Project Management


I think I delivered some shocking news at the Woodinville City Council meeting last night.
youngwomanamazedinsunglasses

During Public Comment I shared some data provided by Realization on their website.  This data documents some of the successes organizations have had when they implement Critical Chain.  For the record, while I believe it’s fair to consider Realization as the top-tier provider of Critical Chain implementations, other companies and organizations that have implemented Critical Chain have reported similar results.

Here are some of the examples I shared last night, with results before and after Critical Chain implementation:

  1. Medtronic USA
    Before: 1 software release every 6 – 9 months.
    After:  1 software release every 2 months.
  2. Medtronic Europe
    Before: Average cycle time 18 months.
    After:  Average cycle time 9 months.
  3. Valley Cabinet Works
    Before: 200 projects / year
    After:  334 projects / nine months
  4. BHP Biliton
    Before: 25,800 man-hours to be completed in 8 months.  2 weeks late.
    After: 19,500 man-hours used.  Projects finished 3 weeks early.  25% productivity increase.
  5. HP Digital Camera Group
    Before: 6 cameras launched per year; only 1 on time
    After: 15 cameras launched per year; all were on time; R&D 25% lower

When you’re inside an organization like the City of Woodinville and enmeshed in the day to day affairs of the City, it’s easy to believe that big jumps in the performance of the organization are not possible.   And so perhaps it’s a bit shocking when reality proves that other organizations, far more complex than yours, have made big jumps in their performance.

Please note that Realization provides many customer testimonials in video format on their website.  If you want to see some of the full presentations you can contact me.  For a period of time, Realization was giving away DVD sets from their yearly conference.  I have some of those for recent years.

On a related note, the Japanese Government’s Public Works reform, for example, has mandated that all contractors working for the Ministry of Land Infrastructure and Transportation must use Critical Chain.

You can also find references to Critical Chain in the Project Management Institute’s “Project Management Body of Knowledge,” also known as the “PMBOK.”  It’s important to understand, however, that because Critical Chain’s “new rules” for managing projects sometimes clash with what PMI has been preaching for many, many years, PMI is not yet embracing Critical Chain as vigorously as (in my opinion) they should.

Finally, I can give you a personal data point.  When I have used Critical Chain to manage projects it has always worked “as advertised.”  It’s hard to put into words the level of focus that Critical Chain (and Buffer Management, which is a critical-but-sometimes-overlooked aspect of Critical Chain) brings to project planning and execution.

I shared this data with the Council last night.  I asked them to look in to Critical Chain.  My hope is that some of them will do this.  If they do, I think they will find that the City of Woodinville could complete projects in about 2/3rd’s the time that those projects would take if managed as projects are managed today within the City.

Finding success stories on Critical Chain is not difficult.   Any City Council member that can use Google can find such results.

Understanding how Critical Chain delivers its improved performance is more challenging.  Council Members that want to understand how Critical Chain really works will have to invest a little “mental sweat equity” to do it.  However, once this investment is made, I think they will understand why I consider Critical Chain to be an integral part of the “pot of gold” that I believe is waiting to be claimed by any City Council Member (current or future) that is willing to do the hard work required to claim it.

Dr. Alistair Cockburn

Dr. Alistair Cockburn

Just a short note — and one that is not totally about the City of Woodinville.  I know that I have been on a tear on that score lately.  But I can’t help it — improving local government is important to me.

This note however is about some very good news.  In the latest edition of “Crosstalk” (Crosstalk is the “Journal of Defense Software Engineering”) Dr. Alistair Cockburn has written an article (“Spending Efficiency to Go Faster“) on how Theory of Constraints principles can be exploited to improve software engineering.

Many of my friends and clients are involved in the software engineering of complex systems.  They are recognizing, more and more, that the Theory of Constraints is a powerful tool for systems improvement.

Our ability to improve is unlimited.

In a previous article I claimed that “Value delayed is value denied.”  This article continues this theme.

In this article I show how multitasking causes the delivery of significant value to citizens to be delayed and thus, ultimately, denied to the citizens.

Value delayed is value denied.

Second, the belief that “The sooner we start, the sooner we finish” is clearly wrong when we are talking about projects being completed by a finite set of resources.  Why does this matter?  It matters because operating on this assumption causes people to drive the system (the City) into greater and greater levels of multitasking.  Today, the release of work to the system is not done properly and as a result, the system is choked with work.

Systems that are choked with work give the illusion of being highly efficient.  But this is only because we are making the assumption that “If everyone is busy all the time, then the system must be operating in a highly efficient manner.”  This is not valid.  The people in a system can be fully utilized and still not be doing anything productive.

Figure one shows an simple, idealized case from the world of projects.  We have three projects (red, green and blue.)  Each project requires two years of focused effort to complete:

womtWhen these projects are executed without multitasking, the red project finishes two years after starting, the green project two years after that, and the blue project finishes at the end of the sixth year.

