EVA (C/SCSC) and Basics of Project Control

Nathaniel Osgood 4/7/2004

Topics

Monitoring contd
EVA (C/SCSC) Definitions and examples Forecasting

Project Control
General Performance-adjustments Target Adjustments Problem diagnosis

Recall: Earned Value Approach (Cost/Schedule Control Systems Criteria) Definitions

Integrating cost, schedule, and work performed by ascribing monetary values to each.
Budgeted Cost of Work Scheduled (BCWS, $) (Earned value of work accomplished) the value of work scheduled to be accomplished in a given period of time. Actual Cost of Work Performed (ACWP, $): the costs actually incurred in accomplishing the work performed within the control time. Budgeted Cost of Work Performed (BCWP, $): the monetary value of the work actually performed within the control time (= Earned Value). Actual Time of Work Performed (ATWP, time) Schedule Time of Work Performed (STWP, time)

Cost Variance

Is project spending more or less money than anticipated for the work that I did? Cost Variance (CV = BCWP - ACWP)
+ (Underrun); - (Overrun); 0 (On Budget)

Cost Index

(CI = BCWP/ACWP)

> 1 (Underrun); < 1 (Overrun); 1 (On Budget)

Schedule Variance

One metric for judging if project making is progressing faster or slower than expected
More precisely: How does the value of the work I have actually performed compare to the work I anticipated performing during this time? Progress here is measured in value of the work ($)

Calculated in $ -- but here this is a proxy for value Schedule Variance (SV = BCWP - BCWS)
+ (Ahead); - (Behind); 0 (On Schedule) Even if just slightly ahead/behind in time, may be large if

working on very expensive component of project

Schedule Index

(SI = BCWP/BCWS)

> 1 (Ahead); < 1 (Behind); 1 (On Schedule)

Time Variance

Is project spending more or less time than anticipated for the work that I did? Measured in units of time May be very close even if big difference in the resource spending Time Variance (TV = STWP - ATWP)
+ (Ahead); - (Delay); 0 (On Schedule)

Time Index

(TI = STWP / ATWP)

i

> 1 (Ahead); < 1 (Delay); 1 (On Schedule)

Resource Flow Variance

Compares how much expecting to spend during this timeframe with what actually spent regardless of how much work got done. Warning: Doesnt indicate bad or good. e.g. = if
Going faster but more cheaply than expected Going slower but more expensively than expected Resource Flow Variance (RV = BCWS - ACWP)

+ (Underrun); - (Overrun); 0 (On Target)

Resource Flow Index ACWP)

(RI = BCWS /

> 1 (Underrun); < 1 (Overrun); 1 (On Target)

Earned Value Chart

Example: Gantt Chart Schedule

A B C D E F G ACTIVITY

10 15 WEEKS

20

25

Non Critical Path Activity Critical Path Activity

Example: Traditional Reporting

Example: Earned Value Reporting

Example: Activity Analysis

Example: Variances
ACTIVITY BCWP - ACWP = CV A B E $ 1,500 - $ 1,500 = $ 0 $ 3,000 - $ 3,000 = $ 0 $ 1,628 - $ 2,900 = -$ 1,272
CUMULATIVE VARIANCE = -$ 1,272

ACTIVITY BCWP - BCWS = SV A B E $ 1,500 - $ 1,200 = $ 300 $ 3,000 - $ 3,000 = $ 0 $ 1,628 - $ 3,256 = -$ 1,628
CUMULATIVE VARIANCE = -$ 1,328

Variances II
ACTIVITY A B E STWP - ATWP = TV 5-4=1 3 - 4= -1 2 - 4= -2 Cumulative Variance = -2

Example: Activity Indexes

Activity A B E

BCWP = SI BCWS 1,500 1,200 3,000 3,000 1,628 3,256 = 1.25 =1 = 0.5

BCWP = CI ACWP 1,500 1,500 3,000 3,000 1,628 2,900 =1 =1 = 0.56

Example: Project Indexes

The Aggregate Cost Index is:
SI = 1,500 + 3,000 + 1,628 1,200 + 3,000 + 3,256 = 0.82

