McKinsey and Company is recognized for its rigorous approach to problem solving. They train their consultants on their seven-step process that anyone can learn.
This resource guides you through that process, largely informed by the McKinsey Staff Paper 66. It also includes a PowerPoint Toolkit with slide templates of each step of the process that you can download and customize for your own use.
You can click any section to go directly there:
- Overview of the McKinsey Approach to Problem Solving
- Problem Solving Process
- Problem Definition & Problem Statement Worksheet
- Stakeholder Analysis Worksheet
- Hypothesis Trees
- Issue Trees
- Analyses and Workplan
- Synthesize Findings
- Craft Recommendations
- Distinctiveness Practices
- Harness the Power of Collaboration
- Sources and additional reading
Download the Umbrex Toolkit on the McKinsey Approach to Problem Solving
Overview of the McKinsey Approach to Problem Solving
Problem solving — finding the optimal solution to a given business opportunity or challenge — is the very heart of how consultants create client impact, and considered the most important skill for success at McKinsey.
The characteristic “McKinsey method” of problem solving is a structured, inductive approach that can be used to solve any problem. Using this standardized process saves us from reinventing the problem-solving wheel, and allows for greater focus on distinctiveness in the solution. Every new McKinsey associate must learn this method on his or her first day with the firm.
There are four fundamental disciplines of the McKinsey method:
1. Problem definition
A thorough understanding and crisp definition of the problem.
2. The problem-solving process
Structuring the problem, prioritizing the issues, planning analyses, conducting analyses, synthesizing findings, and developing recommendations.
3. Distinctiveness practices
Constructing alternative perspectives; identifying relationships; distilling the essence of an issue, analysis, or recommendation; and staying ahead of others in the problem-solving process.
Actively seeking out client, customer, and supplier perspectives, as well as internal and external expert insight and knowledge.
Problem Solving Process
Once the problem has been defined, the problem-solving process proceeds with a series of steps:
- Structure the problem
- Prioritize the issues
- Plan analyses
- Conduct analyses
- Synthesize findings
- Develop recommendations
Not all problems require strict adherence to the process. Some steps may be truncated, such as when specific knowledge or analogies from other industries make it possible to construct hypotheses and associated workplans earlier than their formal place in the process. Nonetheless, it remains important to be capable of executing every step in the basic process.
When confronted with a new and complex problem, this process establishes a path to defining and disaggregating the problem in a way that will allow the team to move to a solution. The process also ensures nothing is missed and concentrates efforts on the highest-impact areas. Adhering to the process gives the client clear steps to follow, building confidence, credibility, and long-term capability.
Problem Definition & Problem Statement Worksheet
The most important step in your entire project is to first carefully define the problem. The problem definition will serve the guide all of the team’s work, so it is critical to ensure that all key stakeholders agree that it is the right problem to be solving.
Problem Statement Worksheet
This is a helpful tool to use to clearly define the problem. There are often dozens of issues that a team could focus on, and it is often not obvious how to define the problem. In any real-life situation, there are many possible problem statements. Your choice of problem statement will serve to constrain the range of possible solutions.
- Use a question. The problem statement should be phrased as a question, such that the answer will be the solution. Make the question SMART: specific, measurable, action-oriented, relevant, and time-bound. Example: “How can XYZ Bank close the $100 million profitability gap in two years?”
- Context. What are the internal and external situations and complications facing the client, such as industry trends, relative position within the industry, capability gaps, financial flexibility, and so on?
- Success criteria. Understand how the client and the team define success and failure. In addition to any quantitative measures identified in the basic question, identify other important quantitative or qualitative measures of success, including timing of impact, visibility of improvement, client capability building required, necessary mindset shifts, and so on.
- Scope and constraints. Scope most commonly covers the markets or segments of interest, whereas constraints govern restrictions on the nature of solutions within those markets or segments.
- Stakeholders. Explore who really makes the decisions — who decides, who can help, and who can block.
- Key sources of insight. What best-practice expertise, knowledge, and engagement approaches already exist? What knowledge from the client, suppliers, and customers needs to be accessed? Be as specific as possible: who, what, when, how, and why.
