Signal-to-Noise ratio and the elimination of the nonessential

One of the coolest, most useful books I have on my shelf is Universal Principles of Design. This is a beautifully simple book and one that is immensely useful, a must for professionals and leaders from any discipline. The subtitle of the book pretty much sums it up: "100 Ways to Enhance Usability, Influence Perception, Increase Appeal, Make Better Design Decisions, and Teach Through Design." Even trained designers will want this book somewhere on their shelf. Each of the 100 principles — most of them applicable to presentation design — is summarized with great clarity and with good visual examples in just two pages. References are given for each principle for those who want to go deeper, but for a quick reference, this book can't be beat.

Signal-to-noise ratio
The principles are presented in alphabetical order, beginning with "80/20 Rule" and "Accessibility," and finishing 210 pages later with "Uncertainty Principle" and "Weakest Link." Between the principles of "Shaping" and "Similarity" you will find a good summary of the "Signal-to-Noise Ratio" (SNR). The SNR principle is borrowed from more technical fields such as radio communications and electronic communication in general, but the principle itself is applicable to design and communication problems in virtually any field. The authors sum up the signal-to-noise ratio this way:

The ratio of relevant to irrelevant information in a display. The highest possible signal-to-noise ratio is desirable in design.

 
Above: A couple of older slides (from a Steve Ballmer keynote in 2005) with rather low signal-to-noise ratios.


Above: Even Steve Jobs can present simple data in a way that complicates rather than simplifies (or did the faux marble texture and 3D help illuminate?). Note too that the baseline starts at 1000 (though it is hard to tell where the baseline is).

"Excess is noise"
Ensuring the highest possible signal-to-noise ratio means communicating (designing) clearly with as little degradation to the message as possible. Degradation to the message can occur in many ways such as with the selection of inappropriate charts, using ambiguous labels and icons, or unnecessarily emphasizing items such as lines, shapes and symbols, etc. that do not play a key role in support of the message. In other words, if the item can be removed without compromising function, then strong consideration should be given to minimizing the element or removing it all together. For example, lines in grids or tables can often be made quite thin, lightened, or even removed. And footers and logos, etc. can usually be removed with good results (assuming your company "allows" you to). In a nutshell, the authors put it this way:

Every element in a design should be expressed to the extent necessary, but not beyond the extent necessary. Excess is noise.

In Visual Explanations, Edward Tufte refers to an important principle in harmony with SNR called "the smallest effective difference":

"Make all visual distinctions as subtle as possible, but still clear and effective."

                                                                    — Edward Tufte

Related to the smallest effective difference is Occam's razor, which both Tufte and the authors of The Universal Principles of Design point out. Tufte sums up the Occam's razor this way:

"What can be done with fewer is done in vain with more."

Tufte goes on to say that a "...happy consequence of an economy of means is a graceful richness of information, for small differences allow more differences."


Above: This is a fake "Before" slide I made up rather quickly. The data being displayed is extremely simple, yet the eyes have to work pretty hard to get at the data. 98 new members in the first quarter of 2007 would be a very significant fact for the club, but perhaps a declarative sentence on the slide rather than a title would be more appropriate. The "noise" can be reduced in myriad ways (below).

 
Above: Examples of simpler ways to show the same data. Even the baseline was removed (left) since the bars define the endpoint, still a thin baseline may be appropriate as Tufte points out in The Visual Display of Quantitative Information. There are many ways to display this kind of basic data; the two slides above are not necessarily "the best" way.

But is the non-essential always "noise"?
Do design elements which are not absolutely essential necessarily detract from a design? Occam's razor says that unnecessary elements decrease the design's efficiency and increase the possibility of unintended consequences. But does this mean that we must be ruthless and remove everything which is not absolutely "essential" to a design? There are those who say a minimalist approach is certainly best (and also beautiful); I tend to fall in to that camp. But efficiency itself is not necessarily an absolute good or always the ideal. (Would one admire the work of an efficient purse-snatcher for example?) Nonetheless, when it comes to the display of quantitative information (charts, tables, graphs, etc.), I strongly favor display designs which include the highest SNR as possible. With other visuals, however, we may want to consider including or retaining elements which serve to support the message at a more emotional level. This may seem like a contradiction with the principle of a high SNR, or the Occam's razor, and the idea that "less is more." However, sometimes emotional elements matter (sometimes a lot). John Maeda pointed this out in his book, The Laws of Simplicity. Maeda insists that the principle of reduction (removing the nonessential) is important, but he also admits that emotion is very important as well and that often more emotion is better than less:

"When emotions are considered above everything else, don't be afraid to add more ornament or layers of meaning."

