You are on page 1of 3

(lighthearted techno music)

- Hi and welcome to Proof that Agile Works


From Spacecraft to Supercomputers.
The first story that we're gonna talk about
in terms of proof that Agile works is Skunk Works.
Skunk Works was the program that produced
the P-80 Shooting Star, the first jet fighter.
This was a program that was developed
under Lockheed Martin in World War II.
Kelly Johnson, the chief architect
for this new aircraft, was given the task
of building the first jet fighter in 1943.
This had never been done before, so in order to do this,
he took his whole team, and he put them into a tent.
They came out with a working prototype
within 143 days, the first working model.
The idea that we could produce an aircraft
of a brand new quality that could be a jet fighter
that could go into battle within less than half a year
is crazy, but it was done back
in World War II using Kelly Johnson's approach.
When we look at the keys success factors,
what we see is that a lot of these
success factors actually match what we propose
is the right way to do work in Agile.
In fact, when we take a look at the key tenets,
we see that he had strong, self-directed,
cross-functional teams; that he had owners
and vendors on the same team, and they trusted
each other, they emphasized that.
They were willing to manage and respond to change
because remember, they had never done this before.
They had never built a jet fighter.
As a result, they would have to minimize
the reports but still record the important work.
This is because so much work was being done concurrently
that everyone in the tent was allowed
to come in and adjust the architecture
as learning and improvement occurred.
Finally, with this incremental development
and self-testing approach, they were able to quickly
identify if they were getting a working product.
This required enormous amount of trust between
the Air Force and Lockheed Martin and Kelly Johnson's team.
If we take a look at how this compares
to the tenets of Agile, we see that it matches very well.
A shared, robust vision that is able to adapt
to change with whole teams that are doing incremental
delivery and continuous testing and integration.
They mirror exactly the principles that
were in Kelly Johnson's Skunk Works program.
Skunk Works continues to deliver with success today.
Now, this next program is actually one that
I was on, and I ran this program
for about four years when I was working
for Booz Allen Hamilton at CNIC.
CNIC is the Commander Navy
Installations Command for the Navy.
It's kinda funny, it's got Navy in the name twice.
This is a branch of the Navy that invests
in its facilities in order to reduce energy use
and also to switch from brown power to green power.
Now, this program was all about picking the best projects.
The way that this worked is that first, projects
were sent to the headquarters inside of templates.
Those templates would then evaluate the projects
based upon a hierarchy of criteria that
would produce scores like net present value,
energy savings, the production of new
enabling infrastructure for future projects.
After it went through our optimization formula,
we would come up with which projects were good
and should go ahead and get funded,
which projects needed to be taken a look at again,
and then which projects were just bad,
and they were never gonna get funded.
The output was a selection of the best projects
and the rejection of those that didn't meet muster.
Of course, only the best projects would get the money.
Now, in terms of the basics of this program
and what you should know, this was a program
for the Navy, and it was run as a contract
with Booz Allen Hamilton when I was the project manager.
The scope was to help invest half a billion dollars
a year in terms of investments in the shore.
Much of that money was through third-party investment,
and about a third of it was through direct investment
by the Navy with tax-funded dollars.
The cost of the program that we were running,
the decision support program, was just $5 million over
the course of four years, so about $1.25 million per year.
We had two cross-functional teams of developers
and analysts, which included eight BAH
personnel and five Navy personnel.
So, this was a small team, but in the end,
we were able to produce incredible return on investment.
For every dollar they put into our program,
the Navy got back $50 in exchange.
The way that we did that was first,
we made the data better, and we were able
to save $20 million in bad projects,
followed by we were then able to improve
the selection of projects, so that we were to gain
an additional $30 million in revenue
in terms of saving energy costs.
From there, we were then able to identify
and continue to improve the cycle of investment
with additional modeling to identify the next project
as well as to track the project success.
Now let's talk about projects that are even greater scale.
We're gonna go from the Air Force
with supercomputers to NASA and its spacecraft.
First off, the Air Force build one of the most powerful,
high performance computers, the Condor Cluster.
This Condor Cluster strung together PS3s
with over two million miles of cable.
As a result, it was extremely efficient
and very powerful, able to be flexibly put together
in a lab, and it could even be put on a plane.
This allowed for the real-time processing
of imagery for operators in the field
that had never been provided before.
In fact, they were able to go from
days of processing time to seconds.
So, that actual intelligence was given
to operators, so they could use it in real time.
Then there's the NASA missions under
the Faster, Better, Cheaper initiative.
In total, there were about 10 successful missions
during this time, but two that really stand out
are first, the Stardust mission,
a mission which slung shot a spacecraft
around the Earth and then around the sun,
so that it could trail and catch the stardust
coming off of a comet, as well as Shoemaker,
which was actually a spacecraft that was intended
to fly by an asteroid, but it landed
on the asteroid instead, taking high density readings
that had never been thought possible.
Both of these missions were done
under budget for a tenth of what it cost
to do a mission today using Agile principles.
Because of the reuse and because of the ability
to limit scope during development,
NASA was able to produce many working missions,
10 working missions for the cost of one that it costs today.
So, I hope that these examples
have inspired you and given you some good food for thought.
It should be able to allow you to talk
about Agile and its success in any context.
I look forward to seeing you in the next lesson.

You might also like