Justin O’Beirne

Justin O'Beirne of San Francisco, California. Essays, projects, and contact information.

ROADS, STREETS, & THE SMALLEST EFFECTIVE DIFFERENCE
Why do Google's Street Maps seem Simpler & Easier to Understand?

Simple trumps complete.
NEIL HUNT, Netflix's Chief Product Officer

I’ve often thought Google’s street maps are easier to read than those of the other mapping sites. There’s just something about them that makes them seem a little simpler, a little more approachable. And whenever I’m looking at them, the street grid always seems clearer.

For example, take this comparison of Google Maps and Bing Maps:

Notice how much simpler Google’s roads look, compared to Bing’s. (Google’s are subtle, while Bing’s almost seem to shout at you.)

 

Now compare Google Maps to Yahoo Maps:

Compared to Yahoo, Google’s map once again seems simpler, cleaner, and easier to understand.
 

You can be forgiven for thinking that Bing and Yahoo show more streets than Google. Even though they don’t, it almost appears as if they do. And this why it’s so interesting: the maps above show the exact same streets—the only difference is in their presentation.

So why do Google’s streets seem simpler and easier to understand?

There are, at least, a couple of explanations:

(1) Google Maps has a simpler data model.

(2) Google Maps takes advantage of the “Smallest Effective Difference”.

Let’s look at Google’s simpler data model...

 

EXPLANATION #1: GOOGLE'S SIMPLER DATA MODEL

One of the reasons why Google Maps seems simpler is because it actually has a simpler data model (i.e., fewer road classifications) than Bing and Yahoo.

Let’s take a closer look at the maps from above. First, Bing:

If you look closely, you’ll notice that Bing features five distinct road classes: one type of highway (“1”), three types of arterial road (“2”, “3”, & “4”), and one type of local road (“5”).

 

Meanwhile, Yahoo features six distinct road classes:

Yahoo features two types of highway (“1” & “2”), three types of arterial road (“3”, “4”, & “5”), and one type of local road (“6”).

 

Now look at Google:

Google’s map has only three distinct road types — half Yahoo’s number. There’s one type of highway (“1”), one type of arterial (“2”), and one type of local road (“3”). See what I mean by simple?

 

In summary, Google has three distinct road classifications, while Bing has five and Yahoo has six. Interestingly, Google has only one type of arterial road, while Bing and Yahoo have three arterial classes each:

So Google has the simplest “data model” of the three. What are the consequences of this?

 

*   *   *

 

You may have noticed the Neil Hunt quote at the beginning of this post, “simple trumps complete.” Hunt left this quote as part of a Quora answer and later elaborated on it further:

“Simple trumps complete” - a 5% feature (used by less than 5% of all users) is a distraction for all other users, and is better removed, unless it’s really critical (a small number of users do need to cancel service, for example).

I have this mental model of particles of attention that a user brings, a finite quantity that they will spread around according to what catches their attention. I call them “attentrons”. An extra tab or button will attract a bunch of attentrons that are not then available to focus on other areas. So the tab had better be *better* than the competing areas of the site to avoid diluting the results, or it’s better off removed.

Reading Hunt’s words, I can’t help but wonder if the extra road classifications on Bing and Yahoo are a “5% feature”. It’s likely that only 5% of users – and perhaps even fewer – would ever care about them. And it’s also likely that fewer than 5% of users could even articulate the differences between the classifications. (After all, neither Bing or Yahoo offers a legend explaining the differences.)

Could it instead be that for the remaining 95% of users, the added colorings and widths needed to denote the extra classifications are distractions? And that they add unnecessary complexity to the maps?

Google seems to think so and has boldly reduced the number of road classifications on its maps. [1]

“Simple vs. Complete” — Bing’s and Yahoo’s “completeness” necessitates the use of additional colors and widths. But does this added complexity benefit most users? Or is it a confusing and unnecessary distraction?

 

I was initially hesitant to question Bing’s and Yahoo’s extra road classifications. After all, don’t these extra classifications add more detail, more information to the maps? But then I realized two things:

  • Without a key, it’s unclear what this added information is: The colorings and widths used to differentiate the road classifications offer no clue as to how the roads differ. Are they “truck routes”? Do they have more traffic? Higher speed limits? More lanes? Are they divided or limited-access highways? It’s unclear why a motorist should want to take a yellow road over a white one, or an orange road over a yellow one on Yahoo’s map. Or a gray arterial over a lavender one on Bing’s map. [2]
     
  • Whatever this added information is, it’s probably not useful or actionable enough to warrant increasing the map’s complexity. After all, when was the last time you heard someone complaining that there weren’t enough road classifications on Google Maps? [3]

Google has made a very interesting design decision by choosing a simpler data model for its maps’ streets and roads. Indeed, one of the advantages of this is that Google has fewer road classifications to differentiate, compared to the other sites. Let’s now take a closer look at how each site differentiates its streets… 

 

EXPLANATION #2: GOOGLE TAKES ADVANTAGE OF THE “SMALLEST EFFECTIVE DIFFERENCE”

It’s really interesting how each site chooses to differentiate its streets, and this has a large effect on how “simple” we perceive each map to be. Let’s take a look at an identical area of Manhattan on each of the sites. As you examine the images below, pay close attention to the differences between the streets:

(From top to bottom: Google Maps, Bing Maps, and Yahoo Maps.)

