Justin O'Beirne

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


CARTOGRAPHY COMPARISON INTERLUDE
GOOGLE MAPS’S REDESIGN

Halfway through our Comparison, Google has refreshed its cartography:

What’s changed? And how many of our earlier observations are still true?

In this Interlude from our Comparison, we’ll take a look at Google Maps’s recent changes.

 

*   *   *

INTERLUDE
CONTENTS

Boiling Frogs
   - “Mylar”
   - “V******e”
   - “Tactile”
What’s Changed?
   - Perfect Timing
   - Content Changes
   - A Tour of San Francisco
Next Chapter (Available Soon)

Interlude Summary

 

⚠️ As with our earlier chapters, this is an in-depth exploration.

TL;DR: Halfway through our Comparison, we have a surprising new lineup — and a peek at the future. (To skip to the Interlude Summary, click here.)

 

 

*   *   *

 

⚠️ Tap or click any image to enlarge



BOILING FROGS

A Google Maps executive once told me that changing cartography is like boiling a frog. Change too much, too soon — and users might react or even revolt. But change things slowly over time, and they won’t even notice. The map will feel as though it’s improving, but no one will quite know why—just that it’s getting better and better.

Looking back over the decade so far, there seems to be some truth to this. 

Three of the biggest cartography changes were Bing Maps in 2010, Apple Maps in 2012, and Google Maps in 2013. And in two of those — Bing and Apple — users revolted.

What made Google different?

In other words, why did users revolt against Bing and Apple, but not against Google?

 

 

*   *   *



“Mylar” | Bing’s 2010 Cartography Change

Design has to work; art does not.
— DONALD JUDD

Back in the Summer of 2010, Bing Maps was losing ground to Google Maps. Aiming to attract users to its struggling mapping platform (and keep developers from migrating away), Microsoft needed a bold move — something that would cause people to take a second look, something that wouldn’t go unnoticed.

So it teamed up with Stamen, a San Francisco design studio known for artful and experimental cartography…

…And together, Microsoft and Stamen crafted a striking new vision of online cartography — something sure to make a statement.

But when it was released late one August afternoon, it was a statement that few outside of Microsoft understood. And instead of winning new users, it ended up alienating many of its existing ones.

Fewer labels. Less contrast. Harder to read. To many of Bing’s users, the cartography had regressed.

Now that’s not to say that the new map wasn’t beautiful. But it was a beauty rooted in subtlety — and hence, the problem: If a map is a way of communicating, should that communication be subtle? Or should it be obvious and clear?

In the end, Bing may have had the most striking map — but it didn’t have the clearest…

…and those who were already using Google Maps had little reason to switch away.

But the frustration over Bing’s new cartography would pale in comparison to the outrage that Microsoft’s old arch-rival would face two summers later…



“V∙∙∙∙∙∙e” | Apple’s 2012 Cartography Change

Design cannot rescue failed content.
— EDWARD TUFTE

Two summers after Bing revamped its cartography, Apple sought to make a cartographic splash of its own. It was to be “the most beautiful, powerful mapping service ever” — a design so good, iPhone users would forget Google Maps had even existed.

But no matter how much prettier Apple’s new map was compared to what it replaced, it couldn’t erase the fact that its content was orders of magnitude worse and often comically inaccurate.

David Pogue of The New York Times likened Apple’s new Maps app to a $1,500 professional coffee maker that had had moldy beans poured into it, calling it “an appalling first release… the most embarrassing, least usable piece of software Apple has ever unleashed”. Wired, meanwhile, crowned it 2012's biggest tech fail.

Whereas Bing had regressed the map’s usability, Apple had regressed the map’s accuracy. And users were furious.

But what made Bing’s and Apple’s changes even more infuriating was their sudden rollout. Bing had changed overnight, with no prior warning or announcement, surprising its users. And though Apple was careful not to repeat Bing’s mistake—it had announced its map three months ahead of time—the vast majority of Apple’s 300 million users had never actually used it prior to its launch. (The only way to preview Apple’s new map was to sign up for a $99 developer’s license and install buggy pre-release software onto an iPhone.)

Making matters worse, neither Bing nor Apple left users any way to revert to the old map.

So when Bing and Apple suddenly flipped the switch, their customers felt as if a rug had been pulled from underneath them. And a firestorm ensued. Things eventually got so bad for Apple that its CEO was forced to write a public apology that recommended competing mapping products.

It’s not hard to imagine the Google Maps team watching the Apple Maps revolt in horror, as they were secretly preparing a cartography update of their own — an update that was to be the largest and most ambitious in Google Maps’s history…



“Tactile” | Google’s 2013 Cartography Change

Less than a year after the Apple Maps debacle, Google announced a dramatically redesigned map of its own:

But it released its new map a bit differently than Apple or Bing.

As soon as it was announced, Google allowed anyone interested to request an invitation to a “beta” version of the new map—a brilliant move buying Google three things:

(1) It kept the new map’s initial user-base small and contained.

(2) Because the new map could only be used while signed in, it allowed Google to carefully monitor and study each invitee’s behavior — and make behavior-driven improvements.

(3) Designating the new map as a “beta” allowed Google to manage user expectations; i.e., “beta” = “expect bugs”.

As expected, Google’s most passionate and vocal users (i.e., those most likely to ignite an online firestorm if displeased) requested invitations — as did those who couldn’t resist the exclusivity of invite-only access to a new Google product.

Google then went on to spend more than ten months collecting feedback, fixing bugs, and tweaking and re-tweaking the map before finally releasing it to the rest of its user base (which was nearing one billion users in those days).

But here was another stroke of genius: instead of releasing its map to everyone on the same day, it carefully rolled it out to different segments of users over a three week period.