CI =

1,500 + 3,000 + 1,628 1,500 + 3,000 + 2,900

= 0.83

Example: Earned Value Reporting

Values (in Dollars) of BCWS, BCWP, and ACWP for Weeks 1-4 Week 1 Week 2 Activity BCWS BCWP ACWP BCWS BCWP ACWP BCWS A 300 500 500 300 500 500 300 B 1,000 1,000 1,000 1,000 1,000 1,000 1,000 E 814 300 814 814 400 686 814 2,114 1,800 2,314 2,114 1,900 2,186 2,114

Week 3 BCWP ACWP 300 300 500 500 500 1,000 1,300 1,800

BCWS 300 0 814 1,114

Week 4 BCWP ACWP 200 200 500 500 428 400 1,128 1,100

Example: Earned Value Analysis

Example: Schedule and Cost Index

Schedule Index for the Project
1.00 0.90 0.80 Schedule index, SI 0.70 0.85 0.88 0.79 0.82

Cost Index for the Project

1.00 0.90 0.80 0.70 Cost Index, CI 0.60 0.50 0.40 0.30 0.20 0.10 0.78 0.82 0.79 0.83

0.60 0.50 0.40 0.30 0.20 0.10 0.00 1 2 Week 3 4

0.00

2 Week

Example: Integrating CI and SI

Topics

Monitoring contd
EVA (C/SCSC) Definitions and examples Forecasting

Project Control
General Performance-adjustments Target Adjustments Problem diagnosis

Forecasting Performance

Critical in the performance analysis process, since it can be used to identify future performance variances and design the project control process in advance of facing real performance problems Attempts to predict the conditions at a later time or the end of the project Typically made repeatedly on a regular basis throughout a project

Forecasting Completion Dates

Forecasted completion date
= Current date + (Work remaining / Expected work rate)

Expected work rate

= Expected productivity* Workers

Expected productivity
= [Work accomplished / Workers] / Time spent

Forecasting Total Costs

Forecasted total cost
= Cost spent + (Work remaining * Expected unit cost)

Expected unit cost

= Costs spent / Work accomplished

Cost Updating

Budget at Completion
BAC = Sum BCWS on lower-level OBS BAC = Sum BCWS on lower-level WBS

Work Remaining

WR = BAC BCWP ETC = Update cost for Work remaining EAC = BAC - CV or BAC / CI

Estimate to Complete

Estimate at Completion

EAC Original Estimate Approach

Estimate at Completion: EAC = ACWP + ETC Estimate to Complete: ETC = BAC BCWP EAC = ACWP + (BAC - BCWP) EAC = BAC - (BCWP - ACWP) EAC = BAC - CV

EAC Revise Estimate Approach

ACWP / BCWP = 1 / CI ETC = WR * 1 / CI ETC = (BAC

- BCWP) * 1 / CI EAC = ACWP + (BAC - BCWP) * 1 / CI EAC = ACWP + (BAC / CI) - (BCWP / CI) ACWP = BCWP / CI EAC = BAC / CI EAC = BAC * ACWP / BCWP

Example (after a month)

BCWS BCWP ACWP

= $7,456 = $6,128 = $7,400

CV SV CI SI

= $1,272 = $1,328 = 0.83 = 0.82

Original

Estimate Approach EAC = ACWP + BAC - BCWP = BAC -CV = $31,000 - (- $1,272) = $32,272 Revised Estimate Approach EAC = BAC / CI = $31,000 / 0.83 = $37,349

Beware of Delays

Financial, time indicators are necessary but not sufficient to alert to problems In most cases of serious problems and normal reporting, the problem may be very serious by the time that it is noticed in the formal reports Rapid qualitative judgment is often much more effective than delayed quantitative reporting

Topics

Monitoring contd
EVA (C/SCSC) Definitions and examples Forecasting

Project Control
General Performance-adjustments Target Adjustments Problem diagnosis

Project Control: Managing Risks

Monitoring alerts us to when theres a problem Key elements of control

Problem diagnosis (discussed later) Either

Plan correction

(often political process) Problem correction (often technical & managerial)

All of the above must be undertaken rapidly to effectively control a project

Need to see if they correct the problem and react accordingly Control without rapid monitoring highly handicapped

Value of Flexibility

Flexibility is primary defense against risk Planning too tightly may highly complicate control Already Discussed: Flexibility value to the owner

(Expandability via clearspanning, larger # conduits,

Flexibility in construction is key during control

Want enough give to change plans if necessary Usual tradeoff: Overoptimizing for cost can limit flexibility

E.g. Equipment, materials, personnel

Be careful on value engineering that limits flexibility!

