Choosing in conditions of uncertainty
Feedback from the cafe
As each chapter of this book is written, copies are sent to a virtual cafe. This virtual cafe consists of about six tables of readers who each get emailed a copy to read. More shall be said about this virtual cafe later, but, sufficient to say at this time that the tables provide valuable feedback and inspiration for subsequent chapters.
Two responses to the first two chapters need to be addressed before we carry on. The first was from Mark Baartse:
My understanding is that at a very simple level, decision theory is basically a tree diagram. You map out different possibilities, different outcomes, etc., in a tree. As you can imagine, with lots of options and lots of outcomes, the tree can get pretty complex.
Branches are created as an activity to do this or do that: basically, a potential 40% chance we get the deal with x and 60% we don't. You also attach values to each event or activity (eg decision: we do a TV campaign for 5m or a radio campaign for 1m). At the end, you will have multiple (potentially dozens) of options with a value attached to each which allows you to create the best decision.
My question is: what is the advantage of game theory over this tried and tested system that's been taught in business schools for decades? Why doesn't this system work in e-business?
Sure, there are a lot of possibilities and possible outcomes, and yes, you often make a best guess (60% you'll get the deal? a bit subjective), but isn't it quite similar to game theory - you're picking the one most likely to be profitable based on information at hand?
Although the method described by Mark may seem to be very similar to the technique of selecting an appropriate e-business niche - described in chapter two - it is not suitable for environments where there are unpredictable uncertainties. This is because the method relies on a reasonably accurate estimation of the probabilities involved in a problem. This confines the method to reasonably stable environments - such as those found in much of the business environment of the second half of the twentieth century.
Game Theory uses a different kind of conceptual framework, which can deal with situations where calculations of probabilities are not always reliable . The difference is subtle, but, critically important. If you look at Mark's description of qualitative analysis he begins: "You map out different possibilities, different outcomes, etc...".
In a reasonably predictable world, where you can use a structural top down approach, you are able to isolate a range of options, complete with information relating to possible outcomes, and use a statistical approach to choose between them. In the world of e-business, not only is it virtually impossible to calculate probabilities, some of the best options often emerge only after projects are underway.
Events move so fast in the world of e-business, that whatever happened in the past doesn't necessarily have any bearing on what happens in the future. Think how many businesses were planned around charging for Internet Services. The idea of creating a customer base, of several thousand people paying a few hundred dollars a year each for an Internet link, seemed an attractive basis for a viable business enterprise. Many entrepreneurs bought hardware and designed software specifically to cater for this potentially lucrative market.
Unfortunately, for these early start-ups, there were so many people setting up Internet Services with the same idea, that competition soon forced the charges down to around one hundred dollars a year. This diluted the profits, but, the situation was still viable. So the rise in the number of I.S.P.s continued to grow, keeping up with exponential growth of the Internet.
Then, a few large companies realised that these surfers would be spending money on the Internet. A back-of-the-envelope calculation soon revealed that the cost of supplying people with Internet Service would be a mere fraction of what they would likely be spending buying goods and services through their Internet connection. To gain access to these millions of potential customers, and their combined buying power, it seemed economically sensible to provide them an Internet Service for free. As subsequent equity issues proved, the market places a very high value on what was then seen to be a captive audience of receptive buyers who had the means to purchase at the touch of a button.
The many thousands of start-ups, who had based business plans on gaining revenues by supplying Internet connection services, were left with egg on their faces. Their lucrative market disappeared overnight.
The more adaptable, saw the potential for providing Web space for people and businesses. Very soon, most of those who'd lost or were losing their ISP customers to the free services, were including free Web space in with their ISP services. It didn't take long for the big companies to work out that the relatively small cost of providing Web space for their customers could also be offset by their buying potential. They then began offering free Web space to all their customers.
To compete against the free services being offered by the giants, the smaller hosting services began designing special server side software so that they could keep and attract customers by offering better service and specialised facilities. This spawned a host of software development companies, supplying this server side software to everyone. It wasn't long before these specialised services were also being offered for free by the larger companies, who were by then aggressively competing with each other to build the largest portals.
(Note: there are many types of portals and there is as yet no single agreement as to exactly what the term means, but, in this book the term will be used to mean a point of entry for many people to access an organised sector of the Web. We shall be dealing with portals more specifically later in the book).
With free, or next to nothing, Web hosting available, hordes of people were attracted into e-commerce. Multitudes of Web site design companies grew up to serve this massive potential market. At first it provided a lucrative business for Web designers. It attracted hundreds of thousands into the business. Even high school kids were jumping onto the bandwagon to cash in on the gold rush.
Competition soon started to drive down the cost of Web design and construction. Once again, the envelopes came out and the more savvy Web designers figured out that it would pay to design an initial Web site for free - as this could provide them with a stable base of satisfied customers who would most likely turn to them when they wanted to improve their site or expand their Web based businesses.
Professional web site authoring packages then started to come into their own. Complex Web sites could be designed within the environment of these authoring tools, cheaply and efficiently by anyone who took the trouble to learn how to use them properly. Templates could be made from sophisticated Web site designs, making it easy and inexpensive for similar sites to be produced on mass..
The cost of customising these clones for different clients became minimal. Pretty soon, eTailing sites - which had started out costing clients hundreds of thousands of dollars to be designed specifically for them - were being offered for a fraction of their original design costs to all and sundry.
As entrepreneurs realised the potential of hosting e-business sites - complete with their customer bases - out came the envelopes again. These valuable clients could be captured. It didn't take long for the software mechanisms involved in e-commerce to be offered for free: together with free Web space. There were minimal costs involved in customising these eTailing sites for individual businesses, so, it made economic sense to offer eTailing sites free of charge. Soon, there was a flurry of activity as companies raced each other to create virtual malls, providing traders with free eTailing sites - complete with shopping carts, ordering systems, inventory controls, back end data base management - all completely free of charge.
Now, how can a decision theory approach deal with this kind of situation? How can any future plans be based upon experiences in the past? This is a great enigma for Industrial Age strategists who are used to planning in a predictable universe. How can they plan and organise projects where there can be no idea as to what is going to happen in the future? This is why the Information Age strategists are now turning to Game theory.
Game Theory assumes that the future is uncertain and unpredictable so, unlike decision theory, which starts by making a list of probable solutions, it starts with all possibilities - however unlikely and bizarre - and gradually pares this infinite choice down using broad principles and taking calculated risks.
The difference then, between decision theory as described by Mark Baartse and Game Theory as applied in chapter 2, seems to be that decision theory works by means of intelligent selection whereas Game theory works by intelligent elimination.
Note: This may seem to be a strange way to interpret Game Theory because Game Theory is usually discussed in terms of rules and pay-off matrices. But, an elimination is in fact a rule that takes the form: "If such and such condition holds true, do that or don't do this".
This contrasts subtlety with decision theory, which takes the form: "Do this in preference to that". Which is a selection procedure, rather than a method using rules based upon conditionals.
As we shall see later, Game Theory can be interpreted and applied in many different ways.