McKinsey on the McKinsey cost curve

A decade of strategy consulting work at McKinsey has not made me a big believer in standard frameworks. Most business problem require a tailor-made approach without the buzz words and generic statements you find in most airport business best sellers.

There is another problem with frameworks: a framework to solve a business problem is usually not the framework to communicate a business solution. Problem solving and presentation are two different things.

McKinsey's "enduring ideas" series periodically discusses one of the classic frameworks from the world of management consulting. This month it discussed the cost curve.

• On the vertical axis you show the cost per unit
• On the horizontal axis you line up the competitors in order of their production cost
• (Unusual) you change the width of the column to reflect the production capacity of a player
• Drawing a vertical line where capacity = demand shows you what the market price of the (commodity) product will be, and who is making money/who is not.

This framework is maybe an exception. A slightly modified column chart can serve both as a problem solving tool and a communication instrument. If there is incomplete information, 3 people can spend 2 months to develop it (running all the analysis), but once it is there, it shows what's going on in a (commodity) industry on one very insightful piece of paper (piece of PowerPoint slide). 

How to do a McKinsey-style source of change chart

Some numbers today. The source of change is a tool to explain the delta between two numbers in terms of its components. Assume you need to get the story below across in a crisp presentation.

The first thing is to understand what's going on. Get some more information until you have the full picture in a clear table.

Now let's do the analysis. This is the tricky part, the text below does not do a good job in explaining this, you can click the spreadsheet for a bigger and more visual explanation.

1. Calculate the profit in the "before" scenario using a formula that just uses inputs
2. Now stretch each of the variables that change to their "after" value, jot down the value, and return the value back to its original number
3. Repeat for all the variables and see what delta in profit you managed to explain.
4. Calculate what is left to explain, and allocate that to the individual values.

Finally put the values in a nice waterfall chart.

Chart concept: the 2x2 matrix and other grouping techniques

McKinsey and other management consultants love 2x2 matrices (and obviously 3x3s). Personally, I think they are often overused (framework overload).

Not every categorization can be crammed into this framework.

• The axes need to be logical
• The groups needs to lead to 4 categories, i.e., leaving one or two boxes as "not applicable" does not make sense
• They work particularly well when you want to show things moving from one category to the other
• They are good to show that something stands out (from for example the competition) by popping up in the top-right corner

Here are some other techniques to group items on a PowerPoint slide using line and venn diagrams:

• Diagram 2 - "you cannot have it both ways"
• Diagram 3 - "the best of both worlds"

McKinsey interviews Gartner CFO on communicating with investors

Presentations to analysts and investors are very important for publicly traded companies: they have a direct impact on the share price.

In this interview for the McKinsey Quarterly, Gartner CFO Christopher Lafond lays out his investor relations strategy:

• Segment and prioritize investors you want to talk to (in his case funds that have a long-term interest in Gartner, and do their homework to understand the fundamentals of the business)
• Focus on a very limited number of performance metrics, not shying away from internal, operational metrics that are usually only discussed by management
• Educate investors why these are important

Note: (free) registration to the McKinsey Quarterly might be required to read this article.

Not all presentations are "Zen" - different formats for different settings

Not all presentation settings are the same. A "Presentation Zen" slide show with stunning images and the incidental word on a slide is great for a keynote, but might be a bit too much to discuss last quarter's financial results. The 50 page deck with bullet point slides might be serve better as a printed business plan than the key communication tool for a 20 minute VC funding pitch. I have tried to describe 6 presentation scenarios and categorized them according to:

• Whether the  presenter is present or not
• The amount of detail/data inside the document

Here we go (click image for bigger picture):

1. The key note is the classical "Zen" presentation. Huge fonts, dark background, few words, large images.
2. The pitch is similar to the key note, with the difference that it might be shorter, and does contain some more data to answer questions from the much smaller audience.
3. The meeting presentation is probably done on a light background, and contains much more facts and details. Over-simplified slides with beautiful pictures do not work in the small conference room with people ready to go through raw material. McKinsey and other consulting firm's presentation often fit in this box.
4. The slideshare (or online) presentation is something relatively new. People see it typically in small windows, i.e., fonts should be big, pictures should be nice. The audience of this presentation is highly impatient, clicking rapidly to reach the end, and aboning your presentation if it is not interesting enough. No animations here.
5. The email attachment is similar to the key note presentatation with an important difference that it needs to stand on its own, titles need to explain the messages in the charts. Some animation could be used here (sparingly though). Detail is less than the handout.
6. The handout contains the full detail, the full text. It should be prepared on a white background (people will often print it) and use no animation (again, does not come out in print). For VC pitch situations, the good handout makes the business plan "brick" obsolete (hardly anyone reads these anyway).