                                                                  — John Maeda

Design makes things clear, Maeda says, but art — the stuff of emotions — makes us wonder. Design can bring clarity to a message, art can help bring meaning. "Sometimes...clarity alone is not the best design solution." Presentation design is as much art as it is science, and, of course, aesthetics do matter. The Aesthetic-Usability Effect (also in the Universal Principles of Design) says that "aesthetic designs are perceived as easier to use than less-aesthetic designs." First impressions based on aesthetics are important, for example, as the way people think about or interact with a design will be influenced at some level by how the design looks or feels (to them). Clarity and the reduction of the nonessential are important, but we need to remember too that how a design looks influences perception.


The Google website is often mentioned when talking about sites that have cut everything visually except that which is essential. Yet there is emotion there too; the logo is large, colorful, and even "playfully seasonal."

As with all things...balance
Understanding what is noise and what is excess is an important design consideration. As pointed out in this paper by Michael Albers on SNR in documentation, what is considered noise will depend on the context. "An excess of noise can occur from either too much or too little information," says Albers. "... much of the real difficulties in communicating information do not fall within the technical realm. [They] fall within the people realm which revolves around the contextual aspects of the information." Use depends, then, on our particular circumstance, audience, and objectives. Ideas like SNR are good principles, but not rules to be blindly followed. In my opinion, designs with a high signal-to-noise ratio are not only generally clear, they often look good as well. But in the end, SNR is one principle among many to consider when creating visual messages.

Related links from Presentation Zen
• 2-D or not 2-D
• Noise and the elimination of the non-essential
• Wabi-Sabi and Presentation Visuals (part I)
• Wabi-Sabi and Presentation Visuals (part II)
• Wabi-sabi simplicity as it relates to interface design (from 37 Signals)

Comments

[Paul]

I really liked this post. Being a fellow presenter, I too lean towards minimalistic design in presentations. I point people to your site often to learn how to clean up there power point. Now I plan to go book shopping for some of your recommendations. Thanks.

[Gray Miller]

Garr, I used to address this when I was working with University students in the field of dance and technology. In those cases, the "noise" often came from the the extravagant use of technology (video projectors, keyboards, TVs onstage, etc) that tended to draw focus away from the actual dancework being performed. Not that they can't be integrated--but in dance, as in presenting, it's important to focus on the message you want to convey. If that message IS the medium (to invoke His Excellency MacLuhan) then fine, go nuts on the stage. But if it's not...know that having cables and prominent projectors and DVD menus showing will decrease the signal-to-noise ratio.

Do you think the technology used to give presentations detracts from the message? I see some of the old TED talks, and I cringe at the potpourri of equipment strewn around the stage...but people don't seem to mind.

[Marcel van Oosterwijk]

The quest for the 'empty' slides ends often up in a struggle against an obligatory company template which is full of irrelevant stuff that takes the attention away from the actual content.

Sometimes in such a struggle, I threaten to stop using slides together and use whiteboard to paint the pictures 'live.'

What do you guys do to avoid the Terror of the Template?

[Jorge Camoes]

Let me start by quoting MacEachren ("How Maps Work"): "This approach [signal to noise] was doomed to failure, if for no other reason than that the user can combine map information with previous knowledge to produce conclusions that were not part of the initial map message."

We too often import concepts and theories from the technical fields without proper assessment of the differences to the way mind and society work. To some extent this works, but at some point it is a barrier to a better and more complex knowledge.

