Narrated Version Dallas MPUG
Dallas MPUG slides with speaker narration
Published on: Mar 3, 2016
Transcripts - Narrated Version Dallas MPUG
When I say the Probability of Success it means that projects are always in some kind of
trouble. No matter how experienced, and well planned, how well funded, and especially
how good the tools are.
There is always a probability that the project will be in some sort of trouble.
This is a fact of life for projects.
What we need to understand is that managing projects in the presence of this uncertainty
the probability that there will be some sort of trouble – requires careful application of
b bilit th t th ill b t f t bl i fl li ti f
tools and the processes around the tools.
Tonight’s talk is not really about tools. It’s about applying processes and tools to increase
the probability of success.
I’m going to speak to you tonight as a tool skeptic. By skeptic I mean as someone who is
deeply committed to using tools, using them on the complex, mission critical project and
But all the same a skeptic for one simple reason. Putting tools to work requires a credible
process. By credible I mean a project management process that has a chance of actually
getting the project out the door on‐time, on‐budget, and on‐specification.
Tools are necessary, but they are far from sufficient.
So for this evening, I’m going to focus on the credible processes that can be enabled by
So for this evening I’m going to focus on the credible processes that can be enabled by
tools and not the tools them selves.
I know you’ve seen lots of demonstration of tools.
Heard speakers tell you about how to use tools to manage a project. Listened to all the
benefits of one tools over another.
Having managed projects since the early 80’s I’m here to tell you it’s not about tools.
It’s also not about people, but a reason that may surprise you. If you don’t have the right
people, no tool or process is going to help.
If you don’t have the right tool, the project can still get done. We do that everyday on the
program we manage. The tools are far from ideal in many cases, but work still gets out the
It’s about the processes that are implemented by the tools, using the right people
But this talk is not about the demonstration of tools. It’s about how to implement the right
processes using some tools your familiar with.
processes using some tools your familiar with
Let’s first see at what “done” looks like.
This statement – “what does done look like,” is a critical success factor on all projects. If
you don’t know what done looks like, you'll never recognize it when you get there – if you
every get there.
So what are we here to learn about:
The first thing is that cost, schedule, and technical performance are inseparable.
Technical Performance is a word that may not be familiar to some of you. In the Defense
T h i lP f i d th t t b f ili t f I th D f
and Space program management business – where I spend most of my time – Technical
Performance means all the attributes of a program that are not cost and schedule. It’s the
attributes of the product or service that the customer has bought. The “abilities” of the
product or service. These could be as simple as processing data or as complex as flying to
the moon and getting alive.
For these three attributes process is what integrates them Not tools Tools are the
For these three attributes – process is what integrates them. Not tools. Tools are the
physical mechanisms but processes providing meaning to those mechanisms.
People are the source and consumers of information. They are critical of course, but it’s
the processes that give meaning to this information that allows the people to make
In the end the tools are just pieces of equipment used by the project. Nothing more
nothing less. A fool with a tool is still a fool – the saying goes.
hl fl h l ll f l h
So what’s the problem domain in which we want to use tools to help us with our processes.
Well these are probably common issues in the domains you work in.
They certainly are common problems in the domain of defense systems and manned space
So no matter where you go, the problems are pretty much the same. You can make your
At the core of the casual sources is usually some missing process.
• What the customer wants is not clear, concise, or stable
• The system architecture is changing under neither the project plan
The first thing we need to understand when we’re talking about process and tools is that
the world of project management is a probabilistic world. There are very few if any fixed
outcomes. Nearly everything is random in some way.
Random means that the measurements of numbers is actual a statistical process, where
the value you measure is drawn from an underlying probability distribution.
Let’s remember this later where we’ll see some tools that provide information about
managing in the presence of this randomness. Because this randomness cannot be
removed. It is part of the world we live in.
We cannot manage the uncertainties that come out of this randomness, we can only
manage in the presence of these uncertainties.
We’ve got the three obvious elements of any project.
People, process, and technology.
That’s been stated up front. But let’s review quickly where we’re going before we start out
• People pay for solutions, they don’t pay for processes or technology. People produce the
solutions. The processes and technology are just a means to the end.
• Processes guide the people in the development of a product or service. Some processes
are more rigorous than others. Some processes have more formality than others. Some
processes come with the tools, some are independent of any tool. Tonight I’m going to
speak about processes that are independent of any tool. The tool is simply a means to
make the application of the process easier.
• Tools are glue between a process and the people who use the process. Tools transform
one view of the project into another view of the project. That’s all. They provide data.
