Making decisions often means selecting the one project that will fetch you the most advantage. Advantage may be monetary gains or cost savings or even bigger user base for your product. Most often it happens to be monetary gains.
How many times it happens so that the problem statement is complex with inter-related if-else relations and you find it hard to make a choice in a rational way? Here are some tricks you can leverage to make the process easier and help yourself.
There are multiple fundamental tools, which are extremely useful in fairly regular decision-making. One of them is decision tree. This is an effective tool because of the focus it has on chances of occurrence of every possible what-if scenario.
The Café problem
Let us say you are the owner “Le Grand Café” coffee shop. You were a pioneer in your town and you have worked hard in past three years to make a successful business out of your pet project.
.jpg)
While you have enjoyed the first comer advantage to see positive cash flows in the past two years, you now are seeing a slowdown. As per your analysis, it is mainly because of
- Direct competition from a couple of new entrants who are inspired by you
- There has been an increase demand for fruit juices and ice creams lately
- You have not introduced any new flavor lately
After weeks of thinking you have come up with a list of possible actions to rejuvenate your business.
First, you can invest in trying out and developing a new flavor. It will take significant effort for your talented barista to work on the R&D job while still doing the regular stuff. He is okay to do it, but demands 8,000 to do it. Instead, you can also consult a coffee specialist who also consults for Starbucks. He charges a hefty 28,000. If your barista creates a likeable flavor, it will result in an estimated profit raise of 110,000. If he fails to impress, you will make only 7,000. Coffee specialist, if he makes a likeable flavor, you can earn 120,000. But you are afraid, as he is not really aware of the taste of your town, a failure will set you back by 32,000 because of weakening of brand. In your opinion, there is 40% chance that your barista will create likeable flavor. On the other hand, the consultant is highly skilled and will create a likeable flavor with a 70% probability.
Second option could be to diversify your business into ice cream domain. This requires a new chef, as you do not currently have the skills in house. It will cost 70,000. If you were to market this venture successfully, estimated benefit is around 160,000. But, if you fail, you will make only 60,000. As the market is competitive, there is only a 60% chance of you succeeding in this venture.
Third option is to expand to newer market with your existing expertise. You know a nearby town where you can open you second shop of “Le Grand Café” chains. This will take an investment of 150,000. If it succeeds you expect to make 220,000. If it does not click, you will only make 180,000. Knowing the culture of that town, you think the chance of success is about 55%.
You could also hold on and not do anything. This will give you a benefit of 40,000.
What would you do in this case?
As it should be evident now, the complexity escalated quickly with just 4 simple alternatives at hand. Decision tree will help you see things clearly and get an objective perspective of the situation.
Decision Tree
Decision tree is a fairly simple tool once you understand the fundamental elements of it. There are following graphical elements.
First is a square, which represents a set of decision alternatives. For example the square number 3 shows you have four choices to have go at in our café case.
Second is a circle, which represents chances of certain events occurring. Each event must be associated with a probability.
Finally, the lines are branches, which represent either one of your choices or a possible event depending on if you are at a decision node or a chance node.
Here is how our problem looks like once modelled.
Decision tree works with the concept of Expected Monetary Value (EMV).
EMV of each chance node is calculated by taking the probability weighted arithmetic sum of monetary values of each branch. In statistical terms, this is simply expected value or average of all values.
EMV of each decision node is simply greatest monetary value amongst all the branches. Remember we discussed about “monetary gains”? You are simply trying to earn maximum money by making a choice.
So in our café case, you would be better off by opening a new café in the nearby town rather than going for R&D or diversification as of now.
Food for Thought
While all this looks very objective and effective, there some fundamental challenges with this tool.
- How do you get the probabilities?
It is very tricky to determine the probabilities events. One way to do it is to have a look at historical events. But often that can be very misleading. Another option is to simply make a guess, like we did in the above example. These may also be qualitative decisions one has to make based on experience and historical data put together.
Sometimes, it does not make any sense to make a guess for the probability. In such cases there you may use strategies like Maximin, Maximax or Minimax Regret. These can help you see different logical options you have in a given situation.
- How do you accurately estimate the benefits of each choice?
This is another difficult job. In our example, we simply assumed some monetary values. In reality one has to estimate the cash flows for the full lifetime of the alternatives. Again, this needs extensive forecasting.
- Does this model hold even if you are evaluating a very long-term decision?
If you are making a strategic decision like in our example, it will have to stay for many coming years. So, the time value of money must also be considered based on the forecasted cash flows to make the decision accurate.
Thought Farming 