2. PMI-002 11/5 11/6/03 9:20 AM Page 13
There has been a high degree of • A c tual c o s t (A C ): This is the cumula- this work package is EV = $ 20,000 x
EVM acceptance among current and tive AC spent to a given point in time 1.00 = $ 20,000. W ork package 1.2 has
past users of the method. They tend to accomplish an activity, work pack- a total budget of $ 4 0,000 and is 5 0%
to agree that EVM can improve cost, age, or project and to earn the related complete as of the status date.
schedule, and technical performance value. This was previously called the Therefore, the earned value for this
of their projects. EVM nonusers indi- actual cost of work performed work package is EV = $ 4 0,000 x 0.5 0 =
cate that the method is hard to use, (ACW P). Figure 2 illustrates a project $ 20,000. The earned value for the
that it applies primarily to federal in which the planned value as of the entire project is EV = $ 20,000 +
projects, and that they do not need it project status date is PV = $ 5 0,000 and $ 20,000 = $ 4 0,000.
(Fleming & Koppelman, 2000; the actual cost is AC = $ 60,000. The preceding formula converts
Kim, 2000). project accomplishments from physi-
This paper simplifies EVM and cal units of measure, e.g., cubic yards
shows its applicability to public and C o st Planned of concrete, linear feet of cable, percent
($ 0 0 0 ) V alu e (PV )
private sector projects, regardless of Budget At complete, milestones achieved, or
siz e. The paper uses the simplified ter- C o m p letio n deliverables completed, to financial
60 (BAC )
minology and provides graphical 50 units of measure. These financial
A c tu al C o s t
tools, extensions, and applications of (A C ) measurements of value can be in dol-
S tatu s D ate
EVM to enhance the use and effective- lars (or any other currency), labor
T im e
ness of this important project manage- hours, work hours, worker days, or any
ment method. other similar quantity that can be used
F igure 2 . Planned Value and A c tual C o s t as a common measurement of the
EVM Key Components value and cost associated with project
EVM uses the following project param- • E ar ned value (E V): This is the cumu- work. Figure 3 illustrates the above
eters to evaluate project performance: lative earned value for the work com- project, in which the total budget at
• Planned value (PV): This is the time- pleted up to a point in time. It completion is B AC = $ 100,000, the
phased budget baseline (Figure 1). It is represents the amount budgeted for planned value as of the status date is
the approved budget for accomplish- performing the work that was accom- PV = $ 5 0,000, the actual cost is AC =
ing the activity, work package, or proj- plished by a given point in time. This $ 60,000, and the earned value is EV =
ect related to the schedule. It can be was previously called the budgeted $ 4 0,000. These are the main basic enti-
viewed as the value to be earned as a cost of work performed (B CW P). To ties in EVM.
function of project work accomplish- obtain EV for an item, multiply its
ments up to a given point in time. This total budget by its completed propor- P erformance Measurement
graph of cumulative PV is often tion. Table 1 shows the work break- Cost performance is determined by
referred to as the S-curve (because, down structure (W B S) of a project with comparing the EV to the AC of the
with a little imagination, it looks like a total budget of $ 100,000. W ork pack- activity, work package, or project.
the letter S, or as an abbreviation of the age 1.1 has a total budget of $ 20,000 Schedule performance is determined
Spending-curve). This was previously and is 100% complete as of the status by comparing the EV to the PV. This
called the budgeted cost of work date. Therefore, the earned value for can be accomplished by calculating the
scheduled (B CW S).
• B udg et at c o m p letio n (B A C ): This is ($ 0 0 0 )
the total budget baseline for the activi-
ty, work package, or project (Figure 1). Budget % C o m p lete E a rn ed
P ro jec t
It is the highest value of PV and the last V a lue
P h a se1
point on the cumulative PV curve.
W o rk Pac k ag e 1 .1 20 100 20
W o rk Pac k ag e 1 .2 40 50 20
........
C o st
Planned P h a se2
v alu e (PV ) Budget At
C o m p letio n W o rk Pac k ag e 2 .1 ...
(BAC )
W o rk Pac k ag e 2 .2 ...
........ ...
........ ...
T im e
T o ta l 100 40
F igure 1 . Planned Value and Budget
at C o m p letio n T ab le 1 . W BS , Budget, % C o m p lete, and E ar ned Value
December 2003 Project Management Journal • 13
3. PMI-002 11/5 11/6/03 9:20 AM Page 14
PV R ate = $100,000 / 40 For the above project, SVP =
Cost Planned
= $2,500 per week -$10,000 / $50,000 = -20 %. This
($000) Value (PV) TV = -$10,000 / $2,500 = means that the project is 20%
Budget At
Completion -4 weeks behind schedule.
60 (BAC)
50
40
Actual Cost TV measurement also can be per-
(AC) E arned Status date
Value (E V) formed and reported graphically. This 40
CV and S V ($000)
Time is accomplished by drawing a hori- 20 G ood
zontal line from the intersection of 0
SV
Figure 3 . Planned Value, Actual Cost and the EV curve with the status date to the -20 Poor
CV
Earned Value
PV curve and reading the distance on -40
variances, the variance percentages, the horizontal time axis (Fleming &
Time
and the performance indices at the Koppelman, 2000), as shown in
desired levels of the WBS. It is interest- Figure 4.
ing to note that these comparisons are In the above formulas, 0 indicates Figure 5 . CV and SV G raph
made to the EV, rather than to the that performance is on target. A posi- H owever, it may be appropriate
baseline PV. tive value indicates good perform- to use EV rather than PV in the
It is important to synchronize the ance. A negative value indicates poor denominator of this formula (J. J.
status date for data in the analysis. This performance. Moder in Cleland & King, 1988). The
can be accomplished by using the con- SVP based on the earned value
cept of accrued cost, which includes (SVev% or SVPev) would be defined
expenditures made but not yet reflect- as: SVPev = SV / EV. For the above
ed in the financial system, to accom- Cost Planned project, SVPev = -$10,000 / $40,000 =
($000) Value (PV)
plish work up to the status date. Budget At -25%. This indicates that the project is
Completion
60 (BAC) 25% behind schedule.
