The McKinsey Mind (4/5)
I found this book to be particularly useful for framing and putting a structure to the thinking process that we go through when we try to solve problems or seek answers to complex problems. While the McKinsey-ites mainly use these techniques to solve business problem on which they usually know very little about, I find this thought process to be very useful when forming an investment thesis on a new company or industry. I already use a similar framework when I analyse a new investment candidate but this book has definitely helped me in refining this process and I hope that readers of this blog will find that summary useful in their business/professional life. Let’s get started…
For the purpose of this post, let’s assume that we are a McKinsey task force that was hired by Charter Communication and we need to answer the following burning question: “are cord-cutting and OTT a real threat to cable companies?” I will not answer this question in this post but I’ll use it to illustrate the thinking process that one has to go through in order to answer that.
Step 1: Framing the problem
Before we tackle a problem we need to understand what the problem is and frame it. The first step is to put a structure to the problem. One element to which I was exposed during business school is the MECE thinking (pronounced “mee-see”, and an acronym for Mutually Exclusive, Collectively Exhaustive). This means separating the problem into distinct, non overlapping issues while making sure that no issues relevant to the problem have been overlooked. Another important aspect of the framing part is structure: without structure your ideas won’t stand up. One way to structure a problem is by drawing a logic tree that maps out the different components and factors that impact the business or the segment in question, the logic tree, of course, has to be MECE.
Let’s draw a MECE logic tree to map the situation that our client is facing:
Cable — Broadband
Video ——— revenues ——— traditional sources
This could be expanded into many other levels (we can do the same for the competitors) but for the purpose of this post we will leave it at that.
We will drill a bit deeper in just a bit.
Mckinsey-ites learn that it is much more efficient to analyse the facts of a problem with the intent of proving or disproving a hypothesis than to “boil the ocean.” For that we need to form a hypothesis. To do that, one has to be properly prepared to brainstorm after having absorbed all the facts that are relevant to the case. For this process to be efficient, it is better to check your preconceptions at the door and think in all directions. Finally, we have to remember that the problem is not always the problem! Don’t take any assumptions for facts and challenge every assumption, dig deeper, ask questions and get the facts. Let’s say that in our case we come up with the hypothesis that OTT is a real threat to the business. Next we need to build an issue tree:
Issue: Can OTT reduce our profitability?
Subissue: Will it decrease our revenue?
Subissue: Will it increase our costs?
Subissue: Can we do anything about it?
We can further develop the subissues, for instance for the last one we can go one level deeper:
Subissue2: Does it require special skills/resources?
Step 2: Designing the analysis
Once we have our problem duly framed we need to come up with a battle plan for how to analyze the issues at hand. Few points to remember:
- Find the key drivers: there many items and variables that impact every issue, but some are more important than others. We need to figure out which factors are the most important and focus on these.
- Look at the big picture: when trying to solve a complex problem it is very easy to lose sight of your goal. We have to be focused on the big picture and make sure that everything that we do moves us towards the goal.
- Focus: I had a boss in the past that had a very interesting name to losing focus in the ocean of data that is always in front of us—he called it “analysis paralysis.” We have to get our priorities straight and let our hypothesis determine our analysis by avoiding analysing things that are not related to our hypothesis.
- Forget about absolute precision: while we have to stick to a fact-based analysis, solving business problems does not require the same precision level of solving problems in math and physics. In fact, in most situations, achieving a scientific level of exactitude is counterproductive. You will spend an inordinate amount of time and effort getting from mostly right to precisely wrong. In most business problem, the right way to go is to find a comfort zone—that is a range in which we are comfortable with our decision. That range can be very wide as in just deciding if it’s $5m, $50m or $500m and it won’t matter if it’s $50m or $75m even if $75m is 50% off from $50m.
Bearing all this in mind, the next step is designing a work plan for the analysis of the issue at hand. The table shall contain the Issue, Hypothesis, Suggested Analysis and Data Sources that we will use in order to conduct the analysis that will prove or disprove our hypothesis.
Issue: Can OTT hurt our profitability?
Suggested analysis: look at competitive product offering and try to synthesize the desired offering via substitute products
Data sources: competitors price plans, cost of a stand alone internet package, surveys that show what is the typical desired offering for a family etc.
The profitability issue can be divided into two subissues:
- Will it increase our costs?
- Will it decrease our revenues?
We can design a work plan for each of the subissues until we are certain that we covered the whole land.
