Copyright 2009 Corvus International Inc.  All Rights Reserved

Home      About Corvus       Contact Us       Articles/Resources       Clients          Affiliations  

Plans are nothing,  planning is everything.
                        Gen. Dwight D. Eisenhower
quoting Prussian General Helmuth von Moltke

No battle plan survives contact with the enemy
                        Gen. Helmuth von Moltke quoting himself

No software project plan completely survives
contact with the project
                        Phillip G. Armour
quoting Prussian General Helmuth von Moltke


ACM DL Author-ize serviceTo plan, two plans

Phillip G. Armour
Communications of the ACM - Special issue: RFID, 2005









Men of Company E, 16th Regiment of the 1st Infantry Division take cover behind a "hedgehog" beach defense
at Easy Red sector,  Omaha Beach, June 6 1944.  Copyright (c) Robert Capa/Magnum Photos
"Es una cosa muy seria"------"This is a very serious business"

One of the most visible planning "failures" in the history of the world occurred starting around 6:30am on June 6 1944 on the beach between the villages of Vierville and Coleville on the northern coast of France.  If "success" is defined by "adherence to plan" this part of the D-Day invasion could not be considered to be successful.

Predicting the future                                             

To some extent all planning is an attempt to predict the future.  This is not easy; as Nobel physicist Niels Bohr once remarked: "Prediction is difficult, especially if it involves the future."    This is true of projects.  The reason why we cannot predict the future is its intrinsic uncertainty: things may happen that we don't know about or think will happen; things that we do think will happen don't happen or don't happen as much.  This uncertainty has the effect of invalidating plans.

Why plan?                                                           

So if plans are almost always ex post facto wrong, why bother planning?  Why not charge ahead and just do it?  The reason is that plans have other purposes, other than to be "right".  They also exist to:

  •  Publish project goals--so everyone knows what "should" happen, if not in events and actions, then at least in the end-state

  •  Prepare for "likely" events--the things that probably will (but maybe won't) happen

  •  Anticipate "unlikely" events--the things that probably won't (but maybe will) happen

  •  Provide a reasonable quantity of resources--to deal with the likely and unlikely events

What the plan cannot deal with, except inferentially, is the wholly unexpected events which we, well, didn't expect. [SEE NOTE AT END]

These are very valuable aspects of a plan over and above the plan's post hoc "correctness"

Sources of variance on projects                       

There are a number of sources of variance (the reason why the plan and the project turn out not to be the same), including:

  •  Scope--how big and complicated the final product turns out to be, including time-based requirements changes

  •  Performance--how "good" we are at build this kind of system

  •  Technology--what it brings us and what it costs us

  •  Market--the simple fact that what was required when the project started is not what is required when it is finished

The aggregate of the effect of all the sources of variance can be called the "intrinsic uncertainty" of the project.  So how can we deal with this uncertainty?  The only real answer is to provide more resources of time, budget, and people to deal with it.  Unfortunately, management often views this as "padding" the project.  It is not, it is a determination of the "cost of risk" being carried by the project, and if the cost of risk is not covered, then the risk may compromise the project.

Two plans                                                   

The solution is to create not one, but TWO plans: one that openly expresses the risk and one which does not

The Commitment Plan "hides" the risk by including the cost of risk in the cost of work.  The Work Plan separates the two, the "size" of the Management Risk Reserve being calculated from an assessment of the intrinsic uncertainty mapped up against an acceptable probability of completion (the plans show here are assumed to be schedule-based, but the x-axis could also be budget or staff or some other resource).

As the project progresses, we should expect to see the following:

A "Detailed Replan" of the initial plan should show "slippage"--the reason being that the effort expended in the Detailed Replan activity has created a more detailed plan and has inevitably resolved some of the uncertainty.  By doing some of the work on the project, we actually build the artifacts and work products required and we also inevitably resolve some of the uncertainty.  Resolving this uncertainty costs time and effort, and therefore we should expect to use up our risk reserve.  That is, of course, why it is there.

In the ideal situation, if we have correctly calculated the intrinsic uncertainty, and our performance at resolving that uncertainty, by the end of the project we will use up all the assigned budget for work (yellow) producing artifacts and work products and all the assigned budget for risk (blue) resolving those things we did not know about when we initially planned the project.  If we have under-estimated the risk (including incidentally the risk of under-sizing the yellow work component) we would expect the project to run over.  Equally if we have over-estimated the risk, we should finish early, or with a higher quality and more functional product.

Results, not plans                                                  

The critical thing is that we don't  expect the plans to actually work.  Their job is not to be "correct", it is to be valuable which is a different criterion.  The plan for the Omaha Beach landing on D-Day was huge and almost none of it actually worked the way it was expected to.  But the end result was that despite a large loss of life on the beach, Hitler's vaunted "Atlantic Wall" held up the US Army for less than one day.

We should measure the effectiveness of planning, not by how close the plan came to predicting the future, but by how it assisted in facilitating a positive outcome.


One More Quote                                       

Everyone has a plan,...
...until they get hit
Mike Tyson
Professional Heavyweight Boxer


NOTE: Likely and unlikely things                       

These correspond to:
Likely Things--this is 0OI and some of 1OI: the answer and the "predictable questions" (those questions which generate answers that we thought had a high probability)
Unlikely Things
--this corresponds to the remainder of 1OI: the "unpredictable questions" that generate answers that had a low probability
Unexpected Things--this is 2OI and 3OI: the things we didn't know we didn't know
See this article for more information on The Five Orders of Ignorance 



To Plan, Two Plans