Regarding charts: I do like minimalistic designs, but, since I am not an artist, it doesn't really matter what I like or don't like. Rather, I have to evaluate the "rhetoric situation" where communication takes place and choose the visualization that meets the specific requirements (without compromising data integrity and some other basic values). Of course, it is my job to decide what works and what doesn't and to evaluate the possible consequences of my choices.

[Rich...!]

Awesome post. I really do agree with the principle message, I do however feel that sometimes we'd do well to support the metaphor in the visual too.

So if Joe Presenter wanted to talk about the amount of steel used in their new building. He could use a simple bold number to display the amount, however if he added a pic of the Eiffel Tower and said, "that's more steel that the Eiffel Tower, the may forget the number (they usually will), but that superfluous pic will help the remember the metaphor, and that's usually enough...!

[Matt James]

As usual, your post was excellent!

I have 2 comments:

1) I find it interesting that everytime I read a quote from Tufte, I find it extraordinarily difficult to understand what he is saying. Usually, the surrounding blog post explains it much better than he. I think the big reason is that he seems to be efficient with his word choice, but that doing so often makes what he is trying to communicate more dense that it needs to be. I often wonder why he can't choose to present his message in more colloquial terms.

2) The only part of the post that I disagreed with was the selection of Google's homepage as a representation of an aesthetic victory. In fact, seeing the graphic at the bottom of the screen before I read the surrounding text, I expected you to say that Google is an example of far too much simplicity without attention to the emotional impact of design. I am glad that you chose point to the logo as an example for your argument as I find it to be the ONLY aesthetically pleasing thing on the page. The rest of the page, by contrast, is extremely bland to me. They don't even change the default color of the links (which could be considered a good thing from a completely utilitarian perspective). Google's web applications are another example of the obvious need for a graphic designer. Apple is one of few companies that I think really masters the art aesthetic minimalism. They can capture the tranquility of emptyness without making their designs seem empty. Google can usually hit the target of simplicity, but rarely that of good visual design.

[Kian Ann]

Awesome post! I am an advocate for minimalist designs - whether it is for powerpoint slides or web pages. The extra stuff simply clutters the page.

While a simple design may not always catch the eye of your reader, it serves its purpose of bringing across a message.

[Dmitry Linkov]

Thanks for this fantastic post! I'll definitely look for the book.

[Marc Duchesne]

Hi Garr,

I would like to take the opportunity of this great (as usual ;-) post to emphasize the key role of Raymond Loëwy in today's Fundamentals of Design at large. Remember Loëwy's book : "Never Leave Well Enough Alone". It's all there. Even the front cover : an egg, standing on a table.

Quote : "It would seem that more than function itself, simplicity is the deciding factor in the aesthetic equation. One might call the process beauty through function and simplification."
Your post is about slides-based presentations : the function of a slide is to convey a message. So, to do it in the most simple manner is what Raymond Loëwy told us decades ago.

Quote Two : "Between two products equal in price, function and quality, the one with the most attractive exterior will win."
In your post, we can see a typical application of that principle with Ballmer vs. Jobs.

Quote Three : "I once said that the most difficult things to design are the simplest. For instance, to improve the form of a scalpel or a needle is extremely difficult, if not impossible. To improve the appearance of a threshing machine is easy. There are so many components on which one can work."
Here, we can get an explanation of Ballmer's and Jobs's styles : Microsoft does threshing machines ;-)

[Niko Neugebauer]

Excellent post, Garr,
which reminded of one my favorite quotes of Antoine de Saint-Exupéry: "Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away"

[Vietnam tours]

Good afternoon, everybody. I'm a Vietnamese. Do you toured whatever times? If not already please come with our country to discover Vietnam travel agencies, Vietnam tour operators and Vietnam holidays.

[Vera Bomfim]

Por favor solicito que as infomações acima sejam traduzidas para a língua portuguesa.
Obrigada

[Todd]

Garr,
Brilliant! You are a most generous blogger. The richness of your posts is a welcome interlude in my day. Just posting to say thanks for taking the time to share.
Cheers,
Todd

[sandhya]

good moring sir........
am doing a project in wireless communication .
i need som einformation on SNR ratio can u send me some plz
thanx..
bye