50 SV = -10 CV = -20
Variances 40
Actual Cost
The following formulas are used to cal- (AC) E arned Status date
Value (E V)
culate the variances, generally based Time
4
on cumulative data, also called incep-
2 G ood
TV (W eek s)
tion-to-date data and project-to-date Figure 4 . Variances 0
data (Figure 4, using the data from the
-2 Poor
above project): G raph ical D isplays
-4 TV
The cost variance (CV) is a meas- G raphs of variances over time provide
ure of the budgetary conformance of valuable indicators of trends in project Time
actual cost of work performed: CV = performance and of the impact of any
EV – AC. For the above project, CV = corrective actions (Figures 5 and 6). Figure 6 . T V G raph
$40,000 – $60,000 = -$20,000.
The schedule variance (SV) is a Variance Percentages SVPev is consistent with the for-
measure of the conformance of actual The following formulas are used to mula for CVP. It points out that SV
progress to the schedule: S V = EV – PV. calculate the variance percentages, occurred while accomplishing EV.
For the above project, SV = $40,000 – generally based on cumulative data Therefore, it may be a better indicator
$50,000 = -$10,000. (Figure 4, using the data from the of project schedule status, as shown
T ime variance: The average AC per above project): later in the calculation of the time esti-
time period is often called the spend The cost variance percent (CV% or mate at completion (TEAC).
rate or burn rate. Similarly, the average CVP) is a measure of the budgetary In the above formulas, 0 indicates
PV per time period can be called the conformance of actual cost of work that performance is on target. A posi-
planned accomplishment rate, performed: CVP = CV / EV. For the tive value indicates good perform-
planned value rate, or the PV rate. It is above project, CVP = -$20,000 / ance. A negative value indicates poor
defined as the baseline BAC divided by $40,000 = -50%, which indicates that performance.
the baseline schedule at completion the project is 50% over budget.
(SAC). As a formula, PV R ate = BAC / The schedule variance percent Performance Indices
SAC. Thus, SV can be translated into (SV% or SVP) is a measure of the con- The following formulas are used to cal-
time units by dividing SV by the PV formance of actual progress to the culate the performance indices, general-
R ate. The result is the SV in time units schedule. The following formula has ly based on cumulative data (Figure 4,
or the TV. As a formula, T V = SV / PV been generally used to calculate it using the data from the above project):
R ate. If the above project were sched- (Project Management Institute, 2000): The cost performance index (CPI)
uled for forty weeks, then: S VP = SV / PV. is a measure of the budgetary confor-
14 • Project Management Journal December 2003
4. PMI-002 11/5 11/6/03 9:20 AM Page 15
mance of actual cost of work per- ures of project performance and to
formed: CPI = EV / AC. For the above 1.4 quantify the “ traffic light approach.”
project, CPI = $40,000 / $60,000 = 1.2 Good We can include the line that indicates
CPI and SPI
0.67. 1.0 on target performance with the area
The schedule performance index .8 0 SPI Poor that indicates good performance and
(SPI) is a measure of the conformance .60 CPI use the color green to indicate on tar-
of actual progress to the schedule: SPI get and good (better than target) per-
Time
= EV / PV. For the above project, SPI = formance. We can break the poor
$40,000 / $50,000 = 0.80. performance area into two and use
Figure 7 . CPI and SPI Graph
Performance indices can be the color yellow to indicate somewhat
thought of as efficiency ratios. In the SPI being close to target, or, if one of below target performance, and the
above formulas, 1.00 indicates that these indices suggests poor perform- color red to indicate poor perform-
performance is efficient and on target. ance, the other must be indicating ance. It is important for the organiza-
More than 1.00 indicates excellent, good performance. This allows some tion to carefully establish meaningful
highly efficient performance, and less trade-offs to reach the desired project thresholds, acceptable tolerances, or
than 1.00 indicates poor, inefficient goals. critical limits for action on project
performance. A CR of more than 1.00 indicates performance. This helps ensure that
The inverse of the formulas given that the overall project performance is when action is needed, it is highlight-
above has also been used (Anbari, excellent. This may result from both ed, and when action is not needed,
1980; Egan, 1982; Cioffi, 2002; the CPI and SPI being better than tar- tampering and micromanagement are
Webster, 2002). This facilitates use of get, or, if one of these indices is indi- minimized.
the indices in forecasting. Using the cating poor performance, the other For example, performance indices
inverse definition, the CPI for the must be indicating outstanding per- and critical ratios of 1.00 or above can
above project would be $60,000 / formance. This allows extensive trade- be considered green; performance
$40,000 = 1.50, indicating that the offs to reach the desired project goals. indices and critical ratios equal to or
project is running 50% over budget. A CR of less than 1.00 indicates greater than 0.80, but less than 1.00,
Completion of the project would be that the overall project performance is can be considered yellow; and per-
forecasted at $150,000, if performance poor. This may result from both the formance indices below 0.80 can be
continues at this rate. Similarly, the SPI CPI and SPI being worse than target, considered red. In this paper, a black
would be $50,000 / $40,000 = 1.25, or, if one of these indices suggests and white chart depicting this concept
indicating the project is running 25% good performance, the other must be is shown in Figure 9, and is called the
behind schedule. The project would be indicating extremely poor perform- target performance chart. It can be pro-
forecasted to take 25% longer than the ance. This limits the use of effective duced in color and may also be nick-
original schedule, with completion at trade-offs, and highlights significant named the rainbow chart.