Step 3: gathering the data
The data we use to support/disprove our hypothesis have to be fact-based. We need to read a lot of material (annual reports, industry reports, publications) and another potent tool is interviews. Eventually, we want to get to a point where we have sufficient data to support the arguments that we will eventually make. Arguments without supporting data have no value. As for interviews, here are a few tips:
- Be prepared: you must come with a list of questions that topics that you want to discuss. Put your thoughts on paper as it helps you to organize your ideas and make sure you won’t miss anything, finally, end interviews with “Is there anything else I forgot to ask?” it pays big time once in every while.
- Pre/post interview: feel free to send your interview guide to the interviewee, most people don’t like surprises and by giving the other side time to prepare you may get a lot in return. Finally, write a thank you note after the interview to thank the interviewee for his input and time and maintain the relationship.
For our OTT case we can gather data by studying the cable TV packages, OTT packages (Netflix, Amazon Prime etc.). But let’s not forget that an internet package would have to come along with our OTT package. We can dive a bit deeper and look for content cost for cable TV operators and compare those with content cost of OTT operators to understand if we are comparing apples to apples. We can look at the contents of each package. “Movie Package” may refer to Hollywood’s latest hits but can also refer to movies from ten years ago.
Step 4: interpreting the results
So now that we have piles of data and many Excel spreadsheets it is time to make sense of that data. Two important things that are important to bear in mind at this stage are:
- The 80/20 principle: 80% of an effect under study will be generated by 20% of the examples analyzed.
- Don’t make the facts fit your hypothesis, if the data shows that your hypothesis is wrong then go back steps 1 and 2 and start fresh.
It is very important not to get lost in the data, my ex-boss put it best and he always told me that I should avoid “analysis paralysis.” In McKinsey, they always say “What’s the so what?” for each analysis or data that is being processed. If you can’t answer the “so what?” question, it probably shouldn’t be done, doesn’t matter how fancy the analysis is. Next, sanity checks are very important and finally one must remember that there are limits to any analysis so intuition and common sense must take part too. Or, in the author’s own words: “Data without intuition are merely raw information, and intuition without data is just guesswork. Put the two together, however, and you have the basis for sound decision-making.”
For our cable case one might ask how does the 80/20 principle applies? is it that 80% of the content cost comes from 20% of the content? If so, how much would separate content packages for that premium content (think NFL) cost? You probably get the drift.
After completing our top-notch analysis and combining facts with intuition we need to arrive at the end product. In doing that, we must make sure that the solution fits the client and that we see both the problem and the suggested solution through the client’s eyes. This is done by understanding the client’s constraints and abilities, which sometimes have to do with politics, power and other things that don’t show up in the numbers. While this might appear to have little to do with investing I see a clear parallel when some activists armed with spreadsheets show how can a certain company create a lot of value by following recommendations that look great in the activist’s Excel spreadsheet but are impossible in reality. Some of the recent spinoffs of companies that merged only a few years ago in the face of pressure from activist investors—in some cases these are the same investors who pushed for the merger—are an excellent evidence for great analysis that is too grounded on numbers and not grounded enough on non-quantifiable factors and drivers.
Step 5: Presenting the ideas
“It’s amazing how many successful people cannot simply focus on two or three key points and articulate them well”
This part is all about giving structure to our ideas and getting buy-in from the client. I also learned this during my MBA: structuring your communication in a certain way helps to create a maximum effect and a logical flow of ideas can be easily absorbed and effectively get your message across. This has to do something with the way our brains work but I don’t remember the whole mechanism of how our brains digests data and ideas that are presented to it.
This chapter opens by going into what is known as the Pyramid Principal (the author didn’t use this name) of presentations. It is basically an inductive reasoning structure: “We believe X of reasons A, B, and C” instead of deductive reasoning that most people use in presentations, which is: “A is true, B is true, C is true; therefore, we believe X.” The deductive reasoning is boring to follow and you might lose your crowd if A,B and C contain a lot of data. The listeners might be wondering all the time “where is he going with this?” so it’s good to first tell the audience down which path you are going to take them so they see the rest of the presentation in the correct context.
For our OTT example the first slide may contain the following:
OTT will isn’t a threat that cable TV operators should be worried about:
- It doesn’t save money
- It is more complicated to handle
- Even if pay-TV customers leave cable, cable companies’ P&L won’t change much and may actually improve because they won’t have to buy the expensive content
Then other slide may look like this:
It doesn’t save money
- If we build a bundle that provides similar content to the pay-TV package it will be more expensive
- If we go for OTT, we need to purchase a broadband package not as a part of the triple-play
The book then goes into various presentation tactics which I won’t get into here.
The next steps relate to managing the client, the team and yourself and have less to do with investments so I won’t get into it. Overall, I would recommend this book for anyone who has to analyze things for living and come up with actionable ideas. The book was a good read and I hope that you can refer to this post next time that you have to analyze something and improve the structure and quality of your analysis and ideas.