Below is an excerpt from the basic edition of our Succeeding as a Management Consultant book which explains how to use MECE and 80/20 principles to structure and analyze a real consulting project.
Engagement teams usually consist of experienced and inexperienced members. Consultants may be inexperienced because they do not understand a sector, client or particular type of analyses.
This initial week in the office provides inexperienced consultants a chance to understand the client and sector, and not to appear “green” in front of the client.
This is the general format of the week:
Reading and Research
Planning commences with the engagement team reviewing all relevant newspaper articles, research reports, equity research reports, annual reports and regulatory filings about the client.
They will also read competitor information and obtain advice from other consultants who have worked on similar engagements. The objective is to gain a broad understanding of the issues without going into too much detail.
Throughout the entire engagement, the team will follow the 80/20 principle. They will only focus on 20% of opportunities which will generate 80% of the value created.
As the team conducts its research, each member will write down issues they think affect the client. The team is not worried about the accuracy of the lists of issues at this point. The aim is to generate a list of issues based on educated guesses and a careful reading of the material available. The team will try to go broad and identify as many issues and across as many areas of the business as possible.
Over a series of one or two, 2-hour meetings the team will use post-its to list every issue on a white board. The Goldy team generates over 140 post-it notes. They can be in any order, any priority etc. The aim is to capture all possible issues affecting the client. It is similar to an educated brainstorming session. Luther facilitates the sessions to generate discussion and to capture the issues. Usually the team will work together in one large room during this planning week. This allows them to create a war room, put up information and immerse themselves in the engagement.
Putting issues up on the white board helps everyone see a common set of items, potential patterns and themes developing. In the second or a third meeting, the team starts discussing the common themes from the list of issues. The session focuses on clustering the issues into common themes. For example, the following cluster of 13 issues scattered around the whiteboard can be listed under a theme called “rising costs” or “costs”:
• Four salary increases to reduce labour walk-outs.
• Fuel prices doubled in 16 months.
• Shortage of equipment and spare parts driving up costs.
• Mining taxes have increased.
• Water levy introduced 7 months ago.
• Extended use of contractors.
• Tunnelling machines are in high demand and difficult to secure.
• Unrest in Thailand is forcing Goldy to buy electrical reticulation equipment from expensive Malaysian manufacturers.
• Brazilian Real is strengthening against the US Dollar, Australian Dollar, Russian Ruble and South African Rand.
• New mining act is creating uncertainty about future environmental costs.
• Steel and copper prices have increased 17% in two months.
• New shared services centre only rolled out at 20% of the mining sites.
• Rising inflation driving up domestic costs.
The team will go through every issue and add it to a theme. Once this is done, some themes may be combined while others may be split apart. Generally there are rarely more than 8 themes at any one client and for any one problem to be solved. Sometimes consultants take each theme and develop a hypothesis to test the theme.
The challenge with this approach is developing an analyses framework on which to lay the hypotheses. Without the framework it is difficult to apply the rules and techniques management consultants use to structure their analyses. To overcome this critical problem, another approach is presented below which allows for the development of a hypotheses framework.
The group uses this discussion on themes and key issues to spur debate and gain a better understanding of the engagement. This robust debate forces consultants to ask important, probing and tough questions. It is as much planning for the study as an educational session for the consultants. It serves as a filter to weed out poorly formed ideas or weak thinking. After they are comfortable with the themes, the team puts aside the themes and key issues for a moment.
The group then takes the key question presented by the client and tests if this is indeed the key question they need to answer in the engagement. Sometimes clients raise the wrong question, which the client thinks must be answered. The engagement team tests if this is the right question.
Many consultants get too focused on answering the question posed by the client. However, perfectly answering the wrong question will not help the client. The right answer to the wrong question will still not solve the root-cause problem. Therefore the team takes time to ensure they are asking the correct questions in the engagement.
They ask themselves, “If we solved this question, would the problems at the client be resolved?”
If the answer is yes, the key engagement question is captured. In this case, the engagement team believes this is indeed the correct question to answer and puts it up on the whiteboard.
Key question: How can Goldy improve its production value?
Assuming they have the correct question, they need to think about how they would go about answering this question. Answering this one question is a difficult task. To make it easier and manageable, the team takes this question and splits it into smaller questions in a logical format. The next, level two, set of questions would be:
Level 2 Question: Can Goldy increase its revenue?
Level 2 Question: Can Gold reduce its costs?
Either way is a means to raising Goldy’s production value.
As the team develops each layer of questions, they test each layer (layer 2, layer 3, layer 4 etc) by asking themselves two further questions.
First, are these the complete list of questions in this layer which can impact the previous question? Therefore, is raising Goldy’s revenues and reducing Goldy’s costs the only way to increase production value? Is there any other way which should be added as a level 2 question? If the answer is no, this is called a collectively exhaustive list of questions or hypotheses (“CE” part of MECE).