1.25 x 40 weeks = 50 weeks, if per- difficulty in attempting to reach the O ther colors can also be added.
formance continues at this rate. These desired project goals. For example, orange or amber can be
forecasts are discussed in more detail A graph of the critical ratio over used between yellow and red, or in the
in the forecasting section of this paper. time provides a quick indicator of yellow area to indicate that the item in
Graphs of performance indices trends in the overall project perform- trouble has been previously reviewed.
over time provide valuable indicators ance, and of the impact of any correc- Blue can be used to indicate the super-
of trends in project performance and tive actions. These graphs may be very stars— items with performance indices
the impact of any corrective actions. effective in project reviews (Figure 8). above 1.20, for example (Figure 9).
These graphs can be very effective in Some may say that such superstar
project reviews (Figure 7). items must have had inflated baseline
budgets and schedules. However, there
1.4
The Critical Ratio Good
may be important lessons to be
1.2
The critical ratio (CR) is the product of learned from these items in terms of
1.0
CR
CPI and SPI (Anbari, 2001; L ewis, estimating, budgeting, performance
.8 0 Poor
2001). It can also be called the cost- management, and cost control.
.60
schedule index (CSI) (Barr, 1996; CR Reallocation of organizational
Meredith & Mantel, 2000). It is used as Time resources may be another outcome
an indicator of the overall project from such analyses (L ewis, 2001).
health: CR = CPI x SPI. For the above Figure 8 . CR Graph An activity, work package, or proj-
project, CR = 0.67 x 0.80 = 0.53. ect should be carefully reviewed when
A CR of 1.00 indicates that the Q uantifying the Traffic L ight A pproach it enters the yellow zone, with the
overall project performance is on tar- Graphs of CPI, SPI, and CR can be intent of finding the root cause(s) of
get. This may result from both CPI and used to further highlight these meas- performance or planning problems
December 2003 Project Management Journal • 15
5. PMI-002 11/5 11/6/03 9:20 AM Page 16
and eliminating them. When an item affecting the activity, work package, or ciencies observed to date will prevail to
in the red zone is reviewed, this should project, a new ETC needs to be devel- completion, the EAC3 is the sum of
generally be a status report on oped; EAC1 is the sum of the cumula- the cumulative AC plus the original
action(s) taken or not taken when that tive AC plus the ETC. As a formula, budget for the remaining work (BAC –
item was in the yellow zone. When an EAC1 = AC + ETC. For the example EV), modified by a performance factor,
item enters the blue zone, it also project used in this paper, EAC1 = which is usually the cumulative CPI.
would be appropriate to review it, to $60,000 + ETC. This applies where As a formula, EAC3 = AC + (BAC – EV)
obtain information on the root ETC is developed for the remaining / CPI. For the above project:
cause(s) of the super performance, and work. EAC1 may also be called the EAC3 = $60,000 + ($100,000 –
incorporate the lessons learned into revised cost estimate (RCE), latest $40,000) / 0.67
future work. revised estimate (LRE), or current = $60,000 + $60,000 / 0.67
working estimate (CWE). = $60,000 + $90,000
Using the above assumption, the = $150,000
ETC for the remainder of the activity, The above formula can be
1.4
work package, or project usually is simplified as follows:
CPI, SPI, and CR
Super Stars
1.2
G ood developed using various cost estimat- EAC3 = AC + (BAC – EV) / CPI
1.0
Caution ing methods. Because the work already = AC + BAC / CPI – EV / CPI
.80
SPI is in progress, a detailed, bottom-up = AC + BAC / CPI – AC
.60 CPI Poor
CR
cost estimate for the remaining work is = BAC / CPI
Time common in this case. Thus:
When current analysis shows that EAC3 = BAC / CPI
Figure 9 . Target Performance Chart past performance is not a good predic- The definition of EAC3 can there-
tor of future performance, that prob- fore be simplified to equal the original
Forecasting lems or opportunities which affected BAC divided by the CPI. For the above
Project management is primarily con- performance in the past will not occur project, EAC3 = $100,000 / 0.67 =
cerned with decisions affecting the in the future, and that future perform- $150,000. EAC3 may also be called the
future. Therefore, forecasting and pre- ance will parallel the original plan, the statistical estimate at completion
diction are extremely important EAC2 is the sum of the cumulative AC (EACstat), the mathematical estimate at
aspects of project management. EVM is plus the original budget for the completion (EACmath), or simply the
particularly useful in forecasting the remaining work (BAC – EV): EAC2 = cost at completion (CAC).
cost and time of the project at comple- AC + BAC – EV. For the above project, Using the above assumption, the
tion, based on actual performance up EAC2 = $60,000 + $100,000 – $40,000 estimated cost to complete the remain-
to any given point in the project. = $120,000. der of the activity, work package, or
The above formula can be simpli- project is the original budget for the
Forecasting of Cost at Completion fied as follows: remaining work divided by the CPI. As
The EAC may also be called cost esti- EAC2 = AC + BAC – EV a formula, ETC = (BAC – EV) / CPI.
mate at completion (CEAC). The esti- = BAC + (AC – EV) This may be called statistical estimate
mated cost to complete the remainder = BAC – (EV – AC) to complete (ETCstat) or the mathe-
of the project is usually called the esti- = BAC – CV matical estimate to complete
mate to complete (ETC). Both can be Thus: (ETCmath).
developed using various cost estimat- EAC2 = BAC – CV A graph of the EAC over time
ing methods or calculated mathemati- The definition of EAC2 can there- provides a valuable indicator of
cally using EVM. fore be simplified to equal the original trends in project cost performance
EACs may differ based on the baseline BAC minus the CV. For the and the impact of any corrective
assumptions made about future per- above project, EAC2 = $100,000 – (- actions. This graph can be particular-
formance. The PMBOK® Guide (Project $20,000) = $100,000 + $20,000 = ly effective in project reviews. Figure
Management Institute, 2000) provides $120,000. 10 shows a graph of EAC for the
three such estimates, based on three Using the above assumption, the example project used in this paper,
different assumptions. In this section, ETC for the remainder of the activity, using the above assumption.
these estimates are reviewed, simpli- work package, or project is the original
fied and enhanced. They are given a budget for the remaining work (BAC – Additional Forecasts of Cost
sequential subscript to differentiate EV). at Completion
among them. When current analysis shows that Other assumptions can be made about
When current analysis shows that past performance is a good predictor future performance and may result in
the assumptions underlying the origi- of future performance, that perform- different estimates at completion. In
nal estimate are flawed, or no longer ance to date will continue into the this section, other assumptions and
applicable due to changed conditions future, and that efficiencies or ineffi- the resulting EACs are presented. They
16 • Project Management Journal December 2003
6. PMI-002 11/5 11/6/03 9:20 AM Page 17
it acknowledges that cost management higher than the original baseline BAC.