The Project Control Process

Initial Targets

Project Targets
Performance Targets

Forecasting

Forecasted Performance

Performance Test
Performance Deficits and Surpluses

Target Changes

Measured Performance

Error Signal Descriptors

Error Signals

Target-Driven Control Loop

Performance-Driven Control Loop

Control Actions
Project Changes

Project Performance

Topics

Monitoring contd
EVA (C/SCSC) Definitions and examples Forecasting

Project Control
General Performance-adjustments Target Adjustments Problem diagnosis

Performance-Driven Control
Project Targets
Performance Targets

Forecasting

Forecasted Performance

Performance Test
Performance Deficits and Surpluses

Measured Performance

Error Signal Descriptors

Error Signals

Performance-Driven Control Loop

Control Actions
Project Changes

Project Performance

Performance Driven Control

Planned Versus Actual Expenditures on a Project
100% Planned Expenditure 80% Revised Estimate of Future Expenditure Actual Expenditure

PROJECT TIME (%)

60%

40%

20%

0%

0%

20%

40%

60%

80%

100%

BUDGETED EXPENDITURE (%)

Performance-Driven Control Methods

Awkward fact: Can typically only correct for one attribute of a problem at a time
Time Cost Quality

Need to understand tradeoffs and triage Most easy wins will already be in place

Exception: Sometimes new information is available that can enable improved performance now

Acceleration $ (Overtime, shift work, Rework, higher-end equipment, better crews etc.)

Attribute Linkages

Slow progress $ Delayed occupation, Higher interest on Cost Schedule const. loan Difficulty in getting financing Trying to save $ Loss of tenants Can lead to substitution, Default of parties Opportunity cost lower quality Suspension of work Selection of poor quality workers workmanship

Need for rework imposes high expenses High quality needs can lead to costly miscalculations on labor time

Schedule

Quality level impacts speed of work, Level of rework

Quality

Quality problems result from overtime, shift work, new hires

Caveats on Overreacting

When trying to correct, often bump up against other limiting factors

Space constraints Skill set breadth Hiring time Morale Coordination difficulties Confusion Cascading unanticipated effects

Often improvisation can lead to

Job rhythm and learning curves make big difference!

Schedule Performance Control

Project managers can use resources to increase work rate mainly in two ways: 1) adding new project resources (eg., Schedule Crashing) and 2) reallocating available resources (eg., Linear Scheduling Method ),

Schedule Performance Control

Change operating conditions by altering the location of the work Change operation conditions by altering the precedence, sequence, or timing of work Change the technology used Changes in the tools and methods

Project Acceleration I

Multiple-shift work
Lack of coordination Hiring Environmental/safety constraints

Overtime/Extended workdays
Fatigue Lower morale Rework

Project Acceleration II

Using larger or more productive equipment

Training/learning curve Procurement time Space constraints

Increasing # of workers
Training (takes time of most experienced!) Space constraints Hiring time Lack of knowledge of processes

Project Acceleration III

Using faster-installing materials

Procurement

Alternate construction methods

Skill set Learning curve Unknown side-effects

Project Acceleration IV

Summon on-call contractor

Learning curve Friction between teams Unknown personality situation

Activity Time-Cost Tradeoffs

Frequently we have a tradeoff b.t. $ and time

Time is money Can finish more rapidly if have

More highly skilled labor More expensive equipment More

workmen More highly paid (motivated!) labor

Project Time-Cost Tradeoff

Project Time-Cost Trade-off Curve 25 20 15 ($) 10 5 0 1 2 3 4 5 6 7 8 9 10

project penalty Overhead Total

Total cost is non-monotonic: Sometimes using less time globally NOTE: If activity time-cost curves are linear, then finding the optimal z

Link to Earlier Topic: Resource Scheduling

Recall: Earlier we discussed some resource time tradeoffs

Resource leveling Resource scheduling

At that time, we considered activity atomic: We did not consider changing activity durations/resource use profiles Time-curves often serve as a proxy for intraactivity resource reasoning

Time-Cost Tradeoffs: Key Concepts

Two components (either or both)

Reduce duration for activities on critical path

Try to increase $ as little as possible in process!