Of course, no real projects fit into such nice, neat packages, but let’s stick with this idealized case for now.  We can always make things more complex later.

Now, look at figure two:

wmtIn this case, we have the same three projects, and each project still requires two years of focused effort to complete.  However, in this case, we are showing the effect of multitasking, where the resources work on one project for three months, then switch to another project for the next three months, and then to the next project, and so on until all projects are completed.

Notice that the total time to complete all three projects is still six years.  Note that there was no loss of efficiency in this idealized case, although in the real world it’s unlikely that we would be so fortunate.  In the real world, multitasking does cause a loss of efficiency, and the loss is not small.  But still, it’s a secondary effect to what I’m writing about here.

So multitasking didn’t cause a loss of efficiency.  But wait a second — do what want the City to be efficient, or is being efficient really just a requirement (a necessary condition) to achieving something else?

In my view the goal of the City is “To deliver more value to citizens, now and in the future.”  I don’t care if the City has to be efficient to achieve this or not.  If they can achieve it without being efficient, fine.  If they need to be efficient to achieve it, then again, fine.  Either way, the standard that I hold them to is whether they have achieved the goal.  That is what matters.

So when we consider these two very different modes of operation (multitasking vs. non-multitasking) is there any difference in value delivered to citizen?

Well, in the non-multitasking case (figure one), the citizens started to receive the benefits of the red project at the end of year two.  At the end of year four, they started to receive the benefits of the green project, and at the end of year six, they started to receive the benefits of the blue project.

But look at what happened in the multitasking case (figure two)!  Even though there was no loss of efficiency, the citizens were denied the benefits of the red project until the end of June in the sixth year.  They were also denied the benefits of the green project until October of the sixth year.  There was no change in when they began to receive the benefits of the blue project.

It is equally clear that the assumption “The sooner we start, the sooner we finish” is not valid when multitasking is in effect.  Look at the green project.  When we allowed multitasking, that project started sooner, but finished later.  The blue project also started sooner, but still finished no earlier than before.

Our concern should not be when projects start, but instead, when they finish.  The City Council should be concerned as to whether citizens are getting all of the value they are paying for, as soon as it can be delivered to them.  Members of the Council should know whether the City is working projects in a manner that guarantees that the projects that will generate the greatest value to citizens are the ones which will be completed 1) first and 2) as quickly as possible.

(As an aside, let me say that in a future article I will write about small, localized tweaks to various policies, and why such tweaks will be utterly insufficient in terms of bringing the City to the point where it is consistently achieving it’s goal.)

Let’s say that in this example (which is again, an idealized case), all projects are equally costly in terms of the investment that must be made, and in terms of the benefit that each will begin to bring when it is completed.  Let’s assume that each project costs $6M to complete, and that over a useful life of 30 years, the project will create $18M in value to the citizens of Woodinville.

If the project costs $6M to complete and ultimately brings $18M in value, then the project creates +$12M in benefits (value to citizens.)  If the useful life of the project is 30 years, then the project generates value to citizens at the rate of $12M / 30 years = $400K per year.  Again, this is an assumption.  Every project will have a payback curve that we can only really guess at.

Now, in the case of the red project, the delivery of value to citizens is delayed by three years and six months, due to multitasking.  This amounts to $1.4M in value that is totally wasted.

Completion of the green project under multitasking is also delayed, in this case by one year and nine months.  This amounts to another $700K of value to citizens that is totally wasted due to multitasking.

Adding the two losses results in about $2.1M in value to citizens that is wasted and cannot (for all practical purposes) ever be delivered to them:

Value delayed is value denied.

Now, you can and almost certainly will object to this informal analysis.  You can attack it on many different fronts and in all cases I’d love to read your reasoned comments.

My bottom line conclusion, however, is unlikely to be changed.  That bottom line is this:

There is a world of difference between being efficient and being effective.  There is much that the City could do to operate in a manner that is far more effective than is the current mode of operation.  But the quest for efficiency leads good managers to do the wrong things.

The real solution to the problem outlined here is to recognize that the City is a multi-project environment.  It should therefore be using a multi-project management method that has been shown to be effective in representative environments.  If this was done, the City would deliver more value to citizens sooner, and, as a pleasant side-effect, the level of chaos present in the City’s operations would drop by 2/3rds or more.  The City Manager and the City Council would also (routinely) have the information they needed to manage well.

Finally, it’s important to understand that the people managing the system today are genuinely trying to do the best possible job for the citizens.  In fact, in a perverse way, we would likely be better off if they were not quite as fastidious as they are today.  If you know the problems inherent with the so-called “Waterfall” model of software development, you can get a sense of what I mean by this.