The following colorful diagram makes an attempt to visualize the above. A bit busy, the main message is that things are different in each scenario.

Critiquing a McKinsey exhibit - analysis charts are not presentation slides

PowerPoint slides that are used to analyze data and solve problems inside a team are not always the best exhibits to be used in presentations. To illustrate the point, I am using a chart out of a publicly available McKinsey document: the 2008 Chinese Consumer Report published by McKinsey's Insights China initiative. 

The report itself is formated very nicely with beautiful pictures, the content of the report looks very interesting. I am just focusing on the cosmetics of one chart as an example. (Click image for a larger picture)

• Overall, there is too much information on the slide, but given that this chart is used inside a long-hand document, that might not be too bad
• Because of the red background, the title does not stand out clearly
• The boxes around the chart are too heavy, to separate the 3 charts, it would be better to use a very light background shading
• The box on the right is narrower than the other two.
• Because there are so many bars in this slide, they become very narrow, and hence they almost stop helping the reader see the differences in size, maybe boxes with simple numbers work better
• The gradient fill of the bar does not add to the understanding the chart
• The legend in the top right looks slightly lost
• The text labels are too long: “the store makes me feel reliable/trustworthy” is appropriate for a consumer questionnaire, but can be shortened in the chart
• The text labels “dance”, they are not right aligned, some labels take 1 line, some 2 lines, some 3
• The middle graph has 8 bars, the other two 7, disturbing the visual calm of the slide

In short, a great chart for a team brainstorm, but not for external audiences.

More posts on McKinsey and management consulting presentations

What do consultants mean when they say "you're not MEESEE?"

MECE (pronounced "meesee") stands for mutually exclusive, collectively exhaustive. Management consultants (McKinsey, BCG, etc.) use it a lot. What does it mean? In simple language, a text is MECE where are no holes, no overlaps. For example, the following structure is not MECE:

• Prices ....
• Volume .....
• Revenues ....

Revenues is the result of prices x volume, there is an overlap in this list. Why does being MECE matter? If you want to analyze a problem, it is important that you get all the component structure of the problem right: you need to get them all on the radar screen, and there should not be any dependency between the components. For example if you concluded that "costs are not the issue" and close the analysis here, you don't want it to re-emerge somewhere else in your problem solving effort. Why does being MECE matter in PowerPoint presentations? If you want to tell the audience a story  you don't want to confuse them by bringing up points again in an unclear structure. The logic should be clear. Should you apply it blindly in your PowerPoint presentations? No. MECE stories can have the perfect logical structure, but can also be boring. For example, in order to be complete, you need to address the rest of the world aftering discussing the most exciting markets US and Asia, even when they are not relevant. Be MECE when designing a story line structure, then adjust to make your presentation interesting and compelling.

The strategy consultant's review of The back of the Napkin by Dan Roam

We talked about sketching chart ideas on paper before, but Dan Roam takes visual problem solving to the next level in his book The back of the Napkin. This book was an interesting read for me not only because of the presentation concepts discussed, but also because The back of the napkin aims to provide a complete framework to solve business problems. (The key frameworks can be downloaded here for free). I think the book did really well on the presentation front, the goal of a generic strategic problem solving kit is not really reached. Dan does a great job convincing us that we should use our drawing/visual thinking skills that most of us have been neglecting since we started formal education. On top of that he provides practical guidelines to get going

• Have the courage to use a more informal drawing style (away from the computer) to get to the essence of problems, focus not on form but on content
• Help us think about what type of drawings are best to be used in which situations (who, what, when, why, etc.) and to what audiences (the visionary CEO, the detailed operations manager)

As a problem solving tool kit, he provides useful tools but falls short of providing a generic solution framework for all business problems (which impossible anyway I think).