160 and schedule management are insepa- As a formula, EAC6 > > BAC. This esti-
EAC
140 rable (Kerzner, 2001). As examples: mate is generally not quantified, but is
Poor Project schedules can be crashed at an referred to by project team members
EAC ($000)
120
BAC
100 additional cost, or less skilled with statements such as: “If you think
80 Good resources may be used on the project, this is bad, wait till you see the next
60 which may reduce the cost and possi- report! Y ou ain’t seen nothing yet! ” or
bly extend the duration. “The cost is going sky high. If this
Time The assumption implied by the project ever finishes, it would be a
above formula is that if the activity, miracle! ”
Figure 10. EAC Graph
work package, or project were behind This case may result from delaying
are given a continuing sequential sub- schedule, additional cost would be corrective action and believing for too
script to differentiate among them. incurred to bring the project back on long that the actual cost at completion
In some organizations, it is com- schedule, through the use of overtime, somehow would end up close to the
mon to state that the activity, work additional resources, expediting ship- original baseline BAC, regardless of
package, or project will meet the orig- ments, and similar actions. On the prior poor performance. Higher costs,
inal targets upon completion, regard- other hand, if the activity, work pack- lower levels of accomplishment, and
less of prior performance. This age, or project were ahead of schedule, inefficient spending patterns become
frequently occurs early in the project opportunities for significant cost sav- practically irreversible and the project’s
when prior performance has been ings may be pursued, although they fate is sealed. Statistics of challenged
poor. The EAC4 would be the original may require more time as a result of and failed projects testify that this case
baseline BAC. As a formula, EAC4 = using resources that are fewer in num- is much more common than we would
BAC. Statements such as the following ber, less experienced, and/or less like to believe.
may be heard: “We had some mobi- skilled. Additional time may also be
lization problems, but we took care of required to find better prices for equip- EACs
them. We expect the project to finish ment and material, negotiate better 160
on schedule and on budget.” or “The contract terms, use more economical 140
EACs ($000)
original specs were unclear. So we shipping methods, or take similar 120 Poor
BAC
took additional time to clarify them. actions. This formula may provide a 100
We are planning to meet project tar- better indication of estimated cost at 80 Good
gets at this time.” completion, when adherence to a 60
The above statements should be schedule is critical to the organization.
Time
challenged firmly, with a response Using the earlier definition of CR
such as: “What we hear you say is that = CPI x SPI, and further defining EAC5
Figure 11. EACs Graph
future performance will be so much or EACs as the EAC adjusted for sched-
better than the original plan and will ule performance, the above formula The Standish Group conducted
make up for prior cost overruns (and can be restated as follows: EAC5 = surveys and interviews to explore what
delays). So far, we have not performed EACs = BAC / CR. For the above proj- causes information technology (IT)
to the original plan and would like to ect, EAC5 = EACs = $100,000 / 0.53 = software development projects to be
know how this superior performance $187,500. challenged and why these projects fail.
will be achieved.” Using the above assumption, the These studies classified projects into
EAC 4 is rarely achieved. ETC for the remainder of the activity, three types:
Unmanaged projects do not fix them- work package, or project is the original Successful: The project is complet-
selves. They only tend to overrun budget for the remaining work divided ed on time and on budget, with all fea-
their budgets, fall behind their sched- by the CR: (BAC – EV) / CR. This may tures and functions as originally
ules, and often miss other scope and be called the ETC adjusted for schedule specified;
quality targets. performance (ETCs). A graph of the Challenged: The project is com-
Heinze (1996) provides the follow- EACs over time provides a valuable pleted and operational but is over
ing additional formula for calculating indicator of trends in project cost per- budget, beyond the time estimate, and
the EAC: EAC = BAC / CPI x SPI. formance and the impact of any cor- offers fewer features and functions
Fleming & Koppelman (2000) provide rective actions. This graph can be very than initially specified;
a similar formula and support it by effective in project reviews. Figure 11 Failed: The project is canceled
indicating that there is a human ten- shows a graph of EACs for the example before completion.
dency to get back on schedule, even if project used in this paper, using the The Standish Group study con-
that requires more resources for the above assumption. ducted in 1994 and published in 1995
same work. The above formula may be A case that is not often mentioned (The Standish Group, 1995) had a
mathematically questionable. However, occurs when the EAC6 is substantially total sample of 365 respondents repre-
December 2003 Project Management Journal • 17
7. PMI-002 11/5 11/6/03 9:20 AM Page 18
senting 8,380 projects. The results of A graph of the VAC over time pro- weeks, meaning that the cumulative AT
that research showed that 16% of IT vides a valuable indicator of trends in is 20 weeks. Therefore: TEAC1 = 20 +
projects were successful, 53% were project cost performance and the TETC weeks. In this case, TETC needs
challenged, and 31% failed. impact of any corrective actions. This to be developed for the remaining
Comparisons to subsequent studies graph can be effective in project work. TEAC1 may also be called the
are shown in Table 2 (The Standish reviews. Figure 12 shows a VAC graph revised schedule or current schedule.