 

You could say that the differences between the streets are much more emphasized on the Bing and Yahoo maps than they are on the Google map. And this is not without consequence: When you first looked at the Yahoo map, chances are that the differences in the colors and the widths were the first things you noticed—perhaps even more than the “map” itself. And they were probably also the first things you noticed when you looked at Bing’s map.

Google’s maps look calmer and simpler because they exemplify a design strategy that Edward Tufte calls “the Smallest Effective Difference”. Here, in Tufte’s words, is the basic idea:

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

The trick here is to “use just noticeable differences, visual elements that make a clear difference but no more—contrasts that are definite, effective, and minimal.” [4] Google, for instance, does this by having its local and arterial roads differ only by color:

The local and arterial roads on Google’s map are nearly identical, except for their coloring: the locals are white, while the arterials are yellow. They look so similar, in fact, that the arterials look like local roads that have been highlighted with a yellow highlighter. The differences are clear, but minimal.

 

In contrast, the differences between the local and arterial roads (and even between the arterials, themselves) are far from minimal on Yahoo’s map:

The colors and widths—and not the “map”—become the main attraction here, and it’s harder to read the map as a result. As Tufte would say, “when everything is emphasized, nothing is emphasized” and that’s certainly the case above: there’s so much that’s emphasized on Yahoo’s map, that the impact of any single element is diluted.

 

Contrast the “loudness” of Yahoo’s map to the calmness of Google’s:

Isn’t it interesting?: The differences between Google’s streets are much subtler than Yahoo’s—and yet it’s just as easy to tell Google’s streets apart.
 

*   *   *
 

One of the main ideas behind the “Smallest Effective Difference” is that the design should disappear in favor of the information. And when you think about it, you don’t really notice the “design” of Google’s streets. (They almost seem “undesigned”, in fact.) Just look at how plain they look:

This is not necessarily the case with Bing’s streets, though:

When you look at Bing’s map, the contrasting street widths are the first things that catch your eye. The differences are so pronounced that it’s almost as if certain streets are shouting at you.

Now that’s not say to that Bing’s map is bad; it’s just that the differences between the streets are so great that the streets visually dominate the image—far more than anything else. Contrast Bing’s “dominant” streets to Google’s comparatively boring ones:

Google’s streets look boring, compared to Bing’s. But this is to Google’s advantage: the “boring” streets allow Google’s map to be layered with additional information, and all without any particular element dominating the image. In other words, the “boring-ness” of Google’s streets actually increases the map’s capacity for other information.

Tufte has something to say about this, as well:

Small contrasts work to enrich the overall visual signal by increasing the number of distinctions that can be made within a single image; thus design by means of small effective differences helps to increase the resolution of our images. [5]

The lesson here is that by using subtle differences, Google is actually increasing the amount of information that can be shown effectively on its maps. Which brings me to my final point…
 

*   *   *
 

“Minimal Differences Allow More Differences” 

One of my very favorite Edward Tufte quotes is that “minimal differences allow more differences.” And this is certainly the case on Google’s maps:

By employing smaller differences (in this case, smaller differences in road widths), Google is able to increase its maps’ detail. Compare the detail shown on Google’s image to that shown on the Yahoo and Bing images.

 

 

NOTES

[1] “Google seems to think so and has boldly reduced the number of road classifications on its maps”… I wasn’t kidding when I said that Google has taken a bold move by featuring only three road classifications on its maps. Almost every consumer-grade map (online or print) features more than three basic road types. Bing, Yahoo, and MapQuest, for instance, all feature multiple arterial classes. And look at how many distinct classifications appear on this legend from a Rand McNally paper map:

[2] “Without a key, it’s unclear what this added information is”… I get the feeling that Google Maps was designed, specifically, to be used without a key or legend (i.e., to be so simple, that nearly anyone could understand the maps). And that to reach this goal, great pains were taken to reduce the maps’ complexity.

[3] “Whatever this added information is, it’s probably not useful or actionable enough to warrant increasing the map’s complexity”… There is one exception here: Yahoo’s green toll roads, which add “actionable” information (i.e., toll road information) to the maps. Then again, it’s not as if Google’s maps are devoid of toll information: for instance, instead of coloring its toll roads green, Google uses text labels to denote toll roads:

And when Google Maps is used for directions, Google indicates if any of the recommended routes have tolls:

[4] Quoted from Edward R. Tufte’s Visual Explanations: Images and Quantities, Evidence and Narrative. 1st Edition. Page 73.

[5] Quoted from Edward R. Tufte’s Visual Explanations: Images and Quantities, Evidence and Narrative. 1st Edition. Page 77.


🏠 Home  ·  ✍️ Essays

Copyright © 2010–2017 Justin O’Beirne