Reduce costs on activities not on critical path

Often involves increasing duration but want to keep off

Critical path!

Explicit activity time-costs tradeoffs examine direct, local activity costs only
Ignore (important) indirect costs of project extension These are global costs that depend on the entire project duration rather than activity duration

Time-Cost Trade Off Curve

A single Activity trade off curve:
20 18 16 14 12 ($) 10 8 6 4 2 0 Time
Normal Duration

Time-Cost Trade-off Curve

What duration would you choose?

Trading Money for Time Activity Crashing

Critical path tells us time-limiting activities

No benefit from reducing time of all activities up front just those on critical path NB: This is an important area in which CPM has contributed to construction understanding

Previously, many managers put effort into accelerating

whole project

Critical path may change as crashing changes activity durations

Time-Cost Algorithms

If activity time-cost curves are linear, then finding the optimal duration of the project is a linear program (LP). If not, then it is an NLP. Common assumptions
Time-Cost tradeoff is convex No binding resource constraints Normal activity cost is lowest-cost point

Kelly & Walker Crashing Heuristic

1. 2. 3. 4. 5. 6.

Solve CPM with normal durations

For critical activities: Find marginal cost of crashing (i.e., additional cost of shortening duration 1 time unit) Reduce by one time step the critical activity with the lowest marginal cost of crashing Record resulting activity project duration and cost Repeat [3] until another path becomes critical. Repeat [1] until project cost increases.

NOTE: Good, but not necessarily optimal solutions Problems? Concerns?

Issues with Heuristic

What about resource constraints?

If our preferences were determined partly by resource constraints, we are no longer guaranteed to have a legal schedule! The resulting schedule could have highly irregular (and thus costly) resource use

Number of nodes multiples as more detailed cost tradeoffs required Monotonically decreasing but non-convex timecost curves require different algorithm

Cost Performance Control

Resource use, allocation again central Effective and timely cost control is crucial to ensuring the project cost performance. It should be an on-going process, taking into consideration the following:

Change resources to remove excess capacity Change operating conditions to increase work efficiency and product quality Change methods by outsourcing different operations Re-price the work, equipment, or materials Substitute with less expensive but acceptable materials or equipment

Trading Time for Money Slack Management

Remember: Time imposes extra indirect costs! Slack Management : when budget is limited during a certain time period, rescheduling the project by changing activity timing and associated expenditure or income. Activity Timing Change: Non-critical activities first (having FF -> TF), then critical activities.
Activities having Free Float Activities having Total Float Activities on Critical path

Adapted from Pena-Mora 2003

Recall: Resource Leveling

Insight: a more steady usage of resources leads to lower resource costs. Labor: costs associated with hire, fire, and training Material: storage requirement, planning and controlling efforts Resource Leveling : the reallocation of slack (TF or FF) in non-critical activities to minimize fluctuations in the resource requirement profile.

Topics

Monitoring contd
EVA (C/SCSC) Definitions and examples Forecasting

Project Control
General Performance-adjustments Target Adjustments Problem diagnosis

Target-Driven Control: More Political Process

Initial Targets

Project Targets
Performance Targets

Forecasting

Forecasted Performance

Performance Test
Performance Deficits and Surpluses

Target Changes

Measured Performance

Error Signal Descriptors

Error Signals

Target-Driven Control Loop

Performance-Driven Control Loop

Control Actions
Project Changes

Project Performance

Topics

Monitoring contd
EVA (C/SCSC) Definitions and examples Forecasting

Project Control
General Performance-adjustments Target Adjustments Problem diagnosis