Group, 1999): for the example project used in this When current analysis shows that
paper, using the above assumption. past schedule performance is not a
good predictor of future schedule per-
Y ear of Study Successful Challenged F ailed
Completion Time Forecasting formance, that problems or opportuni-
19 9 4 16% 53 % 3 1%
EVM has not been widely used to esti- ties which affected schedule
19 9 6 27 % 33% 40%
mate the total time at completion, performance in the past will not occur
19 9 8 26% 46% 28%
total project duration, or schedule for in the future, and that future schedule
Table 2. Project Resolution H Istory an activity, work package, or project performance will parallel the original
based on actual performance up to a plan, TEAC2 is the sum of the cumula-
The Treasury Board of Canada given point in the project. However, tive AT plus the original scheduled
Secretariat (2000–2002) supported using assumptions and logic similar to time for the remaining work. This can
findings of The Standish Group, indi- those discussed above, the project’s be simplified to the original baseline
cated similarities to results of reviews time estimate at completion (TEAC) SAC minus the TV (Fleming &
of Canadian government IT projects and time variance at completion Koppelman, 2000). As a formula,
and presented a framework for the (TVAC) can be calculated based on the TEAC2 = SAC – TV. For the above proj-
management of these projects. baseline schedule at completion (SAC) ect, TEAC2 = 40 – (-4) = 40 + 4 = 44
A survey of IT projects by Sauer and actual performance up to any weeks.
and Cuthbertson (2002) covered vari- given point in the project (Anbari, The above is the total estimated
ous industry sectors and government 2001 and 2002). schedule duration that would have
in the United Kingdom, and had a been obtained using the critical path
usable sample size of 565 projects. It method (CPM) or the program evalua-
showed that 5% of all projects were 40
tion and review technique (PERT), if
reported to have been abandoned 20 Good the schedule slippage of four weeks
VAC ($000)
prior to or during implementation, 0 were on the critical path.
55% of projects exceeded budget, 27% -20
When current analysis shows that
Poor
came in exactly on budget, and 8% -40
past schedule performance is a good
came in below budget. Performance, VAC predictor of future schedule perform-
measured by attainment of initially Time ance, that performance to date will
agreed upon specifications, averaged continue into the future, and that
above 80%. Across the whole sample, Figure 12. VAC Graph schedule efficiencies, or inefficiencies,
56% delivered 90% to 99% of the observed to date will prevail to com-
specifications, approximately 20% of In this section, various time esti- pletion, TEAC3 is the sum of the
projects delivered less than 80% of the mates are presented and given a cumulative AT plus the original sched-
specifications, and a sprinkling of proj- sequential subscript to differentiate uled time for the remaining work,
ects exceeded the specifications. among them, following the same pat- modified by the cumulative SPI. This
tern used previously for the cost esti- can be simplified to the original base-
Variance at Completion: mate at completion. line SAC divided by the SPI. As a for-
The variance at completion (VAC) When current analysis shows that mula, TEAC3 = SAC / SPI. For the
gives an indication of the estimated assumptions underlying the original above project, TEAC3 = 40 / 0.80 = 50
cost underrun or overrun at the com- time estimate were flawed or no longer weeks.
pletion of the project. As a formula, applicable due to changed conditions The above example indicates that
VAC = BAC – EAC. For the above proj- affecting the activity, work package, or the project is estimated to be complet-
ect, using BAC = 100,000 and EAC3 = project, a new schedule, duration esti- ed 25% behind schedule: (40 weeks –
150,000, VAC = 100,000 – 150,000 = mate, or time estimate to complete 50 weeks) / 40 weeks = -10 weeks / 40
-50,000. (TETC) needs to be developed, and the weeks = -0.25 = -25%.
In the above equation, 0 indi- TEAC1 is the sum of the cumulative AT A graph of the TEAC over time
cates that the project is forecasted to plus the TETC. As a formula, TEAC1 = provides a valuable indicator of
be completed on budget. A positive AT + TETC. trends in project schedule perform-
value indicates a forecasted under- The example project used in this ance and the impact of any corrective
run. A negative value indicates a paper has an original baseline SAC of actions. This graph can be effective in
forecasted overrun. 40 weeks, and its status date is 20 project reviews. Figure 13 shows a
18 • Project Management Journal December 2003
8. PMI-002 11/5 11/6/03 9:20 AM Page 19
graph of TEAC, for the example proj- prices for equipment and material, of prior poor performance. Longer
ect used in this paper, using the above negotiate better contract terms using durations, lower levels of accomplish-
assumption. more economical shipping methods, ment, and inefficient schedule achieve-
and similar actions. On the other ment patterns become practically
hand, if an activity, work package, or irreversible, and the project’s fate is
52 TEAC project were running below budget, sealed. Statistics of challenged and
48 opportunities for reducing completion failed projects testify that this case is
TEAC (Weeks)
44 Poor time, reducing cycle time, and expedit- much more common than we would
SAC
40 ing time to market may be pursued, like to believe, as previously discussed
36 Good although they may incur more cost. in the development of EAC6.
32 This may be accomplished through the
use of overtime, additional resources,
Time 70 TEACc
and expediting shipments.