The information needed to manage the project is provided by the process. The meaning if
the data is part of the process.
Here’s my view of the process world. This view comes for experience, training, and more
experience. This view has evolved over several decades, mostly from failed projects.
Projects or programs that failed to perform one or more of this processes and the related
These five process areas – if we use the CMMI Process Improvement framework vocabulary
1. We need to know what done looks like in terms of the user. What useful work is the
1W dt k h td l k lik i t f th Wh t fl k i th
user going to do with the products or services delivered by the project? This question is
not always answered. Pretend you had the resulting system next Monday, it was free,
and it met all the requirements stated in the requirements document. What would you
do with the system? This is call the “System Capabilities.”
2. With this information we can derived the requirements. All requirements are derived
from the Capabilities of the system. If they are not, they have no home. No reason for
from the Capabilities of the system If they are not they have no home No reason for
3. With the requirements we need a plan and a schedule. A plan is the strategy for
successfully delivering the capabilities. The schedule is the sequence of work
performed to deliver the capability. They are not the same, both are needed.
4. Next we need to execute the Plan and the Schedule
5. Finally we need to perform continuous risk management for the four process areas.
I’ll show you next how tools can support all five process areas.
So let’s connect our 5 process areas with the three attributes of a project.
Here’s a notional picture of these connections.
In the defense and space business – notional means a picture that “could” the real way to
do something, but it is not actually the way it is done in practice.
In other words it’s a nice power point slide of what might take place. But there’s more to it
than what’s shown in the power point slide.
Since our time is short tonight, let’s skip to the end.
Let’s start into the tools talk, even though I said this talk is not about the tools per se.
But without the tools, the processes become hard to deploy and the people start looking
somewhere else for solutions.
Remember tools are the glue that connects people with the process.
First is our requirements management tools.
Actually first was the capabilities management tools, but that topic is outside our interest
The development of system capabilities is business side function in IT and a Systems
Engineering function in other domain.
But capabilities are easy in certain ways. Here’s an example. The “owner” of the Hubble
Space Telescope – F k C
S Tl Frank Ceponilla – h d th
ill had three capabilities he needed for a robotic serving
biliti h d df b ti i
1. Do no harm to Hubble
2. Replace the Wide Field Camera – it was busted
3. Connect an umbilical cord to the batteries to extend their life
How much does that cost? And when can you have the solution ready.
H hd h ?A d h h h li d
Now there were 1,000’s of requirements that came from those 3 simple capabilities. But
without the capabilities being stated first, each requirement had no reason for being there.
So requirements tools are needed to keep track of why and what questions for the project.
Once we have the requirements in some form we need to answer the HOW and WHEN
questions. This is the purpose of the schedule.
But this schedule needs to posses certain attributes to be of any use to the project in
forecasting the future.
This forecasting process is what schedules are actually about. They are useful for
determining what should happen in the future. But their real value is to reveal where there
are trouble spots that are not visible in the static description of the linked work.
I’ll show you how to use a Monte Carlo Simulation (MCS) tool in later slides. This tool
provides information about the schedule that can not be determine with a static structure.
HOW MUCH is another question that needs to be answered.
Connecting cost and schedule is the minimal acceptable process of any credible project
But this effort always starts with the schedule. So when we says “cost and schedule,” we
really mean “schedule and cost.”
One question that not frequently answered correctly in IT projects is “what are the
impediments to the success of my plan?”
What are the risks, how can they be handled, and what does it cost to handle them in
terms of time and money?
We need a risk management tool. Actually we needed several of them. One manages the
lists of risk, the inventory of risks, who’s looking after them. Another tool models how
these risks impact the cost and schedule.
Remember this phrase – risk management is how adults manage projects. It may be crass,
but it’s true.
These four tools areas need to come together in one place for them to be effective.
They can be separated, but it is better for the project if they’re integrated.
Now this integration can tale many forms. Project Server is one form. Separate tools
integrated “on the glass” is another.
But before any integration of the tools takes place, the work processes for managing the
project needs to be in place. Only then should tools be added to the process flow.
If you start with the tools, the needed capabilities of the project – the business process
capabilities are locked. And making changes to them becomes more difficult.
Wait as long as you can before deploying a tool. Any increases in the probability of project
success comes from the process side, not the tools side. Remember tools are the glue
between people and processes. Glue alone adds little value, other than spend money for
the tools and raise everyone's hope that the project probability of success will increase.
Being integrated is necessary but not sufficient. The integration needs to produce value for
Here’s a list of value statements we should look for when we say “we have an integrated
project management system based on tools.”