The problem definition should not be vague, without clear measures of success. Rather, it should be a SMART definition:
Example situation – A family on Friday evening
Scenario: A mother, a father, and their two teenage children have all arrived home on a Friday at 6 p.m. The family has not prepared dinner for Friday evening. The daughter has lacrosse practice on Saturday and an essay to write for English class due on Monday. The son has theatre rehearsal on both Saturday and Sunday and will need one parent to drive him to the high school both days, though he can get a ride home with a friend. The family dog, a poodle, must be taken to the groomer on Saturday morning. The mother will need to spend time this weekend working on assignments for her finance class she is taking as part of her Executive MBA. The father plans to go on a 100-mile bike ride, which he can do either Saturday or Sunday. The family has two cars, but one is at the body shop. They are trying to save money to pay for an addition to their house.
What is the problem definition?
A statement of facts does not focus the problem solving:
It is 6 p.m. The family has not made plans for dinner, and
they are hungry.
A question guides the team towards a solution:
1. What should the family do for dinner on Friday night?
2. Should the family cook dinner or order delivery?
3. What should the family cook for dinner?
4. What should the family cook for dinner that will not require spending more than $40 on groceries?
5. To cook dinner, what do they need to pick up from the supermarket?
6. How can the family prepare dinner within the next hour using ingredients they already have in the house?
Stakeholder Analysis Worksheet
In completing the Problem Statement Worksheet, you are prompted to define the key stakeholders.
As you become involved in the problem-solving process, you should expand the question of key stakeholders to include what the team wants from them and what they want from the team, their values and motivations (helpful and unhelpful), and the communications mechanisms that will be most effective for each of them.
Using the Stakeholder Analysis Worksheet allows you to comprehensively identify:
- What you need from them
- Where they are
- What they need from you
The two most helpful techniques for rigorously structuring any problem are hypothesis trees and issue trees. Each of these techniques disaggregates the primary question into a cascade of issues or hypotheses that, when addressed, will together answer the primary question.
A hypothesis tree might break down the same question into two or more hypotheses.
Example: Alpha Manufacturing, Inc.
Problem Statement: How can Alpha increase EBITDA by $13M (to $50M) by 2025?
The hypotheses might be:
- Alpha can add $125M revenues by expanding to new customers, adding $8M of EBITDA
- Alpha can reduce costs to improve EBITDA by $5M
These hypotheses will be further disaggregated into subsidiary hypotheses at the next level of the tree.
The aim at this stage is to structure the problem into discrete, mutually exclusive pieces that are small enough to yield to analysis and that, taken together, are collectively exhaustive.
Articulating the problem as hypotheses, rather than issues, is the preferred approach because it leads to a more focused analysis of the problem. Questions to ask include:
- Is it testable – can you prove or disprove it?
- It is open to debate? If it cannot be wrong, it is simply a statement of fact and unlikely to produce keen insight.
- If you reversed your hypothesis – literally, hypothesized that the exact opposite were true – would you care about the difference it would make to your overall logic?
- If you shared your hypothesis with the CEO, would it sound naive or obvious?
- Does it point directly to an action or actions that the client might take?
Quickly developing a powerful hypothesis tree enables us to develop solutions more rapidly that will have real impact. This can sometimes seem premature to clients, who might find the “solution” reached too quickly and want to see the analysis behind it.
Take care to explain the approach (most important, that a hypothesis is not an answer) and its benefits (that a good hypothesis is the basis of a proven means of successful problem solving and avoids “boiling the ocean”).
Often, the team has insufficient knowledge to build a complete hypothesis tree at the start of an engagement. In these cases, it is best to begin by structuring the problem using an issue tree.
An issue tree is best set out as a series of open questions in sentence form. For example, “How can the client minimize its tax burden?” is more useful than “Tax.” Open questions – those that begin with what, how, or why– produce deeper insights than closed ones. In some cases, an issue tree can be sharpened by toggling between issue and hypothesis – working forward from an issue to identify the hypothesis, and back from the hypothesis to sharpen the relevant open question.
