Clues from the World of computer programming
The Christmas tree light problem
The top down, object oriented approach is dependant upon having a planned structure in place. The system has to be visualised structurally in order to be able to decompose it into appropriate objects or modules. This requires that the basic framework, messages and modulular functions have to be planned beforehand.
Using the strategy of spreading risk, any problem areas ought to be covered and allowed for. However, the top down, structured approach has a few inherent problems that it is difficult for even a risk spreading strategy to deal with. It has to do with the effect different parts of a system have upon each other and the performance of the system overall.
We have discussed above how a system can be decomposed into different modules. For the system - and by that is meant the e-business or e-commerce project - as a whole to perform effectively and efficiently all the modules must be working properly. If any one fails, like the link in a chain breaking, the whole system fails.
Seems straight forward enough: to fix it, you just have to mend or replace the link. However, it may not be that simple. Think of a set of Christmas tree lights connected up in series such that if one bulb fails all the lights go out. To solve the problem you'd have to get a new bulb and gradually go along the string of lights replacing each bulb in turn with the new bulb until you reach the faulty bulb. When the faulty light is replaced, all the lights come on again and you breathe a sigh of relief.
Now imagine what would happen if more than one bulb were faulty? Supposing the new bulb itself was faulty? Perhaps the new bulb and several of the original bulbs are faulty? Suddenly, you'd find yourself with a very difficult problem on your hands. How would you attempt to solve it? There are just too many variables involved; the Christmas tree lights would have to be trashed.
This is the fundamental problem of any structured pre-planned design in a fast changing competitive environment. If the system doesn't perform properly it may be too difficult to spot where the fault or faults are located.
The perplexity of this problem is delightfully analysed by a discussion at one of the tables in the virtual cafe where the reviewers were reading the chapter on risk spreading. The discussion centered mainly around three reviewers. Here are excerpts from that discussion:
Chris Heape wrote:
I've taught design students for a number of years now and I try to get them to understand or be aware of the necessity for a design process, when tackling a design project. Apart from introducing them to an awareness of a structure over time, it has been very important to enable them to utilise their world of associations to come up with a richer form or visual world of expression.
This inevitably leads to the description of "Killing Your Darlings" That it just isn't sufficient to proceed with a design project if you've got only ONE idea. What will happen if that idea fails?
I've always encouraged them to come up with at least five *different ideas* that they develop parallel to each other in the design process, and along the way to drop the weak ideas and to use what they can to get a final result.
My point is that I've always used as my teaching motivation the notion that the students should touch on as broad a base of their world of association as possible. What occurs to me now, after reading this last chapter, is that what I've been asking them to do is "to hedge their bets". To allow the students to *Plan for Partial Failure*. That they give themselves enough areas of potential that they can allow some of these to fail in order to come up with a successful design.
The book is about e-commerce and as chapter after chapter comes in I'm gaining a far greater understanding of the *Potential* of this area and at the same time getting a confirmation that my way of thinking can be used within this area. You're making me aware that this isn't an exacting science that I have to learn. No! I can use my own human potential and intuitive understanding to get hold of this new world of possibility and commerce. I can't help feeling that this is where your book will succeed. You're touching other people's intuition and understanding that might be contrary to the up-to-now or Industrial Age's way of doing things, but which now has the potential of being put to use in a fruitful way.
Ben Rollason wrote:
I agree with the concept of stripping away elements, so as not to cloud issues, but sometimes I feel in this sheer compelling simplicity, some of the subtleties of situations are lost. Some of the more fuzzy concepts are overlooked. Swings and roundabouts!
In the last chapter, I found the concepts of success and failure rather black and white. Of course, this is necessary in order to illustrate a mathematical point. But personally, if I were to fail or succeed all the time, I would feel like a great success. The hardest thing of all is to succeed partially... Mediocre success. Failure can be a success too, as the entrepreneur in the probability example is pointing out, but you have to fail well enough. The worst that could happen to him to foil his plan, perhaps would be if one of his enterprises half took off...Stick or twist?
