unknowns'. Known unknowns are risks you can identify and quantify, so you add 'contingency reserve' (time or money) to the schedule. Unknown unknowns are things you cannot foresee, so you add 'management reserve' (a separate buffer not included in the baseline schedule). The exam will give you a scenario involving 'unexpected regulatory change' — that is a known unknown if you knew it might happen. If it is truly unforeseeable, it is an unknown unknown.
Second, the difference between 'duration' and 'effort' is tested frequently. They will give a task: '40 hours of work over two weeks'. They ask: what is the duration if the person works 5 hours per day? The answer is 8 days of duration, not 40 hours of effort. They love to make you calculate using calendar constraints.
Third, calculating float on the critical path. The critical path always has zero float. But the exam may give a network diagram with multiple paths. You must calculate the float for each activity. The formula is Float = Late Start - Early Start, or Late Finish - Early Finish. If you get a negative float, the project is behind schedule.
Fourth, the difference between 'Lead' and 'Lag'. Lead is overlapping two tasks — the second task starts a few days before the first finishes. Lag is a waiting time — the second cannot start until a set time after the first finishes. For example, after pouring concrete, you must wait 3 days for it to cure. That is lag of 3 days.
Fifth, they will test which estimation technique to use in which situation. The traps include:
When a project is new and no historical data exists: use 'Bottom-Up Estimating' (estimating individual work packages and summing them up). This is the most accurate but most time-consuming.
When you need a rough order of magnitude early on: use 'Analogous Estimating' (comparing to past projects).
When the scope is somewhat known but not fully defined: use 'Three-Point Estimating'.
When you have a strong statistical relationship between variables: use 'Parametric Estimating'.
Key definitions to memorise for the exam:
'Critical Path Method' (CPM): calculates longest path and float.
'Program Evaluation and Review Technique' (PERT): uses three estimates for risk.
'Schedule Compression': includes fast tracking (parallel) and crashing (adding resources).
'Resource Leveling': adjusts schedule to manage resource constraints, often extending duration.
'Resource Smoothing': adjusts schedule within the float, not extending the end date.
The exam also tests 'Variance Analysis'. They give you a baseline schedule and a progress report. They ask: is the schedule ahead or behind? The answer uses Earned Value Management (EVM) formulas like Schedule Variance (SV = EV - PV) or Schedule Performance Index (SPI = EV / PV). A negative SV or an SPI less than 1.0 means behind schedule.
Finally, they will test the 'Murphy's Law' scenario: if anything can go wrong, it will. They ask how to protect the schedule. The answer is to use contingency reserves, risk response planning, and buffer management.
The Critical Path is the longest sequence of dependent tasks; any delay on that path delays the entire project.
Float (or Slack) is the amount of time a non-critical task can slip without affecting the project end date.
PERT (Three-Point Estimating) uses an average of optimistic, pessimistic, and most likely estimates to account for uncertainty.
Crashing adds resources to shorten a task and increases cost; fast tracking runs tasks in parallel and increases risk.
Resource leveling can extend the schedule; resource smoothing does not change the end date.
Contingency reserve is added to the schedule baseline for known unknowns; management reserve is separate for unknown unknowns.
Duration is calendar time including off-days; effort is the actual person-hours of work.
The Work Breakdown Structure (WBS) is the foundation for all scheduling as it defines every activity that must be estimated.
Analogous estimating is fast and cheap but less accurate; bottom-up estimating is slow and expensive but most accurate.
Lead allows tasks to overlap; lag enforces a mandatory waiting period between tasks.
These come up on the exam all the time. Here's how to tell them apart.
Critical Path Method (CPM)
Focuses on task dependencies and float.
Assumes unlimited resources; does not explicitly manage resource contention.
Uses float as the buffer for schedule flexibility.
Critical Chain Method
Focuses on resource constraints and adds buffers.
Accounts for resource availability and adds a project buffer at the end.
Uses feeding buffers for non-critical paths to protect the critical chain.
Fast Tracking
Overlaps sequential tasks to run in parallel.