• Dan takes the "S-type"/"sensing" approach to problem solving, spread out all data, put in on the walls, digest it all to see the bigger picture. A way of data processing very similar to the human brain sizing up a new environment. This is actually a useful and fresh approach compared to for example strategy firms such as McKinsey, that apply a very targeted data gathering approach focussed on key questions/issues that have been identified earlier. 
• Another take away for me were diagrams that try to summarize all relationships in a problem. Plot a variable on the x axis, one on the y axis, start adding bubbles in different sizes and different colors to analyze 5-6 dimensions in one diagram. Useful for solving problems, less for communicating results to a "cold" audience that is confronted with the material for the first time.
• I do think however that the book does not provide a simple step-by-step guide to solve problems, you need guidance for this. Running problem solving brainstormings around a white board requires a strong moderator, and picking the right diagrams requires experience. Hiring Dan's firm would probably do the trick, but the novice will find it difficult to apply the techniques after having read the just the book.

As a presentation tool, Dan's ideas are highly valuable in a smaller group setting, where everyone can gather around a white board while the presentor draws the story "live" in front of the audience without any help of PowerPoint. For the big audience however, this approach is high risk.

Classic McKinsey business frameworks

The McKinsey Quarterly added another framework to a series that discusses historic business analysis frameworks: the GE-McKinsey 9 box matrix, developed in the early 70s. Glancing over the article and hearing the audio explanation made me realize: things have moved on. Managers have become much more skilled in absorbing diagrams, and business portfolio analysis is more than plotting industry attractiveness and competitive strength. The elaborate explanation of the 3x3 would almost be an "insult" to a sophisticated management audience in 2008. Still, today I constantly invent "frameworks" for presentation on the spot: how to present 4 steps, 3 distinctive factors versus the competition, 6 phases to launch, the best of both 2 worlds, but none of them are standard. Also, I am increasingly moving away from the urge to summarize all factors and aspects of a solution on one (the first) page. Rather, I take an audience through a step-by-step story to the final answer. The difference between a concise document and an engaging live presentation. For more nostalgic reading:

• Classic frameworks by McKinsey & Company (7-S, Structure-Conduct-Performance [SCP], others)
• Classic frameworks by the Boston Consulting Group (Growth-share matrix [Stars, cash cows, dogs, question marks], others)

The McKinsey - or any consulting - presentation

The vast majority of Google traffic that lands on my site is looking for advice on how to write a "McKinsey presentations". Let's discuss them a little bit more, including my logic why they might NOT be suitable for just any communication situation. Why do they look so good and professional? A few reasons (some of which are good recommendations for any presentation you prepare)

• They stick to a strict slide format: every page is laid out exactly the same, making the whole document look very consistent
• Pages have muted colors and no spectacular animations.
• Consulting presentations are almost always all about numbers, and this quantitative data is displayed and structured in simple and clean data graphs (i.e., not an ugly, busy cut and paste from Excel), and numbers are rounded
• Each chart has a single message, which is written out in the chart title and clearly supported by the numbers in the chart body
• They (sometimes over-)use a lot of frameworks to structure information: a time line, the impact of a number of forces, evaluation of pros and cons, strenghts and weaknesses.
• The presentation has a clear logical structure, taking you step by step through an argument. A lot of energy is invested in the PowerPoint slide sorter: re-shuffling charts until the story is lined up the correct way. This process is not only for communication purposes, it is an integral part of problem solving. Trying to articulate a logical story will inevitably highlights flaws in logic, sending you back to the drawing boards to do additional analysis or change your recommendations.
• It is full of summaries. If you have 30 seconds to read a document, you will find the full story on page 1, if you have 5 minutes, you can read the summaries of the next subsections (each section explaining 1 paragraph of the summary in more detail), if you have more time you can read the whole document.

Where do these presentations work best? Not surprisingly: to present the results of a consulting project. The "answer" on page 1 supported by all the backup and analysis for people who need to be convinced, or to find the source of that 1 number a year after the project is finished.