Let’s revisit some questions we want our tools to help us answer.
The tool itself doesn’t provide the answer. People provide the answer. The tool provide the
data and the information that provides the answer.
Do we know what the customer wants? Capabilities and requirements are good ways to
answer that. Descriptions of the work needed to produce the requirements are next.
How much will it cost to do this work?
When will this work be done? This is the role of the Plan and Schedule.
What are the impediments to progress – a risk management tool helps here.
Can we talk to the customer in units of measurement meaningful to them? Physical
Percent Complete is a good unit of measure.
Now this is a real snoozer topic – work breakdown structures. For some there could not be
a fate worse – be in charge of the work breakdown structure.
I’ll say this with as much emphasis as possible. You don’t have a work breakdown structure
in some form – you don't have a project.
I came here tonight from a client site where the WBS is noncompliant with the DoD
requirements stated in ANSI‐748B. Now for small IT projects that may not mean anything.
But a billion dollar development program, not having a compliant description of what work
is being done, by whom, with what resulting product is a big problem.
I can’t emphasize enough that the project starts with the WBS.
From the WBS all other project attributes follow. No WBS no project plan.
If your executing a project without a credible WBS you’re wasting money. If it’s your money
shame on you. If it’s someone else's money shame on them for letting you spend it without
This may sound harsh, but in the end someone, somewhere is going to ask – “what will it
get for my money?” The WBS tells you that.
But more important, the WBS and the associated WBS dictionary tells you what done looks
And knowing what done looks like is what project management and project management
tools are all about.
So let’s look at a really good tool for building a WBS.
No it’s not MSFT project, it’s MindJet’s Mind Manager. You may not have heard of
MindManager, but you probably have done “mind mapping” before.
Mi dM bt b bl h d “ id i ”b f
It’s what you do when you draw pictures of how things are related. In this case related
through the hierarchy– a Work Breakdown Structure is a hierarchal description of what
done looks like.
But first what is a good WBS.
Here’s some guidance.
This some obvious, but building a WBS, a good WBS is hard work. We’re on a 7 week
engagement to rebuild a WBS for a multi million dollar software project. It is hard work.
These guidelines help us every day.
Remember though the WBS is a cost analysis tool. From the WBS we need to construct the
Time Analysis – the schedule.
Both are needed. And a third description – the Technical Performance Measures are
needed too, but that’s outside tonight's talk.
Here’s a screen shot of MindManager. It’s a simple tool to use.
It has one powerful feature beyond making these kinds of pictures.
It can export (and import) this hierarchy to MSFT Project.
This is a visual way of building the WBS used in Product Development Kaizen shops. This
avoids committing to project before the visual impact of the WBS can be seen.
In the tree form, the participants can move things around with ease. The Kaizen approach
moves the ownership of the product to a public place, removing the personal connection.
Do this for awhile, until you’re sure you want to move to project.
Here’s the schedule produced with a mouse click from MindManager to Project.
Now this sounds simple. Too simple. Why do this.
Well, what if you have 50 elements in your WBS. no problem do it in what ever tool you
What if you have 5,000 elements in your WBS. Well now that’s a different problem.
I mentioned before the WBS Dictionary.
This is the narrative of what done looks like. This is done in MSFT project or in
MindManager. Your choice
But the WBS Dictionary is the core of connecting the Plan and Schedule, with the
Requirements and the system Capabilities.
In this example the Dictionary uses the somewhat limited formatting of Project’s Notes
This is fine, it’s meant to have content, not necessarily style.
Now let’s put all this together.
Starting with the WBS, the terminal nodes are represented in project with Work Packages.
This may not be a term you’ve heard of before. A Work Package is a lump of work that
produces a single outcome. A “package of work” that produces something.
For the schedule putting the Work Packages in the proper order is the starting point for a
credible schedule. If you can get the Work Packages in the right order, with their durations
defined, then you have the start of the credible schedule.
d fi d th h th t t f th dibl hdl
The next step is to not do nay more scheduling in MSFT project. Instead let the Work
Package manager look have the activities inside the Work Package and be done with it.
This is how large Defense and Space programs schedule. There are three levels of schedules
mandated by a government “rule,” DID 81650.
The Master schedule a top level picture what is happening in what order
The Master schedule – a top level “picture” what is happening in what order
The Intermediate Schedule – the sequence of Work Packages
The Detailed Schedule – the day to day activities of the project.