Once the problem has been structured, the next step is to prioritize the issues or hypotheses on which the team will focus its work. When prioritizing, it is common to use a two-by-two matrix – e.g., a matrix featuring “impact” and “ease of impact” as the two axes.
Applying some of these prioritization criteria will knock out portions of the issue tree altogether. Consider testing the issues against them all, albeit quickly, to help drive the prioritization process.
Once the criteria are defined, prioritizing should be straightforward: Simply map the issues to the framework and focus on those that score highest against the criteria.
As the team conducts analysis and learns more about the problem and the potential solution, make sure to revisit the prioritization matrix so as to remain focused on the highest-priority issues.
Example: Alpha Manufacturing, Inc.
Problem Statement: How can Alpha increase EBITDA by $13M (to $50M) by 2025?
The issues might be:
- How can Alpha increase revenue?
- How can Alpha reduce cost?
Each of these issues is then further broken down into deeper insights to solutions.
Analyses and Workplan
If the prioritization has been carried out effectively, the team will have clarified the key issues or hypotheses that must be subjected to analysis. The aim of these analyses is to prove the hypotheses true or false, or to develop useful perspectives on each key issue. Now the task is to design an effective and efficient workplan for conducting the analyses.
Transforming the prioritized problem structure into a workplan involves two main tasks:
- Define the blocks of work that need to be undertaken. Articulate as clearly as possible the desired end products and the analysis necessary to produce them, and estimate the resources and time required.
- Sequence the work blocks in a way that matches the available resources to the need to deliver against key engagement milestones (e.g., important meetings, progress reviews), as well as to the overall pacing of the engagement (i.e., weekly or twice-weekly meetings, and so on).
A good workplan will detail the following for each issue or hypothesis: analyses, end products, sources, and timing and responsibility. Developing the workplan takes time; doing it well requires working through the definition of each element of the workplan in a rigorous and methodical fashion.
This is the most difficult element of the problem-solving process. After a period of being immersed in the details, it is crucial to step back and distinguish the important from the merely interesting. Distinctive problem solvers seek the essence of the story that will underpin a crisp recommendation for action.
Although synthesis appears, formally speaking, as the penultimate step in the process, it should happen throughout. Ideally, after you have made almost any analytical progress, you should attempt to articulate the “Day 1” or “Week 1” answer. Continue to synthesize as you go along. This will remind the team of the question you are trying to answer, assist prioritization, highlight the logical links of the emerging solution, and ensure that you have a story ready to articulate at all times during the study.
McKinsey’s primary tool for synthesizing is the pyramid principle. Essentially, this principle asserts that every synthesis should explain a single concept, per the “governing thought.” The supporting ideas in the synthesis form a thought hierarchy proceeding in a logical structure from the most detailed facts to the governing thought, ruthlessly excluding the interesting but irrelevant.
While this hierarchy can be laid out as a tree (like with issue and hypothesis trees), the best problem solvers capture it by creating dot-dash storylines — the Pyramid Structure for Grouping Arguments.
Pyramid Structure for Grouping Arguments
- Focus on action. Articulate the thoughts at each level of the pyramid as declarative sentences, not as topics. For example, “expansion” is a topic; “We need to expand into the European market” is a declarative sentence.
- Use storylines. PowerPoint is poor at highlighting logical connections, therefore is not a good tool for synthesis. A storyline will clarify elements that may be ambiguous in the PowerPoint presentation.
- Keep the emerging storyline visible. Many teams find that posting the storyline or story- board on the team-room wall helps keep the thinking focused. It also helps in bringing the client along.
- Use the situation-complication-resolution structure. The situation is the reason there is action to be taken. The com- plication is why the situation needs thinking through – typically an industry or client challenge. The resolution is the answer.
- Test the pyramid. You can assess the rigor of your pyramid using three simple tests:
- Down the pyramid: does each governing thought pose a single question that is answered completely by the group of boxes below it?
- Across: is each level within the pyramid MECE?
- Up: does each group of boxes, taken together, provide one answer – one “so what?” – that is essentially the governing thought above it?