Increases risk of rework because later tasks start before earlier ones finish.
Does not necessarily increase cost, but may increase coordination effort.
Crashing
Adds more resources to a task to shorten its duration.
Increases cost due to overtime, hiring, or expediting.
Typically does not change task sequence; it compresses individual tasks.
Analogous Estimating
Uses historical data from a similar project as a reference.
Less accurate; relies on similarity which may be subjective.
Fast to perform with minimal data.
Parametric Estimating
Uses statistical relationship between variables (e.g., lines of code per hour).
More accurate if the parameters and historical data are reliable.
Requires reliable metrics (e.g., productivity rates) to calculate.
Resource Leveling
Adjusts the schedule to prevent resource over-allocation.
Can extend the project end date because tasks are moved later.
Useful when resource availability is a critical constraint.
Resource Smoothing
Adjusts activities within their float to balance resource usage.
Does not change the project end date; only uses available slack.
Useful when the end date is fixed and resources are slightly overbooked.
Mistake
If an activity is not on the critical path, its delay never matters.
Correct
If the delay on a non-critical activity exceeds its available float, that activity becomes part of a new critical path and will delay the end date.
Beginners think float is unlimited, but float is a finite buffer. Running out of float is the same as being on the critical path.
Mistake
Three-point estimating (PERT) guarantees the estimate will be accurate.
Correct
PERT provides a statistically weighted average that is more reliable than a single guess, but it is still an estimate, not a guarantee. Actual duration can still differ.
The formula feels mathematical and precise, so people trust it too much. The exam tests that it is a 'probabilistic estimate', not a certainty.
Mistake
Resource leveling always shortens the project schedule.
Correct
Resource leveling often extends the schedule because it prevents resource over-allocation by moving tasks later. Resource smoothing is the technique that keeps the end date while adjusting within float.
The word 'leveling' sounds like 'levelling out problems' which implies faster flow. The exam explicitly tests the difference between leveling and smoothing.
Mistake
Crashing always reduces cost because the project finishes sooner.
Correct
Crashing typically increases cost because it involves paying for overtime, hiring extra staff, or expediting shipping. It is a cost-schedule trade-off, not a cost reduction.
Intuitively, finishing faster seems cheaper. But the exam teaches that crashing incurs extra expense.
Mistake
Fast tracking always reduces risk because work is done earlier.
Correct
Fast tracking increases risk by starting later tasks before earlier ones are complete, creating rework if the earlier task changes. It is a risk-schedule trade-off.
Doing things in parallel intuitively sounds more efficient. The exam shows that parallel work carries hidden risk of rework.
Reveal each answer, then mark whether you got it right. Score 60%+ to unlock the next chapter.
Critical Path Method (CPM) focuses on task dependencies and float. Critical Chain Method focuses on resource constraints and adds buffers (feeding buffers and project buffer) to protect the schedule. They are different approaches to managing uncertainty.
Float (or slack) is the amount of time a task can be delayed without affecting the project end date. It is calculated as Late Finish minus Early Finish, or Late Start minus Early Start. Tasks on the critical path have zero float.
No. Use analogous estimating when you have little detail and need a quick rough estimate early in the project. Use three-point estimating when you have some detail and need to account for uncertainty. The best technique depends on the situation and available data.
Use resource leveling when resources are over-allocated and you are willing to let the schedule slip to avoid burning out the team. Use resource smoothing when you want to keep the end date fixed and can adjust tasks within their float to balance workload.
Lag is a mandatory waiting period between two tasks (e.g., wait 3 days after painting for paint to dry before applying a second coat). Lead is when the second task starts a few days before the first finishes (e.g., start testing before coding is 100% complete).
Contingency reserve is included in the schedule baseline for identified risks (known unknowns). Management reserve is a separate buffer held by management for unforeseen risks (unknown unknowns) and is not part of the baseline schedule.
You've finished Scheduling and Estimation Techniques. Continue through the PMP study guide to build a complete picture of the exam.
Done with this chapter?