The Policy Analyst’s Version of Guy Kawaski’s 10/20/30 Rule—and ideas from Tufte and Cohen

Guy Kawasaki’s now classic article “The 10/20/30 Rule of PowerPoint” provides three simple and powerful rules: “a PowerPoint presentation should have ten slides, last no more than 20 minutes, and contain no font small than thirty point” — hence 10/20/30.

These rules are helpful for almost anyone who must create a slide presentation — but for policy analysts and researchers, I’m going to argue for two tweaks: #1 your deck should be 12/20/24 and #2 the topics you cover should answer a different set of questions.

Let’s start with the number of slides. Why do I argue for 12? Because this allows you to answer the key questions that almost any kind of policy analysis study needs to address. These are:

  1. What is the issue?
  2. Why does the issue matter? To whom and under what circumstances?
  3. What’s the background on this issue?
  4. Where are the “problems” and/or gaps that are most key to consider and why?
  5. What is your research question?
  6. What is your analysis strategy and why did you choose it?
  7. What data will you analyze and why did you choose these data?
  8. What are your findings?
  9. What are your policy recommendations and how do they follow from your findings?
  10. What are the limits of your study?
  11. What is the next step required to reduce uncertainty or begin implementing?
  12. In summary, what is the research question, key findings, and recommendations—and how do I reach you?

A title slide is usually a good idea and, in some cases, I also include a references slide, or even have additional supporting data and figures at the end—but with these caveats, a good analyst should be able to cover all the bases in 12 core slides.

And, in fact, it’s possible to combine sequential pairs of almost any set of the first 11 slides above into a single slide, which can reduce your overall slide count as needed. The exact number of slides should be driven by the content, but a rule is a heuristic to make things easy — so aim for no more than 12.

Kawasaki recommends that you present your slides in 20 minutes—which leaves the rest of a typical hour meeting for questions and discussion. There’s good science backing this general point: We are limited in our ability to understand more than a handful of concepts, and it’s difficult for the average adult to focus for more than about 20 minutes.

And so, even though the context is quite different, 20 minutes is a good goal to aim for. If you need to create a longer presentation, try to break it up into 10- to 20-minute segments and leave time for questions and discussion in between.

The last part of the 10/20/30 rule is that text on slides should be no smaller than 30-point font—but, Guy allows, if that’s too “dogmatic” you can take “the age of the oldest person in your audience and divide it by two” to find the optimal font size.

Since the resolution of the average slide presentation system has increased quite a bit since Kawasaki’s essay was written in 2005—and since analysts sometimes must be a bit wordier, I’m going to argue for making your text no smaller than 24-point—although I usually start at 28-point for bullets and 36-point for slide titles.

Here, too, the exact size doesn’t matter—especially because fonts vary widely in how dense they are. But Kawasaki’s observation resonates with my experience in policy organizations and at conferences: Presenters often use small font because they either “don’t know their material” and need to read the text, or they think that more text is better.

Both are terrible reasons to write a lot of text—but analysts face additional challenges. First, subject matter experts (SME) are prone to “data dump” syndrome; and secondly analysts are often not great writers.

A single blog post can’t turn you into a top-notch scribe (although resources abound online), but aim for, what James Salter called, the lightening rod of “brevity, clarity and wit…” And note that wit doesn’t necessarily mean humor, but also means “keen perception” and the ability to make connections between ideas.

To avoid data dump syndrome, you should always ask “what does this audience need to know to understand my argument?” In a typical project, you may run many tests and generate a lot of data, but when you make your case, prune these branches—after all, you can always link to other documents or provide additional slides after the conclusion.

Are slides the answer? What’s the question?

Kawasaki’s advice is sage, but slideware poses deeper problems as well. Edward Tufte’s fantastic essay on the The Cognitive Style of PowerPoint argues, most fundamentally, that sometimes complex thoughts and ideas are resistant to bullet points. Since a typical slide can’t contain more than about 40 words, presenters often feel the need to create slide after slide to express a concept. This linear and segmented approach, what Tuftes calls “one damn slide after another,” harms our ability to make connections across topics and understand context.

Did an endless array of nested bullets and eye-glazing text cause the Columbia Space Shuttle disaster? Whatever the case, when you have to present a deck, take Tufte’s advice to heart: For each idea you want to communicate, ask yourself “Is there a map, diagram, photo, graphic or figure that will provide more insight or understanding than a set of bullets?” If the answer is yes, then create a simple visualization that avoids “chartjunk”—and instead of projecting your text, tell it to your audience.

Since Kawasaki and Tufte are in agreement that you can’t read your slides, how should you keep track of your notes? Most major packages provide a note area, but I find that once I have written and revised my notes, then I don’t need to look at them. However it’s always a good idea to write your notes down as bullet points and keep a copy close at hand.

Let me close this essay by adding on some advice from Steve Cohen, Executive Director of Columbia University’s Earth Institute: For important meetings, always print one copy of your deck for each attendee—and print a copy for yourself with your notes.

If the meeting size is large, or you want to conserve precious resources, then print at least two copies—one with your notes and one with just the slides. This backup copy can be use to show visuals in a small meeting or it can be run through a copy machine at the last minute. I would add that you should always also have your deck available as a PDF along with the native file in case of software issues. And even though cloud systems are ubiquitous, have a backup on your email or, ideally, a thumb drive as well. Even in our hyper-connected age, technical glitches crop up all the time.

For many topics and in many settings, presenting slides is a fact of life. Policy analysts and social science researchers should follow my 12/20/24 rule—but keep Tufte in mind and always ask how you can visualize text. And don’t forget Dr. Cohen’s advice: When the computer crashes and the projector can’t connect, paper can save the day.


Useful Resources: Statistics and Data Visualization

In out current era, even nontechnical people need to have some understanding of data science and statistics. What’s confusing is that so many terms—data science, big data, machine learning, statistics—get bandied about with little explanation. To understand how the pieces fit together, I highly recommend this article from the Win Vector blog.

The R programming language is another tool that is making headlines. From Google’s use of R for analysis to Microsoft buying IDE maker Revolution Analytics, R is all the buzz. What is it? Why use R? The R Bloggers website answers these questions and their email list provides a nicely curated flow of articles that will keep you up-to-date.

For folks who are programming already, the Quick-R page is very useful, especially when you are trying to find equivalent R code for a Stata routine.

As a former graphic designer and communications officer, I’ve spent a lot of time thinking about how to present research. As this nice overview article by the WSJ explains, in an era of Big Data, visualization is going to be a key area of inquiry for firms, researchers and the public.

A good introduction to data visualization can be found on Edward Tufte’s website and in his books (the Visual Display of Quantitative Information is well worth buying and his traveling seminar is also great).

For a more contemporary spin on these issues, the Flowing Data website and book are also worth a look. For true visualization geeks, the Parsons Journal for Information Mapping is a fantastic (and free) publication put out the the Parsons Institute.

If you are working on a project and need a resource to inspire creative approaches to visualization, I highly recommend the Visual Literacy project’s Periodic Table of Visualization Methods.