For IT projects, the Intermediate schedule is a sweet spot. One that connects cost with
work and deliverables. One that minimally imposes effort on the team and fits well with
the agile world of Corporate IT software development.
To restate the importance of the schedule again, let’s look at it from another point of view.
A credible schedule is one that describes the work to be done, the order of that work, the
measures of maturity as the schedule moves from left to right.
The units of measure of credibility are not always obvious. But a starting point could be
that the schedule is “believable”
The schedule should show you want done looks like. Not just the work effort.
“Done” needs to be measured in terms of physical percent complete, not the passage of
time and consumption of resources.
Getting the schedule to tell you what done looks like is hard work. And beyond this limited
On our programs creating, ,maintaining, reporting, updating the schedule to make sure it is
Credible is a full time job. Sometimes a full time job for dozens of people are large
C dibl i f ll ti jbS ti f ll ti j bf d f l l
If you’re spending other people’s money it’s actually your obligation to build a credible Plan
and associated Schedule that sows “done” in units of measure meaningful to the buyer.
You can buy SteelRay or you can build the analysis from SteelRay your self.
Either way you must know things about the schedule in order to answer the question – “is
this schedule credible?”
The point of these tools – and there are others than SteelRay – is to have a consistent
assessment of a project “health.” to not rely on personal opinion. To guide the
development of good schedules.
There are several risk management paradigms. This one belongs to the Software
Engineering Institute. The Continuous Risk Management process.
Another paradigm is the Department of Defense Risk Management Guidebook.
These two approaches provide essentially the same framework. One place is the Project
Management Institute's Guide to the Project Management Body of Knowledge®.
Another is the Software Engineering Institute’s Continuous Risk Management process.
I’d recommend the SEI CRM for its completeness.
But along the way remember there are five easy piece of risk management.
1. Hope is not a strategy. Only actionable outcomes
2. All measurements are measures of random variables. You must know the underlying
statistics and variances of those statistics
3. Never forget cost, schedule, and technical performance are inseparable. Change one,
the others change too, in some way.
4. You have to follow a know risk management process. SEI is good, so is the DoD process.
PMI is OK. It’s better than nothing
5. If no one knows your looking after the risks, than you’re wasting your time.
The risk management processes needed for a credible schedule come in two flavors.
1. Technical risk – the risk that the technical behaviors do not meet the expected
2. Programmatic risk – the risk that the project will be over budget and behind schedule.
The technical risks need to be identified in a risk management tool and have mitigation or
retirement activities embedded in the schedule. The reason they need to be embedded is:
• The risk and its mitigations must be visible
• Funding g and resources must be committed for risk handling
• The impact of the risk on downstream activities must be assessable. The only way to do
this is to have the risk handling work in line with the delivery work
The first risk management tool is free and runs in Excel.
It’s the risk register. This is where the risks are held while they are being processed
according to the SEI Continuous Risk Management process.
This tool can be downloaded from the web site along with a manual.
Doing risk management is a skill. It is not about making lists and checking them twice. It’s
about actually managing the risk in the same way you manage work.
The core process for programmatic risk management is a Monte Carlo Simulator (MCS).
This example uses Risk+. @Risk For Project and PERTMaster are two others in use for
A Monte Carlo Simulator starts with a baselined schedule and evaluates each task for a
probabilistic set of task durations drawn from an underlying statistical distribution. The tool
draws samples and plugs them into the DURATION field of the project . For each run of the
simulator sampled durations for all the tasks
simulator – sampled durations for all the tasks – the resulting completion date is recorded.
After many sample runs – 500 to 5000 – a histogram of the completion dates are
The resulting graph shows the probability of the project completing “on or before” a
This information is used to inform the management team about the confidence in the
planned finished date.
The Monte Carlo Simulation tool can be used for cost modeling as well.
Cost modeling is a second order activity on many IT projects. Since they are focused on
writing software and planning to write software. Much of the cost on internal IT is touch
labor. Because of this the cost projects are really absorbed labor projections.
The concept of resource loading is often confused with resource leveling.
The management of resources many times starts with the wrong paradigm. That paradigm
is “what can we do with the resources we’ve got?
Rather, you ask “what resources does the project need, and when do I need them?” If you
take the first approach, you’ll always be behind on the resource curve. Using the first
approach we can know on day one what the resource requirements are. It’s slowly
becoming clear that multitasking is a Bad Idea. So bad in fact, that project failure can be
directly attributed to the failure to manage resources.
So the first thing to do in MSFT Project is to turn off the resource leveling functions.