- Test the solution. What would it mean if your hypotheses all came true?
Three Horizons of Engagement Planning
It’s useful to match the workplan to three horizons:
- What is expected at the end of the engagement
- What is expected at key progress reviews
- What is due at daily and/or weekly team meetings
The detail in the workplan will typically be greater for the near term (the next week) than for the long term (the study horizon), especially early in a new engagement when considerable ambiguity about the end state remains.
It is at this point that we address the client’s questions: “What do I do, and how do I do it?” This means not offering actionable recommendations, along with a plan and client commitment for implementation.
The essence of this step is to translate the overall solution into the actions required to deliver sustained impact. A pragmatic action plan should include:
- Relevant initiatives, along with a clear sequence, timing, and mapping of activities required
- Clear owners for each initiative
- Key success factors and the challenges involved in delivering on the initiatives
Crucial questions to ask as you build recommendations for organizational change are:
- Does each person who needs to change (from the CEO to the front line) understand what he or she needs to change and why, and is he or she committed to it?
- Are key leaders and role models throughout the organization personally committed to behaving differently?
- Has the client set in place the necessary formal mechanisms to reinforce the desired change?
- Does the client have the skills and confidence to behave in the desired new way?
Great problem solvers identify unique disruptions and discontinuities, novel insights, and step-out opportunities that lead to truly distinctive impact. This is done by applying a number of practices throughout the problem-solving process to help develop these insights.
Expand: Construct multiple perspectives
Identifying alternative ways of looking at the problem expands the range of possibilities, opens you up to innovative ideas, and allows you to formulate more powerful hypotheses. Questions that help here include:
- What changes if I think from the perspective of a customer, or a supplier, or a frontline employee, or a competitor?
- How have other industries viewed and addressed this same problem?
- What would it mean if the client sought to run the company like a low-cost airline or a cosmetics manufacturer?
Link: Identify relationships
Strong problem solvers discern connections and recognize patterns in two different ways:
- They seek out the ways in which different problem elements – issues, hypotheses, analyses, work elements, findings, answers, and recommendations – relate to one another.
- They use these relationships throughout the basic problem-solving process to identify efficient problem-solving approaches, novel solutions, and more powerful syntheses.
Distill: Find the essence
Cutting through complexity to identify the heart of the problem and its solution is a critical skill.
- Identify the critical problem elements. Are there some issues, approaches, or options that can be eliminated completely because they won’t make a significant difference to the solution?
- Consider how complex the different elements are and how long it will take to complete them. Wherever possible, quickly advance simpler parts of the problem that can inform more complex or time-consuming elements.
Lead: Stay ahead/step back
Without getting ahead of the client, you cannot be distinctive. Paradoxically, to get ahead – and stay ahead – it is often necessary to step back from the problem to validate or revalidate the approach and the solution.
- Spend time thinking one or more steps ahead of the client and team.
- Constantly check and challenge the rigor of the underlying data and analysis.
- Stress-test the whole emerging recommendation
- Challenge the solution against a set of hurdles. Does it satisfy the criteria for success as set out on the Problem Statement Worksheet?
Harness the Power of Collaboration
No matter how skilled, knowledgeable, or experienced you are, you will never create the most distinctive solution on your own. The best problem solvers know how to leverage the power of their team, clients, the Firm, and outside parties. Seeking the right expertise at the right time, and leveraging it in the right way, are ultimately how we bring distinctiveness to our work, how we maximize efficiency, and how we learn.
When solving a problem, it is important to ask, “Have I accessed all the sources of insight that are available?” Here are the sources you should consider:
- Your core team
- The client
- The client’s suppliers and customers
- Internal experts and knowledge
- External sources of knowledge
- Communications specialists
The key here is to think open, not closed. Opening up to varied sources of data and perspectives furthers our mission to develop truly innovative and distinctive solutions for our clients.
Sources and additional reading
- McKinsey Staff Paper 66 — not published by McKinsey but possibly found through an internet search
- The McKinsey Way, 1999, by Ethan M. Rasiel