Chris Heape wrote:
I do feel the operative word here is compensating. Or trying to compensate for the unknown. I used the term "hedging their bets" when considering an area of risk that isn't involving such large strategies that, for example, investment companies use on a daily basis. But I feel that the point is still the same. That one starts out on a project or venture with several different fronts, in order to achieve a success, because one recognises the fact that some of these fronts or paths are going to lead to a dead end. That one actively contributes to one's success by planning for partial failure.
Ben, you wrote "partial success". I really feel that there is a fundamental difference. By planning for partial failure, one accepts that, along the way one, can cut away the deadwood or weak parts of a fledgling venture, and this will leave you with a few healthy shoots that can then grow and emerge as a fully fledged business.
By planning for partial success, I feel that it is implicit that one will never succeed fully. Some of Peter's phrases spring to mind: Emergent solutions. A humility of spirit. This I have always felt is at the heart of the matter. That one should have a respect and a humility for the chances and opportunities that present themselves, having set a process going. That one then, with respect for one's original idea or venture, somehow allows oneself to be led/ guided along the way whilst at the same time using the new information to forge some kind of successful result. That one pays attention to the solution that is emerging, and not kill it off or put it into a box by some very structured, preordained or wished-for result.
Could one say - planning for a surprise result? Sounds a bit of a paradox I know. This might sound a bit woolly, but I think it reflects my intuitive sense towards this issue. But that Zen thing springs to mind again. Maybe one has to bring these things up to the surface, examine them, maybe learn them, but then *forget* it all, to finally be able to use what was once a heavy concept, in a light and capricious way.
Ben Rollason wrote:
The idea of spreading risk has been well argued and is no doubt a good strategy. A gamble that isn't a gamble. Although once it's brought down to the level of hedging your bets, it doesn't seem quite so appealing. One of the issues I seem to remember having been raised at the beginning of the book was how larger corporations can emulate the success of start up companies, such as Yahoo. Now I don't know a great deal about the history of Yahoo, but I would guess that they didn't achieve their success by hedging their bets.
What I am trying to say is that there are many sites that will have been started up as a labour of love end up capturing the imaginations of the public at large in the way that it captured the imagination of it's inventors in the first place. Tenacity, vision and single mindedness can suffice where bet hedging may fail (by resulting in a partial success).
As I understood it, at the beginning of the book, the corporates were not totally failing to break into e-commerce, but were simply not flying in the same way that the start ups were. Perhaps, paradoxically it is being prepared to fail totally for something you believe in that distinguishes the small upstart success stories from the corporates, for whom simple, quantifiable, financial success is an end in itself. Undoubtedly many, many start ups did fall flat on their faces. A total failure! Brilliant...next attempt.
The risk spreading argument also assumes unlimited resources, human and financial. This isn't the case always. I also perceive the risk spreading strategy as slightly heavy and certain, whereas the start up mentality is light, capricious and flippant. Perhaps corporations just can't emulate this.
Perry Barile wrote:
Half baked success is hard to deal with. Terminate or continue? I think before measuring how successful a success is, you have to predetermine exactly what will constitute success. At what point will this venture be a success? Knowing these things can be difficult. I think experience would play the biggest part.
Ben Rollason wrote:
My comments on bet hedging were not really a direct response to your posting. I have nothing against the idea of spreading risk at all and I agree the two terms are pretty much the same. It's probably the more prosaic emphasis on the term 'hedging your bets' that prompted my meanderings. Like many things, a small shift in perceptual emphasis leads you along a quite different (and not wrong or worthless) tack. I had forgotten that you had mentioned partial failure, so my comments on partial success were unrelated, but in retrospect, parallel, I think.
Don't get me wrong; I wouldn't advocate planning for partial success. I was trying to say that to me a partial success is more a failure than an all-out , proper, face-down-in-the-mud failure, and hence, something to be avoided. But meanwhile that partial success is an easy situation to slip into and a hard one to get out of. Whether to try and capitalise on the partial success or to fail properly and start again, I suppose is the core dilemma.
I often found that the moment of ditching a partial success, the moment of destruction, often turned out to be the most seminal and creative moment. I recall learning to model portrait heads at college and an elderly teacher by the name of Monty who was famed for keeping an eye on students progressing down the wrong track modelling a feature, a nose for instance, in great detail, having got the fundamentals of proportion and volume quite wrong. He would wait a bit and then take great delight in removing the nose in it's entirety....Eventually, some of the students would start to pre-empt Monty by pulling the features off their own efforts if things were progressing in an unfavourable direction. A good teacher!