Second, have the questions been sufficiently separated, so that changing the variables which impact one question will have NO impact on another question? For example, the price of gold is a variable which impacts revenue. However, the price of gold will rarely have an impact on the costs. If each of the questions does not overlap through their variables, then the team will say that the questions are mutually exclusive (“ME” part of MECE).
The first concept (“CE” in MECE) ensures that no stone is left unturned in analysing the key question of the engagement. For example, imagine there was another way to increase Goldy’s production value but the team overlooked this option. When the final recommendations are presented, it is possible that the overlooked option may have altered the recommendation. Having a collectively exhaustive set of options ensures that all avenues are explored.
When issues overlap and cannot be isolated, it is difficult to know why changes are occurring. It is also difficult to understand the issue. Isolating a question, issue or analysis (ensuring that “ME” part of MECE is taken care of) allows the engagement team to conduct a test whereby they can be sure that x, y or z is responsible for the changes.
If the hypotheses/questions fail the test of being mutually exclusive, then the analyses and findings will be flawed. That is because they are running analyses on hypotheses that have not been isolated for testing.
The team takes the time needed to ensure the analyses fulfill MECE requirements. This is not easy to do. It can take an engagement team up to a week to ensure MECE rules are met for each part of the analyses. The team applies MECE as they continue breaking down the key question.
A level two question/hypotheses can be broken down even further. Let us look at both revenue and costs:
Level 2 Question: Can Goldy increase its revenue?
Level 3 Question: Is there a way to increase the price of gold?
Level 3 Question: Is there a way to increase the volume of gold sold?
Level 3 Question: Is there any other way in which revenue can be increased? (Revenue from other sources such as investment income is included here. Since this branch makes such a small contribution to overall revenue in mining companies, it is removed from the overall decision tree)
Level 2 Question: Can Gold reduce its costs?
Level 3 Question: Is there a way to reduce operating costs?
Level 3 Question: Is there a way to reduce capital costs?
Again the team will check if they are mutually exclusive and collectively exhaustive (MECE). The team will continue building a level four and level five set of questions for each question. In the planning phase some teams build out to a level eight set of questions. Rarely will greater detail be needed at this stage. Later in the engagement more detailed will be added if required for the analyses. When they are laid out from left to right, with the primary question on the left (How Goldy can increase its production value?) and the subsequent levels fanning out to the right, it tends to look like a tree with branches. This is the origination of the term value tree or decision tree.
MECE and 80/20 rule is the foundation on which consulting analyses are built.
Decision Trees, which must be MECE
The team now goes back to the list of issues and themes it developed. Can they find a place on the tree where every issue and theme can be tested? If not, is the tree missing some questions? Are some of the issues irrelevant? Must the questions change?
To accomplish this task, the team takes each issue and numbers it. A stack of post-it notes are used with each post-it note numbered to correspond to an issue. The post-it note is then placed on the decision tree where the team thinks the issue resides.
Using this process, the team can visually see where most of the issues may lie. This can also serve as a check. First, is the analyses skewed to one or two parts of the question. Second, have they been unable to find issues in certain areas. If so, does this mean these areas have no issues?
The team may need to go back and relook at their issues to understand why an area was ignored. Once they are comfortable with their decision tree, they will proceed with the planning.
Mapping Issues on Decision Trees
Developing the decision tree is one of the most important steps of an engagement. Using the decision tree, the team can break down a hypothesis into manageable components for analyses. The team will revisit this in the next few days and test it further until it is finalised as a strawman. Engagements are not static and as information changes, the priorities of answering the questions may change. This may cause the tree to change slightly.
Given the amount of work required to rigorously analyse each branch in the tree, the tree will be split into different branches (set of questions) and different consultants will take over ownership to develop them further. It is critical to document and share all the work done on the decision tree and issues. They will be needed by the engagement team to review their thinking.
Using these detailed decision trees and hypotheses, the engagement team can determine the likely answer to the questions before they arrive at the client site. That is, the team looks at each question in the decision tree, and based on their careful preparation they estimate the likely answer. Using the decision tree, the team also develops the storyboard for the engagement. The storyboard is the message delivered to the client based on the expected results. This is the counterintuitive part.
If this is done so early, then what’s done during the engagement?
The engagement is therefore the process of proving or disproving the hypotheses. The decision tree and hypotheses are written as questions. Over the course of the engagement, the team will develop analyses to test each hypothesis, and thereafter collect the data for the analyses. Depending on the results of the tests, the analyses is either proved or disproved and the storyboard is altered.
The process can be summarised as follows:
1. Determine the key engagement question
2. Develop the decision tree
3. Check for MECE
4. Develop the storyboard
5. Develop analyses to test each question
6. Collect data for the analyses
7. Complete the analyses
8. Refine the storyboard
The process above is iterative. As analyses are completed and new information becomes available, the team may need to go back and create new analyses and repeat the process. It is not a linear process. This is the central technique used by management consultants. It allows the team to see the overall message and focus their work before the engagement begins.
 A strawman is a draft version of a document.
 The client site refers to the client owned premises within which the engagement team will be located over the duration of the study.
Image from Trey Ratcliff under cc.