60
D efining TEAC5 or TEACc as the
TEACc (Weeks)
Figure 13. TEAC Graph
50 Poor
TEAC adjusted for cost performance, SAC
40
In some organizations, it is com- the following formula would reflect
30 Good
mon to state that the activity, work the above assumption: TEAC5 =
20
package, or project will be on schedule TEACc = SAC / CR. For the above proj-
upon completion, regardless of prior ect, TEAC5 = TEACc = 40 / 0.53 = 75 Time
performance. This frequently occurs weeks. This formula may provide a bet-
early in the project, when prior sched- ter indication of estimated time at Figure 14. TEACc Graph
ule performance has been poor. The completion, when adherence to budg-
TEAC4 would be the original baseline et is critical to the organization. TEAC5 Time variance at completion: The
SAC. As a formula, TEAC4 = SAC. may also be called the time estimate at TVAC gives an indication of the esti-
Statements similar to those mentioned completion adjusted for cost perform- mated amount of time that the project
earlier in the cost discussion may be ance (TEACc). will be completed ahead or behind
heard. In some disciplines, such as A graph of TEACc over time pro- schedule: TVAC = SAC – TEAC. For the
software development, it is common vides a valuable indicator of trends in above project, using SAC = 40 and
to conclude these statements saying, project schedule performance and the TEAC3 = 50: TVAC = 40 – 50 = -10
“We’ll catch up during the testing impact of any corrective actions. This weeks.
phase!” Several modifiers to the word graph can be very effective in project In the above equation, 0 indicates
“test” have been developed, which may reviews. Figure 14 shows a graph of that the project is expected to be com-
increase the likelihood of catching up. TEACc, for the example project used pleted on schedule. A positive value
They include: alpha test, beta test, user in this paper, using the above indicates that the project is expected to
test, stress test, acceptance test, and assumption. be completed ahead of schedule. A
parallel test. Such statements should A case that is not mentioned negative value indicates that the proj-
be challenged firmly, with a response often occurs when the TEAC6 is sub- ect is expected to be completed behind
similar to that mentioned earlier in the stantially higher than the original schedule.
cost discussion. baseline SAC. As a formula, TEAC6 > > A graph of TVAC over time pro-
TEAC4 is rarely achieved. Again, SAC. This estimate is generally not vides a valuable indicator of trends
unmanaged projects do not fix them- quantified, but is referred to by proj- in project schedule performance and
selves. They only tend to fall behind ect team members with statements the impact of any corrective actions.
their schedules, overrun their budgets, similar to those mentioned earlier in This graph can be effective in project
and often miss other scope and quality the cost discussion. At times, this case reviews. Figure 15 shows a graph of
targets. occurs in the later phases of a project, TVAC, for the example project used
Recalling that cost performance when team members have no other in this paper, using the above
and schedule performance are insepa- planned assignments, and the organi- assumption.
rable, the assumption can be made zation is “right sizing.” Q uality prob-
that if an activity, work package, or lems become apparent, and CPM, PERT, and EVM
project were running over budget, additional time is requested to fix var- As mentioned above in the develop-
additional time may be needed to ious problems. Sometimes a lot of ment of TEAC2, an underlying
bring the project back on budget. This additional time is needed. assumption of the CPM and the PERT
may be accomplished by reducing Again, this case may result from is that future performance will parallel
resources applied to the project, using delaying corrective action and believ- the original plan, unless changes are
fewer paid resources, many of which ing for too long that the project would made to the original plan time, logic,
are less experienced and less skilled, somehow be completed close to the or cost.
taking additional time to find better original baseline schedule, regardless The example project used in this
December 2003 Project Management Journal • 19
9. PMI-002 11/5 11/6/03 9:20 AM Page 20
paper has an original baseline SAC of meetings. Both mathematical forecasts % Complete = $40,000 / $100,000
40 weeks. With a status date of 20 and subjective forecasts would be = 0.40 = 40%
weeks, TV = -4 weeks. If TV represented included in project performance % Spent = $60,000 / $100,000
a schedule slippage of 4 weeks on the reports. This effort highlights perform- = 0.60 = 60%
critical path, CPM and PERT would esti- ance deviations for work area man- CPI = % Complete / % Spent
mate a completion time of 44 weeks. agers, encourages them to consider = 40 / 60 = 0.67
This is the same as: TEAC2 = SAC – TV appropriate, timely actions, and incor- The above allows a further
= 40 – (-4) = 40 + 4 = 44 weeks. porates their close, detailed knowledge simplification
of performance in their areas, which (Slemaker, 1985) of the EAC3:
may not be evident from the reported EAC3 = BAC / CPI
values. At a minimum, this effort may = BAC / (% Complete /
10 help avoid surprises and arguments % Spent)
TVAC (Weeks)
5 Good
over the numbers during project = (BAC x % Spent) /
0 review meetings. % Complete
-5 Poor Forecasting in project manage- = AC / % Complete
-10 TVAC ment may well be a self-defeating Thus: EAC3 = AC / % Complete
Time prophecy, and that may be good for The definition of EAC3 can be fur-
the organization. Large deviations usu- ther simplified to the AC divided by
ally attract management’s attention the percent complete. For the above
Figure 15. TVAC Graph
and result in corrective action. Small project, EAC3 = $60,000 / 0.40 =
CPM and PERT initially assume deviations are usually left alone. By $150,000.
that problems or opportunities that quantifying and highlighting such Similarly, the TEAC3 can be sim-
affected performance in the past will deviations, EVM helps focus manage- plified to: TEAC3 = AT / % Complete.
not occur in the future and that past ment’s interest on projects or work The example project used in this paper
performance is not a good predictor of packages that need the most attention. has an original baseline SAC of 40
future performance. As a result, EVM supports effective weeks, and the status date is 20 weeks,
The assumption generally associ- management of projects and work which means that the cumulative AT is
ated with EVM is that past perform- packages collectively and enhances 20 weeks. Therefore: TEAC3 = 20 / 0.40
ance is a good predictor of future management of the enterprise’s project = 50 weeks.
