All 4 entries tagged Decision
View all 64 entries tagged Decision on Warwick Blogs | View entries tagged Decision at Technorati | There are no images tagged Decision on this blog
March 18, 2011
After spending a lot of time looking at a lot of different tools lately, I have decided that DT’s are probably one of the best (obviously, depending on the situation and the information that you have. Their ability to get to the heart of a problem through finances (or utility) is quite amazing, and the structuring of the process is brilliant in that it helps you to ensure that you have covered every eventuality. It is especially good at helping you to avoid the confirmation trap. For example, any time you make a branch, you also have to question if there is an opposite alternative at the very least, if not more completely distinct ones. They are also hard to argue with, and help you make informed decisions based on the likelihood of certain scenarios playing out. The power of decision trees is something that I am unlikely to forget, and they are probably something I will integrate into my general System-2 decision-making methodology.
In terms of helping our team to reach a final decision, about whether to continue or not, and where to locate, they were absolutely invaluable. Of course, the tree doesn’t know anything that you don’t tell it, so qualitative factors are very hard to incorporate, and something like Grid Analysis or Analytical Hierarchy Process are much more appropriate. Similarly, while we did include marketing costs, or factory re-sale gains, other costs that are difficult to estimate, such as the potential training costs of workers for the Exmouth factory, were not included.
So it can be argues that DT’s do not always give you the full picture, which is true. But, you don’t have to take the outcome as fact. Even after evaluating the tree, you need a high degree of judgement to decide what is important or not, and how risky certain paths really are. The chance of the worst case-scenario playing out for Waveriders (so, product doesn’t get developed until Dec 2012, and when production begins in Jan 2014, the market conditions are poor) is around 0.15. That is the only situation under which producing at both factories might be a problem, so it is worth balancing the risk and pursuing that course of action anyway.
However, DT’s are clearly an excellent tool for decision making, and when used correctly, are capable of judging a situation in a way that our own cognitive processes are rarely capable of. Of course, there is still potential for bias, and they are no substitute for experience and good judgement. They cannot be relied upon to actually make the decision, as they will certainly NOT be taking the blame for bad decisions!
March 12, 2011
Not that we need to make this task harder, but while reading earlier, I thought of a complication that would often be present in group decision making in the real world. In our groups, no-one really has their own agenda; we are all working together, collaborating for the group in order to do as well as we can. There are no competing agendas or ulterior motives. Additionally, we are all equals - no one person has any more say than another in theory (that might be different in practice!).
However, within a business for example, many groups might be composed of multi-functional teams, or with management of varying levels of positional power present. Each of these might have different areas of concern, i.e. the finance manager's role might be to cut costs, the engineer might be attempting to maximise quality, the marketing manager might wish to preserve the size of the budget available, etc. So how do the competing agendas of these people affect their decision making, and consequently, the ability of the group to make decisions? There is bound to be some bias in the proceedings. We know that Deming would advise that the best way forward would be to break down the barriers between these people, instill constancy of purpose, and get everyone thinking about the organisations goal as a system, rather than their own. But of course, in practice, this is difficult.
So, I wonder, what would this task have been like if we were all to play a role? Pretty difficult I imagine, to the point that it might even defeat the purpose of trying to use all the tools and work together. However, maybe there is cause to have a seminar on this, or some (LE style) role-playing exercise to explore the challenges of a situation like this?
The last few days have been spent clarifying the problem and then trying to understand the tools best suited to solving it. We used a little bit of a methodological approach even in deciding which tools to try (!), and then each person volunteered to try to implement the ones that interested them. Upon meeting again, we attempted to present the results of our findings to each other, most of us thinking that our work had got us to the point where the group might be able to finalise decisions.
How wrong we were!!! Our internal biases had once again led us astray, and into thinking that our work could be without fault, and that each of the others would automatically understand what we had worked on, making the assumption that they had the same tacit knowledge that we did! It's becoming clear to me that managing all kinds of bias is certainly the biggest and most important challenge when making any judgement or decision.
What actually happened was that we found our work either littered with mistakes we hadn't previously seen, or in some cases, the rest of the group held wildly different views when presented with the work we had each done. This meant that much of what we had done turned out to be work in progress rather than the finished article, as so much rework was required! For example, I had largely focused on the decision trees, both for deciding what our options were and working through the possible results of around 8 different resultant scenarios. But, I'd failed to take into account that we might need another tree entirely for choosing the best location, or that sunk costs shouldn't have been included in the expected value calculations...
Once again, that means that after much confusing of our teammates, who might not have known the tools we worked with as well as we did, and realising that even we might not have used them correctly and entirely, or done requisite research, we find ourselves unable to make a decision still! Monday will be an important day for this.
Reflecting back, and thinking about future applications, I think that this will always be a problem. We will often have to present to colleagues entirely unfamiliar with our methods, so it will be crucial to learn how to translate our resuls in a transparent way, making sure to avoid the curse of knowledge. Of course there will be times when we fail to see errors in our work, so it is also very helpful to have colleagues who also know the methods and can check our work. This is quite a common approach in most engineering situations already, but perhaps as managers or leaders, we are not quite so used to this.
March 10, 2011
This module has provided a lot of food for thought so far. I especially always enjoy seeing the links between whatever we are working currently, and what we have seen before. For example, we spoke a lot in Leadership about the ability of leaders to make decisions, and how they should go about doing so, and now we are being given the methods for how. Similarly, we spoke about the difficulties using lessons learned effectively in our PEUSS PMA's, and that the main problem was that tacit knowledge was very difficult to consciously access. Now we see the effects that it can have when the tacit knowledge we hold is based on unconscious biases (related to the concept of mental models in Learning Organisations from CBE!), and these lead to decisions that might seem good at the time, but ultimately result in unforeseen problems.
I am very thankful that we had the chance to test our own judgement. I think that mine came out relatively well, in that I was right a little more than half the time (but don't tell my girlfriend that it was that low! :P), but it was relatively clear to me, and probably most of us, that we're not right about things nearly half as much as we think we are. System 1 lets us down, and that is where the need for System 2-type methods present themselves, of which we are currently discovering so much. However, I have a concern about all this - while it's good that I'm currently questioning a lot more of the decisions I make, is it counter-productive to question yourself constantly. If you can never rely on your System 1, that arguably makes your ability to respond to new developments in general much worse!
Still, I fully accept the need that it is better to question judgement, and understand our internal motivations, than blindly assume correctness, but it also makes it harder to function...