Why CEOs can't play chess
I love chess. It always amazes me how such simple rules for piece movements can give rise to matches of such extraordinary complexity and drama. Often I see corporate strategy analogized as a chess game. On the surface, that seems right – managers adeptly try to outthink their opponents in a rich interplay of competing strategies and asset positions.

However, the cold light of critical analysis casts grave doubt on this analogy. CEOs (or anyone in senior leadership) rarely have the ability to command a direct reaction to a competitive attack in the market, or even to oversee core business transactions (if this was the case we wouldn’t need Sarbanes-Oxley). Economists call this the Principal-Agent Problem. What senior leaders really do is set the stage for an enterprise to do its job in the marketplace. In effect, CxOs create structures by which organizations generate their behavior. Going back to our chess analogy, it would work as if our hypothetical CxO would say, “OK, let’s let Joe handle the Knights and the Rooks. Sally will take care of the Bishops. I want Joe and Sally to huddle every move, especially in cases where their pieces are on the attack. Then I’ll have Jim working the pawns…” – meta-chess, if you will.

Jay Forrester, the creator of the discipline of System Dynamics once described the CEO as an “organization designer”. That is a much better way to describe the role than that of a chess player. Simple rules are the core of what it takes to make an organization work semi-autonomously.

So what are those magic simple rules that will cause an organization to double its stock price? The answer lies in understanding the underlying “physics” of a given organization – different for every firm. Simulation models are an important tool in this regard. If one could abstract the organization into a model, one might test a wide range of simple rules, acted upon by “agents”, to determine the implications of structure (sets of rules) on aggregate organizational performance.

Its your move.
- posted by Industrial Science @ 10:30 PM 0 comments   

 
The many uses of the word "model"
The other day Howard and I were driving back from a client meeting. This client had a software system for managing business transactions and continuously referred to it as a “model”. It seemed odd at first, but it got us thinking…what really is the definition of a model?

Unfortunately for us, the word is so overused that it has lost its real meaning. We have to describe what we do to every client who has a different interpretation of our introductory phrase: “we build simulation models for a living”.

There are mental models – images in someone’s head about how a system works. There are models that describe a process or formulation – “Joe has a certain model for doing this”, “ACME Corp’s business model is …”. There are models of real things that are abstractions, because the real thing is too hard to deal with – geographic maps or equations that describe planetary motion.

What we do here at industrial science is business simulation modeling. Our clients are confronted with a bewildering array of complex problems, and our standard response is to create a simulation replica of the organization at hand. With such a replica, one can experiment in ways that would be problematic and risky to do in real life (“what would happen if we tripled our price for widgets?”). We use models as tools for the discovery of meaningful insight into the behavior of complex systems like companies and markets.

Some of you may want to learn more about simulation modeling. Let me suggest the following reading list:

1. For a good overview of how models and simulations are used in a corporate setting, read Serious Play by Michael Schrage.

2. For an excellent and entertaining primer on agent-based models, read Turtles, Termites, and Traffic Jams by Mitchell Resnick.

3. Stephen Wolfram spent 10 years of his life conducting exhaustive research and model building for his 2002 book A New Kind of Science. If you don’t have time to read all 1200 pages, I suggest you read at least the first 250. An extraordinary body of science.

4. The “bible” of the discipline of System Dynamics is Business Dynamics by John Sterman. Dr. Sterman was my professor at MIT and is quite possibly the smartest man on the planet Earth (Sterman and Wolfram are too close to call).

Once you have tackled this library, I suggest you read the newspaper. If you truly have a modeling mindset, you’ll begin to see “structure” in every article. This kind of thinking is a good test of your progress as a practitioner.

So what is the process for building a model? Are there good models and bad models? What problems are appropriate to a modeling approach? Stay tuned, readers – we will be covering these and more as we describe our experiences working in this amazing field.
- posted by Industrial Science @ 8:03 AM 0 comments   

 
2005: The Year of the Blog
The current BusinessWeek boldly claims that "Bogs Will Change Your Business". The "blogsphere" is filled with all kinds of idle chatter, political ramblings, musings of people you may have nothing in common with. Then there are so-called "corporate blogs". What are such crazy notions? Why would a company want to establish a blog? Will the lawyers allow us to do such a thing?

The BW article is a good read and has some good "Blog 101" information (including RSS, software, podcasting, links to blogs of note, etc.). To make it even more authentic, the article is written in the form of a blog, with sentence fragments, abbreviations, date/time stamps, and of course, links. The only thing it lacks are comments from readers (more on that, in a second). Linking to the article will place you squarely in the middle of the blog phenomenon. As if to prove a point, the authors end the article by "birthing" a real blog (with the ability to post comments) called "Blogspotting", as in Trainspotting. Yes... a blog about blogs (a meta-blog). A blog you can link to from an on-line version of a print article on blogs. It can be very dizzying.