Instead start with these steps:
• Identify, using subject matter experts, what resources are needed for the work to be
• For each work package, identify how much of a specific resource is needed.
• Do not mutli‐task resources. Everyone knows this. Everyone does this. If you do it, you’re
setting yourself up for disappointment. Multitasks between to projects is difficult. It can
be done, ,but it’d difficult. Multitasking between more than two projects leads to failure.
So don’t do it.
So let’s take a quick summary of where we are.
We’ve got five process areas in every project. I’ve introduced some tools that can be used
for each of these process areas.
What’s next is to look at some examples of how these tools can be applied to the process
areas and what some of the results look like.
But let’s remind ourselves before moving on – it’s not about the tools, it’s not about the
tools, it’s not about the tools.
Starting with the schedule, since we’re here at the Microsoft Project Users Groups meeting.
We need a Work Breakdown Structure as the starting point. Without the WBS we can’t
what it is we’re supposed to be building. The WBS must be product oriented. No Design,
Code, Test type of WBS. That is not allowed in MIL‐STD‐881A projects – the WBS Standard.
If should not be allowed in IT projects.
It’s bad project management practice.
From the WBS a collection of Work Packages can be built. These Work Packages should be
self contained, produce one – or a very small number of deliverables, and resource
balanced by the Work Package owner in conjunction with other Work Package owners.
Arranging the Work Package in a logical order is next. This should be Finish to Start (FS) in
every case where it is possible. Don’t get lazy here, do the work needed to arrange the FS
If it doesn’t work that way for the majority of Work Packages, you’ve probability got the
wrong Work Packages.
They are too “coupled.” The measurement of coupling and cohesion, is a standard Object
Oriented Software Development team. The same is applicable to programs with work
Here’s a sample from a real project with embedded risk management, embedded project
margin and a deliverables focus.
The PERTChart Expert tool provides a view of the project not available through MSFT
Each “band” can represent any sort field on MSFT Project. So the view can be tailored to
the needs to the project.
This bands could be People, Roles, Technologies, Method Phases.
The boxes are the tasks in the project schedule, and they move left to right in time and
Any updates to this file are reflected directly in the MSFT Project MPP file. So bi‐directional
linkage is possible.
The use of Monte Carlo Simulation (MCS) is mandated in many project domains.
This is a powerful tool to reveal the statistical behavior of the project under varying
programmatic risk conditions. The Monte Carlo simulator assesses the schedule by
changing the durations of tasks using values drawn from a probability distribution for each
task. It evaluates the resulting changes in the completion dates and records those for a
large number of cycles – 1,000’s.
The tool them builds a histogram of these completion dates and shows the probability of
completing “on or before” a specific date, from the results samples of all completion times.
Here’s a picture of actual output from tools for a major ERP systems integration program.
Starting in the upper left, the needed capabilities are captured through a Product
Development Kaizen with the stakeholders in an offsite session. These Kaizens are full
contact meetings where the participants work “on the wall” to reveal the capabilities and
the attributes of these capabilities.
Using MindManager the relationships between the capabilities and the accomplishments
and criteria for their assessment are assembled into a Work Breakdown Structure.
From the WBS an Integrated Master Schedule is built describing the work needed to cause
the capabilities to appear.
A Master PERT showing the flow of increasing maturity and the Value Stream for the IT
work. This closes the loop from “need” to “delivery.”
Here’s another example of how tools can be integrated to manage a program using the
guidance of Capabilities, Requirements, Performance Baseline, and Execution.
In this example the evidence of Physical Percent Complete is held in the Quantifiable
Backup Data worksheet in the lower left corner. This information is defined during the
development of the Work Packages and used to assess the physical progress of those Work
All these tools and their interactions with the processes and the people needs to be put in
While tools and processes are critically important, having the right people in the right
“seats on the bus” as Jim Collins would say in “Good to Great,” is as critical. These people
must have a deep understanding of how to put the tools to work on their behave or the
tools are of little value to the organization.
We ve seen this many times, where a program management or technical groups purchased
We’ve seen this many times, where a program management or technical groups purchased
a tool as the solution to their problems. Only to discover they did not understand the
problem, so the tool had no chance of success.
A final wrap up is needed, just to remind ourselves how the tools are connected to the
We talked enough about why tools are important. But remember the tool is “glue”
connecting processes with people.
We can see here the MSFT Project Server is a core component of the tool set for Increasing
the Probability of Success. As well tools that augment Project Server are needed.
Risk management and Requirements management are two other processes requiring tools.
Ri k t dR i t t t th ii t l