performance, that performance to date portfolio (Anbari, 1983). Forecasting Similarly, the following is derived
will continue into the future, and that using these techniques provides a uni- (Anbari, 1980):
efficiencies or inefficiencies observed form approach to project reviews, CPI = % Complete / % Spent
to date will prevail to completion. building confidence in the project out- = (Actual Production /
Therefore, the EAC3 is generally asso- come as time progresses. Changing Total Scope) /
ciated with EVM. Similarly, the TEAC3 project evaluation methods during the (Actual Cost /
can be associated with EVM. project duration can result in no Total Budget)
Therefore: TEAC3 = SAC / SPI = 40 / meaningful data for decision-making. = (Actual Production /
0.80 = 50 weeks. Total Scope) x
Which of the above forecasts will Further Ex tensions, Issues (Total Budget /
materialize depends greatly on deci- and Applications Actual Cost)
sions and actions taken by the project = (Total Budget /
manager, the project team, and the Extensions Total Scope) x
organization. Some like to add luck to Using the above definitions, the fol- (Actual Production /
the factors affecting project outcomes. lowing is derived (Slemaker, 1985): Actual Cost)
Others observe that good luck tends to % Complete = EV / BAC = (Total Budget /
be directly associated with better plan- % Spent = AC / BAC Total Scope) /
ning and better decisions. Taking the ratio of the above two (Actual Cost /
formulas Actual Production)
Project Forecasting (Anbari, 1980): = Planned Unit Cost /
It is advisable to ask work package % Complete / % Spent Actual Unit Cost
managers, project leaders, and func- = (EV / BAC) / (AC / BAC) Thus: CPI = Planned Unit Cost /
tional managers to review cost and = EV / AC Actual Unit Cost
schedule mathematical forecasts and = CPI The additional formulas devel-
to provide their own subjective fore- Thus: oped in this section provide a more
casts for their own work areas in CPI = % Complete / % Spent intuitive understanding of CPI based
advance of issuing project performance For the example project used in on information readily available in
reports and conducting project review this paper: many organizations. The first formula
20 • Project Management Journal December 2003
10. PMI-002 11/5 11/6/03 9:20 AM Page 21
for CPI uses information widely ble, collecting 50% of the contract centage of the value associated with
known in project environments, and price for each of these items. the milestone based on subjective esti-
the second formula for CPI uses infor- The 0/100 rule can also be used. mates. When the predefined, tangible
mation widely known in production This rule specifies that the value is criteria for the milestone are met, the
environments: earned only when the item is complet- balance of the value associated with
CPI = % Complete / % Spent ed and is usually used in work pack- the milestone is earned (Fleming &
CPI = Planned Unit Cost / ages having a short duration (Kerzner, Koppelman, 2000). These approaches
Actual Unit Cost 2001). This rule can also be called the may help alleviate the “95% complete
weighted milestone method, where the and stays there forever” syndrome.
Issues in the Determination of value is earned only when the mile- For level of effort items such as
Percent Complete stone is physically completed, and one project management, customer sup-
Determination of the percent complete or more milestones are planned in port, and other support work during a
or proportion complete of an activity, each performance-reporting period given period of time in a project, the
work package, or project is a necessary (Fleming & Koppelman, 2000). effort itself is the end product.
but challenging task in many organiza- Contractors may consider the 0/100 Therefore, the earned value can be con-
tions. This task becomes even more rule harsh. When a contractor is paid sidered to be equal to the effort
demanding when dealing with new, based on this rule, it is reasonable to applied or the actual cost.
emerging, or softer technology proj- expect that the contractor will strive to
ects, such as telecommunications, soft- have a very detailed WBS that breaks Applications
ware development, architectural or the project down to as many items as EVM provides project managers and
engineering design, and research and possible, so that completion of item(s) the organization with triggers or early
development. can be shown regularly and payment warning signals that allow them to take
Alternatives to using the percent can be authorized. timely actions in response to indicators
complete to determine physical Other alternatives for determining of poor performance and enhance the
accomplishments have been used. The physical accomplishments can be opportunities for project success. Such
50/50 rule specifies that 50% of an used. For example, the 10/90 rule, indicators have been found to be reli-
item’s budget is recorded at the time 20/80 rule, and 25/75 rule acknowl- able as early as 15% into a project.
that the work is scheduled to begin, edge that to start a work package, a cer- Better planning and resource allocation
and the remaining 50% is recorded tain amount of preparation and associated with the early periods of a
when the work is scheduled to be com- mobilization are needed. Therefore, project might be the cause of this relia-
pleted. If the project had a large num- 10%, 20%, or 25% of the value would bility (Fleming & Koppelman, 2000).
ber of items, the distortion from the be considered earned when the work is EVM can be used for progress pay-
50/50 rule would be minimal started, and the remaining amount ments to contractors based on the EV
(Kerzner, 2001), because these items would be earned when the work is of contracted or outsourced work.
would be at various stages of comple- completed. If the work package were Because such contractual arrangements
tion. This allows us to calculate PV. front-end loaded, as might be the case create legal and financial obligations, it
Similarly, to calculate EV, 50% of an with certain equipment acquisitions, is important to consider the method
item’s budget is recorded when the then the inverse of these rules might be specified for evaluating progress. The
work begins, and the remaining 50% is appropriate. For example, the 75/25 previously discussed alternatives for
recorded when the work is completed. rule might specify that 75% of the determination of percent complete
To make the 50/50 rule work success- value would be considered earned should be carefully considered and
fully, the project should be broken when the equipment is delivered, and negotiated to achieve a fair and equi-
down into very detailed, short-span the remaining amount is earned when table environment that encourages suc-
work packages (Fleming & installation, testing, and commission- cessful accomplishment of contracted
Koppelman, 2000). ing are completed. or outsourced project items.
The 50/50 rule is a common prac- The percent complete method can For long-term projects, it may be
tice in a number of contractual be used with a buffer that sets a ceiling appropriate to consider incorporating
arrangements, such as those for home of about 80% or 90% upon reported the time value of money and time-dis-
repair. Half of the contract price is completion. A work package may earn counted cash flows into EVM. Inflation
paid up front, and the remaining bal- only up to the specified percent ceiling can be explicitly considered in EVM,
ance is paid upon completion of the based on subjective estimates. When and the inflation variance (IV) can be
work. It should be noted that when the work package is 100% complete, calculated (Farid & Karshenas, 1988).
the 50/50 rule is used in a contractual the balance is earned. A variation of However, these considerations add
arrangement and the contractor is this approach is using a combination complexity to the method and may be
paid based on this rule, it is reason- of the percent complete and a mile- justifiable only for very long-term proj-
able to expect that the contractor will stone gate. A work package may earn ects or in very high inflation periods or
tend to start as many items as possi- only up to a maximum specified per- economies.