As Perry commented, success is a very hard thing to quantify and to do so, you will probably have to set some measure. I think that success goals would be likely to shift though. In this sense that one's perception of one's own success or failure is likely to move and evolve with whatever unfolding venture one is trying to quantify. I must point out as well, that I am not talking specifically about an entrepreneurial experience, never having set up a business. I am talking really about a personal and creative processes, but I think there are probably strong resemblances. If I was to set up some kind of business, it would be this experience that I would expect to draw on most strongly.
In rereading your [Chris Heape's] original response to the last chapter, I came across: "This inevitably leads to the description of 'Killing Your Darlings'". If I understand your meaning correctly, I have always found 'killing your darlings' a laudable aim. The ability to destroy the things that you have invested most time and love in. It's hard, but the best cure for the onset of creative inertia.
Again though, I have always found this hardest when confronted by a partial success. Accident and failure have to be embraced in order to succeed. You're right, embracing partial success is not an option. I think perhaps there are moments when one has to spread risk and times when one has to rush in where angels fear to tread. The glue perhaps is the zen thing. The intuition that tells you when and how to act, when to stick with something and when to 'kill your darlings' and shake yourself right up. At the end of the day, you can own the world's biggest rule book, but without the intuition, it is useless.
Chris Heape wrote:
Whilst making the supper I jotted down these thoughts: This is it! / bouncing back and forth / in and out of intuition / what turns up / chance and logic. Progressing with some kind of rational conclusion is now a little difficult. There is an interplay here. How does one know when to "kill off a darling"? How does one know when a partial success has to be decapitated? When does one use something from a good idea / venture and then throw the rest out with the bath water?
I feel that this has to do with a back and forth between logical rational thinking and maybe more inspired intuitive / creative thinking. On second thoughts logical and analytical thinking is of course just as creative! One engages both these areas, ( intuitive thinking and logical thinking) in a creatively strategic plan? Maybe some kind of: creatively-planned-allowing-oneself-to-try-out-mistakes in order to strengthen-one's-initial-goal plan. I find it difficult to very quickly apply deeply rational arguments for these thoughts. That I feel is Peter's job!!!
I feel OK with an intuitive understanding of this way of relating to a development process. I think the most important thing about this way of constructively using risk and intuition is that one has beforehand allowed / planned-for several approaches. One has defined beforehand some kind of goal. This allows one's subconscious or world of association to link in / latch on. Allows a series of gates to be opened. One has accepted, that within this almost intangible, defined space / railway track, that one will accept whatever turns up, relative to the goal that one has *temporarily* set. One has accepted that there will be a series of choices, from a range of alternatives that one has set in motion, that one will have to make in order to weed out the weak parts.
Ben, getting to your point of: " Whether to try and capitalise on the partial success or to fail properly and start again, I suppose is the core dilemma...." You're not starting again. This the whole point. You're building, expanding on a sense of movement that you've set in motion. This I think, unless I'm completely off the mark, is what the aspect of *engaging the will to apply risk in a creative and growth e-commerce venture* is all about. One learns so much by what has gone wrong, especially if this has been planned for! This is fundamentally how any creative process develops.
This discussion elegantly illustrate all the problems that occur with top down structural solutions, even when object oriented thinking is super imposed over them. Like the Christmas tree lights, the faults are not easy to isolate. The final arrangement, the solution, may only partially work. It may perform, but, only moderately well. It may be filled with modules that are "special darlings" that are emotionally difficult to remove. A brilliant feature like Monty's nose might be blinding the designer to a fundamental fatal flaw in the main structure.
This is the inherent problem with top down design strategies. They can work fine if you are working in stable areas of business and commerce where there are few unknowns and relatively little uncertainty. They work fine in areas where there are set protocols, standards, rules and regimentation. These are the conditions of the Industrial Age.
In the age of information, where everything is changing and uncertain, the top down approach is found to be severely wanting. Here, solutions require the other flavor of object oriented design: the bottom up approach. This will be the subject of the next chapter