Home Monitoring & Controlling Communication Balanced Estimate at Completion, using Earned Value Management
US FTC Compliance
Yes, all these ads are some kind of affiliate link and I get paid a commission if you click or buy.
Not enough to quit my day job, but it keeps the site alive.
- Jeb Riordan, Editor, PROJECTmagazine
Balanced Estimate at Completion, using Earned Value Management Print
One of the most useful outputs of Earned Value Management is the Estimate at Completion; it allows the project manager to keep predicting the project's final cost along the project's life cycle - before the project is completed.

On each time a project manager evaluates the Estimate at Completion he should use all his knowledge in order to get estimations as accurate as he possibly can and this results in a series of estimations usually calculated on different ways.

Although this allows the project manager to do his main job when estimating the project's final cost, he must have the sufficient knowledge to justify each cost deviation.
What if he can't do that?
What if he doesn't know if past deviations are likely to occur in the future?
Which Estimate at Completion should he use then?
My proposal for such situations is the use of the Balanced Estimate at Completion.

Remarks: This paper uses the notation by Quentin W. Fleming and is intended for those that already use Earned Value Management.
Although this paper is focused on project management activities it can be applied anywhere Earned Value Management is applied.

Introduction

When I started using the Project Management Institute's framework was when I first tripped on Earned Value Management (EVM). I was immediately amazed by the way EVM turned things that were specific and subjective into objective and comparable values.
Yes, it is a narrow point of view if you turn something as complex as a project into a single value.
But then again, that enables you to do a few marvelous things such as:

  • See how well your ability as a project manager is growing
  • Detect your worst projects to avoid making the same mistakes
  • Benchmark your projects against a specific market to get a fair judgment of how good you are as a project manager

Common Estimations at Completion

But one thing bothered me and that was using a few different Estimates At Completion (EAC).
"Why wasn't there just one?"
Was the obvious question and "Because different deviations occur due to different causes" was the obvious answer.

If you use EAC as: Image

It means that you basically trust the project's plan (the budget) and consider the project's performance as an accident.
A typical example of such a situation is when a deviation occurs because of an unpredictable cause that isn't expected to have an impact on any other project's work.

If you use EAC as: Image

It means that you trust the project's performance the most.
This estimation is used when the reasons for deviations are likely to affect all other project's work.
There are other ways to find an EAC but all of them are something static in between these two (static in this context means static over time, that is, along the project's progress).
Those EACs that are dynamic also use the same basic EAC but with other factors like CPIxSPI, SPI or 80%CPI+20%SPI instead of CPI alone.

What is the Balanced Estimate at Completion

The Balanced Estimate At Completion (BEAC) is a straightforward, project and context insensitive way to evaluate the EAC when the project manager can't tell

  • why deviations have occurred or
  • if the reasons for past deviations are likely to affect the rest of the project.

 

So how is BEAC constructed?
You may have spotted the answer already.
If you look at the two ways to calculate EAC presented earlier (EACA and EACB) they are different because EACA has PV-EV divided by 1 and EACB has PV-EV divided by the Cost Performance Index (CPI).

That is, EACA uses a factor of 1 and EACB uses CPI as the factor.
So why don't you use EACA at the project's start, EACB at the project's end and PV-EV divided by something in between the two factors (1 and CPI) as the project evolves?
That would give an estimation reflecting more confidence on the CPI as the project progresses; and for that you need to define a factor that varies along the project's progress, that is, that varies with the Schedule Performance Index (SPI).
This means you will get a kind of a dynamic estimation, something that grows as the project itself grows.

 

Project's Progress SPI Factor to use
Beginning 0 1
Ending 1 CPI

 

This makes sense because when the project is just starting many things can happen that make deviations occur. Towards the project's end you are more sure about the project's performance because it has less slack to grow or shrink.