The future isn’t written yet.  I believe that Woodinville could be a model for every other city in the United States.  To become such a model we have to change quite a few things about how the City is managed.  Such change won’t come easily.  But it is possible and it is utterly worth the fight and the struggle that will be required to achieve it.

usmcvideoI’ve been on a tear the past few days writing primarily about our city government here in Woodinville.

I’m continuing this trend but in this short post I simply want to call attention to a (by now well-known) video describing another TOC success story.

The video describes outstanding results achieved by the US Marine Corps repair facility in Albany, Georgia.  This facility overhauls complex equipment for the Marines.  This is a highly complex project environment.

You might think our small city has almost nothing in common with such an environment, but if you did, you’d be wrong.  Both the City of Woodinville and the repair facility are examples of complex multi-project environments.  The big difference is that, today, the repair facility is a high-performing environment where projects flow through the facility quickly and where managers have the information they need to manage well.  We are not yet at this point in Woodinville, but I’m hopeful that someday we will be making our own success story video.

The video can be found here.

denied_1In a previous article I wrote “It’s another thing when the delivery of millions of dollars of value to the citizens of Woodinville is significantly delayed.”

In this article I begin to explain the basis for this claim.

Let’s assume for a moment that both the City Council and the City Staff are competent individuals who are striving to “manage well.”  (Yes, I know, not everyone believes this.  It is so easy to point to “bad people” as an excuse.  Let’s not go there right now.  Thank you.)

Now, one of the most common assumptions in organizations is that “A resource standing idle is a major waste.”

You may believe this yourself. For example, if you hire someone to do some work for you, and that person shows up but refuses to work (he prefers to be idle) then I think it’s likely that you will have a problem with him.  I certainly would.

So, when we hire people (I’ll often use the term “resources”, even though it’s a bit dehumanizing) we usually have an expectation that when they are working, they are producing for us, and when they are idle, they are not.

This assumption (“A resource standing idle is a major waste”) is generally true when we are talking about independent resources. For example, if you hire a plumber to fix a leaking faucet, he is typically an independent resource — he (usually) doesn’t need anyone else to do anything in order to be able to do his work.  If he needs to get a washer for your faucet, he goes out to his truck and gets it.  He doesn’t wait for someone to make a washer for him.

However, when we start to talk about dependent resources, the situation changes drastically. When we have dependent resources, it is no longer necessarily true that “A resource standing idle is a major waste.”

The City is an example of an organization with a lot of interdependencies. For example, only some of the City’s resources handle building permit requests. Other resources are the only ones able to sign certain kinds of requests, such as purchase requests. Other resources are the only ones able to create a master budget, do engineering work,  decide a personnel matter, or  respond to a request for IT support.  Note that the members of the City Council are also dependent resources, from this point of view.

There is also a significant amount of variation in the time required to complete many of the tasks associated with a given project. For example, when preparing documentation for an upcoming Council meeting, this task will sometimes go smoothly and not take a lot of time, but at other times, it may be a struggle and take much longer than anyone expected at the outset.

These two factors — dependency and variation — conspire to make it almost impossible to ensure that a given resource will not have to sometimes stand idle from time to time.

Now, smart people (and the Council and the City Staff are full of smart people) will find a way around this problem. They will find a way to ensure that no resource ever has to stand idle for any significant period of time.

The way to do it is to simply flood the organization with work. Just start far more projects than the City has the capacity to complete in a timely manner and, poof, the problem of idle resources vanishes into thin air.

All of a sudden, no one has to wait anymore. Can’t find a way to make progress on your current project? No problem, just drop what you’re currently doing and pick up another project for a while. Eventually, when you become blocked on your current project, or someone else needs you to do something for them so they don’t have to become blocked, just set your current work aside and switch to another project.

After all “work is work.”  Somehow, it must be that it “all comes out in the wash.”  Certainly, if every resource is working all the time, then there is no waste and we are thus guaranteed to be running the organization in a highly effective manner.

And of course, we know that the sooner we start a project, the sooner we will finish it.   So what’s the harm in starting a number of projects concurrently, in order to ensure that we have enough work available so that resources are never blocked?  Everywhere we look, we see other organizations and businesses operating in this manner.  If everyone is doing what we are doing, isn’t it obvious that we are on the right track?

Of course, the answer is that the assumptions we have just mentioned (“work is work”, “it all comes out in the wash”, “the sooner we start, the sooner we finish”) are actually invalid under the conditions which actually obtain inside City Hall.  These invalid assumptions result in the City operating in a more wasteful manner than is necessary.  And this is what drives the problem that leads me claim that the delivery of value to citizens is being delayed.

When the delivery of value to citizens is delayed, that delay cannot be recovered.  A project that could have been completed in 2009 but was delayed until 2013 cannot go back in time and deliver the value it was supposed to deliver in prior years.

Value delayed is value denied.