December 2003 Project Management Journal • 21
11. PMI-002 11/5 11/6/03 9:20 AM Page 22
An organization may elect to 0.67 = $150,000 Analysis Method [ CD-ROM] .
apply EVM uniformly to all of its proj- EAC3 = AC / % Complete = 60,000 / Proceedings of the Project Management
ects, or only to projects exceeding its 0.40 = $150,000 Institute 2 0 0 1 Seminars & Symp osium,
own thresholds for cost and schedule VAC = BAC - EAC = $100,000 - N ovember 1-10, 2001, N ashville, TN ,
reporting and control. EVM can be $150,000 = -$50,000 USA. N ewtown Square, PA: Project
applied to projects of various types and Management Institute.
sizes in the public and private sectors. EAC5 = EACs = BAC / CR = $100,000/ Anbari, F.T. (2002). Q uantitativ e
It can be applied at various levels of a 0.53 = $187,500 Methods for Project Management, Second
project’s WBS and to various cost com- Edition. N ew York, N Y: International
ponents, such as labor, material and TEAC3 = SAC / SPI = 40 weeks / Institute for Learning.
subcontractors. 0.80 = 50 weeks Barr, Z . (1996). Earned Value
TEAC3 = AT / % Complete = 20 / Analysis: A Case Study. PM N etwork , X
Comprehensive Example 0.40 = 50 weeks (12), 31-37.
A project has a baseline BAC of TVAC = SAC – TEAC = 40 weeks – Cioffi, D. F. (2002). Managing
$100,000 and a baseline SAC of 40 50 weeks = -10 weeks Project Integration. Vienna, VA:
weeks. The baseline indicates that by Management Concepts.
the end of week 20, the project is Conclusion Cleland, D.I., & King, W.R.
planned to be 50% complete. At the EVM helps focus management’s inter- (Editors). (1988). Project Management
end of week 20, it is reported that 40% est on projects that need most atten- H andb ook , (2nd. Ed.). N ew York, N Y:
of the project work has been complet- tion and may aid the prioritization and Van N ostrand Reinhold.
ed at a cost of $60,000. Using the EVM emphasis management gives projects Egan, Jr., D.S. (1982). The
method: within a portfolio, enhancing the Performance Index: Combining Cost
BAC = $100,000 enterprise’s project portfolio manage- and Production Data to Show How
SAC = 40 weeks ment. EVM provides important infor- Good (or Bad) Your Project Really Is!!
mation for project or work package Proceedings of the 7 . Internet W orld
PV = 50% x $100,000 decision-making. Its wider acceptance C ongress on Project Management, 1982,
= $50,000 and effectiveness may depend on bet- Copenhagen, Denmark, PROJEKT-
AC = $60,000 ter understanding of its capabilities. PLAN , The Danish Project
EV = 40% x $100,000 Simplification of EVM calculations, Management Society, The Danish
= $40,000 use of graphical tools to enhance Technical Press, Denmark, 355-364.
AT = 20 weeks understanding of performance trends, Farid, F., & Karshenas, S.
Therefore: and successful application of EVM in (N ovember, 1988). Cost/Schedule
% Complete = EV / BAC industry are important factors for the Control Systems Criteria Under
= $40,000 / $100,000 = 40% growth and effective use of this valu- Inflation. Project Management Journal,
% Spent = AC / BAC = $60,000 / able method in project management. X IX (5), 23-29.
$100,000 = 60% Fleming, Q.W., & Koppelman, J.M.
References (2000). E arned V alue Project
CV = EV – AC = $40,000 – Anbari, F.T. (1980). An Operating Management, (2nd Ed.). N ewtown
$60,000 = -$20,000 Management Control System for Large Square, PA: Project Management
SV = EV – PV = $40,000 – Scale Projects. Unpublished paper pre- Institute.
$50,000 = -$10,000 sented at The Decision Sciences Institute Heinze, K. (1996). C ost
N inth Annual Meeting, N ortheast Management of C ap ital Projects. N ew
PV Rate = BAC / SAC = $100,000 / R egional C onference, Philadelphia, PA. York: Marcel Dekker, Inc.
40 weeks = $2,500 per week Sponsored by the N ortheast Decision Kerzner, H. (2001). Project
TV = SV / PV Rate = - $10,000 / Sciences Institute. Management: A Systems Ap p roach to
$2,500 per week = -4 weeks Anbari, F.T. (1983). An Operating Planning, Scheduling, and C ontrolling,
System for Forecasting Project Cost at (7th Ed.). N ew York, N Y: John Wiley &
CPI = EV / AC = $40,000 / Completion. Unpublished paper pre- Sons.
$60,000 = 0.67 sented at The Third International Kim, E.H. (2000). A Study on the
CPI = % Complete / Symp osium on F orecasting, Philadelphia, E ffectiv e Imp lementation of E arned V alue
% Spent = 40% / 60% = 0.67 PA. Abstract in Program sponsored by Management Methodology, Unpublished
SPI = EV / PV = $40,000 / The International Institute of doctoral dissertation. The George
$50,000 = 0.80 Forecasters in collaboration with the Washington University, Washington,
Wharton School, University of DC.
CR = CPI x SPI = 0.67 x 0.80 Pennsylvania, Philadelphia, PA. Lewis, J.P. (2001). Project Planning,
= 0.53 Anbari, F. T. (2001). Applications Scheduling, & C ontrol: A H ands-On
EAC3 = BAC / CPI = $100,000 / and Extensions of the Earned Value Guide to Bringing Projects In On Time
22 • Project Management Journal December 2003