BEAC geometric meaning

So what you need is a function that transforms the SPI into the factor listed on the table.
The graphic shows a function like that; in fact, it shows the simplest possible continuous one, a straight line.
The horizontal axis (x) represents the project's progress, that is, the SPI.
The vertical axis (y) represents the factor you want to use. Image

To transform one into the other you can now use the formula that represents a line defined by two points Image and Image ,
that is: Image

This is the line shown in bold on the graphic and it is defined by the two points (0,1) and (1,CPI) and you can now adapt the formula using them, obtaining:

Or better yet: Image

Remember that x represents SPI so you can use it instead, getting: Image

And y is what you need to use as a factor. If you do use it as a factor you get: Image

And that is the estimation I propose to be used under certain cases, like those specified earlier.

How does BEAC work

Examples 1 and 2 are based upon the same project.
On the first case a deviation occurs and is never repeated and on the second case the deviation occurred is repeated on each Work Breakdown Structure item.
The measures were taken on two different project stages, the first near the project's start (SPI=18%) and the second near the project's end (SPI=82%).

 

Example Stage BEAC-PC EAC(a)-PC EAC(b)-PC
1 1 309 0 2.3
1 2 88 0 109
2 1 5.19 5.75 0
2 2 379 1.25 0

 

This table shows how far from the Project Cost (PC) these EACs are.
This Project Cost is the actual value of the project cost after the project is completed.
As you can see, this distance involving BEAC is always something in between the ones calculated using EACA and EACB - and this is general, not just an accident provided by these examples.

This is expected by the way BEAC was constructed. BEAC isn't expected to give you the best estimation, not even with these extreme examples; but then again, it never gives the worst estimation either.

Why use BEAC

The bottom line is:
when one doesn't know
(i) why deviations occurred
or
(ii) if past deviations will occur in the future, BEAC is the estimation to use and I can give you four reasons to do it:

  • Even in those cases, a project manager must choose an EAC to use.
  • The project manager should use an EAC that doesn't extreme situations like the ones caused with EACA and EACB. This leaves most EACs out but still includes at least the dynamic ones like the one that uses the factor CPIxSPI.
  • BEAC actually uses EACA and EACB. When the project is just starting, SPI=0 and so BEAC= EACA. When the project is ending, SPI=1 and so BEAC= EACB. That is, if you look at BEAC as a function on SPI, you get:
    Image

    And:
    Image
    This BEAC's property is expected and it is essential because
    • there is no CPI when a project starts
    • and because - BEAC's prediction is the actual cost when a project is over.
  • The last reason to use BEAC, and I believe that it is the most important one, is because it has a geometrical meaning that captures the project manager's needs under these circumstances - namely the uncertainty of the project's progress and the growing confidence as the project evolves.

 

Because of this I propose the use of BEAC as the standard estimation except for the cases in which it is possible to determine
(i) the reasons for deviations and
(ii) if and how they will affect the project's progress.

This may sound like a strong statement to make but it's just a question of semantics - this last sentence has the same meaning than the one used earlier. The difference is just that the first one states when to use BEAC and the second states when not to use BEAC.

Beyond BEAC

You may be considering now if BEAC is the final word on this subject.
The answer is a clear no, simply because you can consider other variables like the speed of growth of confidence - you can consider that confidence grows slower when the project starts and grows faster towards the project's end.
And this makes sense too. But I believe this paper has shown you that BEAC is simple enough and close enough to the purpose of estimating as a straight forward, project and context insensitive indicator.

What this means is that BEAC is the simplest estimate with the following properties:

  • BEAC is an EAC
  • BEAC can be used under uncertainty (it doesn't translate extreme situations)
  • BEAC depends on the project's progress (SPI)
  • BEAC(0)=EACA and BEAC(1)=EACB
  • BEAC has a geometrical meaning

 

So there you have it: BEAC is the simplest indicator with these properties (although it isn't the only one).Image

 

2002 © Luis de Seabra Coelho

Examples

Example 1, Stage 1

Image

Example 2, Stage 1

Image

Example 1, Stage 2

Image

Example 2, Stage 2

 

 

Luis de Seabra Coelho (c) 2002

Share/Save/Bookmark
Comments (0)Add Comment

Write comment
bold italicize underline strike url image quote Smile Wink Laugh Grin Angry Sad Shocked Cool Tongue Kiss Cry
smaller | bigger

busy
 
Copyright © PROJECTmagazine (c) 1998 - 2019 for practical project management information. All rights reserved.