Back to the Industrial Science Blog. What are we trying to do here? We are constantly looking for applications of the right technology to address business and management issues. So whether it's mathematic/scientific tools such as agent-based modeling or Monte Carlo simulation, or on-line tools such as IM or the use of ftp sites to get around coporate email attachment policies, we are always solving problems. We experiment and try new things to promote and elevate the use of scientific methods and tools for addressing business issues.

So we invite you to visit often. Welcome to our clients, friends, mentors, associates, competitors, corporate lawyers, conrtibutors, and members of modeling/simulations community. As for the title of this post? So maybe you started your blog in 2004 or are already part of the phenomenon (if so, you are way ahead!). 2005 will be the year that you will start reading blogs on a regular basis. Perhaps you will start your own, or make contribution on this or other blogs.

The journey continues.
- posted by hcpark @ 10:24 PM 0 comments   

 
catching a fly ball / doing physics in my head
Earlier this evening, I spent a little time with my two-year old playing ball. I threw long arcs with a very soft ball and although he couldn't catch it, he was starting to track the ball in the air.

This reminded me of something I read that discussed how smart we must be to be able to compute 2-dimensional physics equations in our head. Somehow, we know where the ball will land and place ourselves there well before the ball actually arrives. Of course, this is done through experience (trial & error), since most of us will not take measurements, count steps and do s = ut + 1/2 at² in our head. As I was throwing this green soft ball in large arcs, I thought that in reality the problem is actually much easier and at the same time much more difficult.

The problem is made more difficult than our Physics 101 problem since we have to deal with things like wind and terminal velocity. Also, we have height (y<>0), and our eyesight is not connected to the glove itself. In order to do the actual math, we would have to take all these (and a few more) factors into account. But we have a secret weapon. It's more powerful than the equations and ability to accurately measure things on the "fly".

It's called feedback. We can make a good guess (from experience), and make adjustments as the ball is in the air. There's a reason why Coach told us to "keep our eye on the ball". We don't have to recalculate -- we don't even have to do the initial calculation. We can guess, and make smart adjustments.

Now, this is not earthshattering news. This is Baseball 101, Chapter 3: "how to play outfield". Feedback is built into our human nature. It's amazing how well we learn and solve how to do things.

Unfortunately, many models ignore such feedback mechanisms. Think of pricing models that ignore competitor's responses; tax models without the ability of people giving up their US citizenship to avoid taxes. Feedback is an important part of many systems. The world is not static -- the agents in a system are not static either. They react, they learn, they try different things, they fail, they succeed.

- posted by hcpark @ 9:15 PM 1 comments   

 
Predicting the Future
I am often asked by clients, colleagues (and even family), "can your models really predict the future?".

It's enticing to think that it's possible to know, with absolute certainty, how the future will play out--in a business dealing, in world politics, in your stock portfolio, tommorow's game, or tonight's dinner. If we're "smart", and we make our models "smart", couldn't we use it to tell us what will happen? In fact, the early days of computing was marked by an attempt to better predict (and perhaps control!) weather. We know today that our weather reports can be wrong.

"We're not in the business of predicting the future. Instead our models help you prepare and, if you're bold enough, help you shape your future." is the sort of reponse I give. In fact, we find that in many organizations there are varying ideas of what the future is and why it's important. Even when confronted with lots of data (and perhaps because there is too much data) it's difficult to put it all together into a coherent viewpoint.

Sometimes its best to think about the range of possible futures to see what's possible. Think of the three ghosts who visit Scrooge--there is but one future (if we ignore the parallel universe argument), but I'll show you what COULD happen. Why is this important? Becuase it causes Scrooge to change behavior NOW. Dickens ends the story here without fast-forwarding to the actual future. He doesn't have to. We the readers "get it".
- posted by hcpark @ 9:23 AM 1 comments   

 
Welcome to our journey
We are entering yet another exciting chapter in the history of Industrial Science. Founded just under three years ago, we have witnessed an amazing number of complex business organizations and the problems that they face. We have learned a great deal, and this forum is ideal for us to share our experiences with you, as well as to get your reactions and thoughts.

Our mission here at Industrial Science is to help our clients navigate the extraordinarily complex world of engaging durable strategies in a constantly changing climate. The answer, in a word, is science. Yet there also exists a great deal of "art" in the application of appropriate science to the particular problem at hand. Moreover, the process of building computer simulations (our primary tool) is as valuable as the numerical result. What we are privileged to see on a daily basis is astounding - and too important to keep just to ourselves.

In this blog, we will be writing about a whole range of subjects under the broad umbrella of the application of science to business strategy and operations. We will be inviting guest authors whom we consider to be undiscovered thought leaders, to weigh in as well. And finally, we invite you to come along with us, by granting us the favor of your reply to our posts.

Let the journey begin.

George E. Danner
President
Industrial Science, LLC
- posted by Industrial Science @ 8:09 PM 0 comments