What can you learn from them? Even if you are not a strategy consultant, your presentatations greatly benefit from consistent formats, colors, 1 message per chart, clean data graphs etc.

Where can you be different? Still assuming you are not a strategy consultant, your presentation style could be different in a number of ways.

• Don't completely give it all away on page 1. Especially for large audiences, try to create interest and take people along an interesting story only giving a hint of what you are going to do.
• Structure is not all about logic. A story line should be interesting and surprising for the audience, not a mathematically tight proof of a solution
• (Standard) frameworks are boring. Frameworks are great for solving problems, and not very good for communicating solutions. To be avoided.
• Summaries are not repeats. Don't give your presentation 3 times: on page 1, the body, and on the last page
• Use creative graphics, professional images, be more bold
• Avoid lingo, use your own language

In short, make the presentation your own!

First content, then structure

It might sound obvious, but it is not. Consulting projects start with a structure to lay out all the questions that need to be solved (let's look at the market, let's check our competitive advantages, etc. etc.) Some branches of theses "issue trees" (I made hundreds of them at McKinsey, Google "McKinsey issue tree" and you get a lot of examples) might turn out to be dead end streets, some branches might have to be expanded, re-written later on in the project. As more analysis comes in, the solution of the project will emerge, documented alongside the structure of your initial workplan. At this point I would say: cut, take a break, throw away the structure. Now that you have something to say, worry about how to say it. With the benefit of hine sight, build the structure that tells your story from scratch. Problem structure does not equal presentation structure.

How to create a McKinsey-style waterfall chart

The "water fall" chart is an effective way to summarize the quantitative impact of a number of drivers. For example, you need to put the following story in a chart: "Our profits went up by 7, the positive effect of higher prices and lower cost was offset by a lower sales volume." A waterfall chart would look something like this: For illustration purposes I left the light grey color and data labels of the supporting series in so you can see how to make the chart: it is basically a stacked bar chart with 3 series:

• A "white" series to support the drivers
• One series for the drivers
• One series for the (sub)totals

The data table for this chart (Powerpoint 2007): For a final touch, make the color of the light grey series white, take out the data tables and that's it. There is the temptation to make automated tools (in Excel) that do the work for you. Like almost all my charts, I start with a piece of paper and make my waterfalls manually, to make sure that they

• Are correct (negative numbers can make these charts a bit tricky to get right sometimes)
• The chart tells the story I want it to tell (what subtotals to use, in what order to list factors, etc.).

(Click on images for a larger picture)

Googling for free McKinsey PowerPoint templates

Doing a Google search for "McKinsey PowerPoint templates" highlights many entries that are almost all a violation of copy right. Moreoever, the templates are of little use to someone who is not working at McKinsey team on a client engagement. First of all: presentation starts with substance, then follow the frameworks (if any) A bit of historical context. I recognize the frameworks from my time at McKinsey, almost all of them are from the early 90s, when McKinsey was still working with an early pre-PowerPoint presentation tool called "Solo". Solo was developed specifically for McKinsey, later marketed as an independent application. It vanished when PowerPoint emerged, not because PowerPoint at that time was neccesarily better, but all of McKinsey's clients were running it and using it to edit presentations. (A slightly outdated looking site is still offering it for sale?). Your graphics assistant (nobody knew how to make charts themselves then), would dive into the template database to find "something that uses 4 arrows". All these frameworks were meant to be used in densely written strategy/micro-economics documents, not in convincing on-screen presentations. If you would like to learn about McKinsey's approach to graphics and presentations, try this:

• The content available on the McKinsey Quarterly site (most of which look actually better than "day-to-day" work)
• For the chart "Bible" that was used in the early 90s, flick through a copy of the book "Say it with charts" by Gene Zelazny. (I see he's updated it since I last saw it).
• The foundation of McKinsey's approach to writing logical story lines (but not always the most compelling stories that are important in presentations) can be found in Barbera Minto's book "The Pyramid Principle"
• UPDATE: I posted about a question I get often: how to make a McKinsey waterfall chart here.