This was in stark contrast to Bing and Apple, who had released their maps on a single day — and in doing so, created instant mobs out of their user-bases.

But in having twenty-one release days (instead of just one), Google effectively compartmentalized its user-base, guaranteeing that any mob that formed would be small and that any fires that broke out would quickly be contained. There would be no massive pile on as there was in the wake of Apple’s release. And if anything started to go bad during the first few days, Google could simply switch off the spigot.

But perhaps the most important thing that Google did differently was that it gave its users an escape hatch—an easy way to revert back to the old version anytime they wanted, no questions asked. In other words, no one was forced to use Google’s new map. And it made all the difference.

 

*   *   *

 

Bing and Apple had thrown their users into scalding water—water so hot, that their users jumped right back out. But Google had boiled them slowly, so slowly that they never even noticed. And in comparison to Bing and Apple, Google’s release was anti-climactic. There was no revolt. No outcry. No uproar. It was one of the most graceful cartography releases in history—a release as artful as the map itself.

Boiling frogs, indeed.

 

*   *   *

 

Looking back over Google Maps’s history, the cautious approach we saw in 2013 seems to have been Google’s modus operandi all along.

Though Google Maps looks completely different today than it did at its launch in 2005…

…it has gotten that way by evolving slowly and gradually:

The maps above and below show a series of small, iterative enhancements — never anything too big or surprising.

Just a slow, but steady progression.

So if history is any guide, the map has likely changed very little with Google’s most recent iteration—with minimal impact to our Comparison.

But that’s not going to stop us from taking a look anyway…

 

 

*   *   *



WHAT’S CHANGED?

Perfect Timing

We couldn’t have started our Comparison at a better time.

When Google rolled out its most recent changes, it overwrote its old map, leaving us no way to see what it used to look like.

But thanks to our Comparison, we have an ace in the hole: a detailed record of how the map used to look at each zoom in New York, San Francisco, and London:

We’ll use the maps from our earlier investigations to find everything Google changed.

See what I mean about our perfect timing? 😉

 

 

*   *   *



Content Changes

We were halfway through our Comparison when Google rolled out its changes. So let’s backtrack and see if Google changed anything we’ve already looked at.

So far we’ve compared the number of things on Google Maps and Apple Maps — things like CitiesRoads, and Places:

So let’s start there.

Are the same things on Google Maps as before? Or are there different things now?

And have the amounts changed?

(We won’t pay attention to how anything looks just yet; we’ll just see if the same things are on the map.)

Let’s take one of our three cities and do a quick sweep across each zoom. Since our Comparison has had more readers in San Francisco than any other city, we’ll use San Francisco:



A Tour of San Francisco

📷 San Francisco | Photo by Justin O’Beirne

 

Let’s start with our earliest San Francisco zoom, z2:

Here at z2, the changes are subtle: the water color is slightly different now and the shadow along the coastline has been removed. But apart from those small changes, we see all the same things as before: all the same countries and oceans are labeled and all the same lines are drawn.

Let’s go to z3… 

Here again, we see all the same things as before: all the same lines and all the same labels.

Does the pattern continue on z4?

Just as we saw on z2 and z3, z4 also has all the same lines and labels as before.

In fact, if we look at the entire first half of our San Francisco zooms (z2 — z10)…

…we see the exact same things as before:

As the images show, not a single label has changed across our first nine zooms.

And it isn’t until we get to z11 that we see any change at all...

At z11, our map is almost the same as before — except for a small change at the bottom. Do you see it?

San Bruno Mountain State Park has been removed and two cities have been added in its place (Colma and Brisbane):

San Bruno Mountain State Park is still there in real life—it has just disappeared from Google’s map (along with several other parks). 🤔

In any event, the label changes we see at z11 mark the start of a trend—and as we continue across our remaining eight zooms (z12 through z19), we see a number of other label changes:

As the images above show, all the zooms from z11 onward seem to have some sort of label change (unlike the first half of our zooms, which had no changes at all):

So with all of the label changes, how different is the map now? And what’s changed?

Let’s count the number of label changes on each zoom and plot them in a graph:

Across all of our San Francisco zooms, 28 new labels have been added to the map, while 38 other labels have been removed—giving us a total of 66 label changes across all our zooms.

But our graph above is deceptive: even though there have been 66 changes, all but five of our zooms have the same number of labels as before:

Surprising, isn’t it?

And adding all eighteen zooms together, the new map has just ten fewer labels than before:

Ten fewer labels? That doesn’t seem like a big change—especially with those labels spread across eighteen zooms. And it means that each zoom has an average of just 0.6 fewer labels than before:

0.6 fewer labels per zoom? It doesn’t seem like the map has changed at all, in spite of those 66 changes.

So what’s really going on?

And what kinds of labels have changed?

As we did in Part 2, let’s divide all of the map’s labels into seven categories:

Now let’s plot all of the labels that have been added and removed—but this time, instead of plotting them by zoom, let’s plot them by category:

Our graph shows us that only four kinds of labels have had any changes: CitiesCity Sub-AreasRoads, and Places.

Each of these four categories has had labels added or removed, and there are now two more Cities, eight fewer City Sub-Areas, and four fewer Roads labeled on the map:

But these changes are small in the grand scheme of things: there's only a 2% decrease in Road labels (from 209 to 205) and a 3% increase in City labels (from 66 to 68). And spread out across eighteen zoom-levels, neither of these changes seems significant—or even noticeable to the average user.

But the story is a little different with City Sub-Areas, where there’s been a 22% decrease in labels (29 now versus 37 before) . And the decrease is most noticeable on the 13th and 14th zoom-levels:

The City Sub-Area labels are noticeably larger on the new map. Perhaps this is why there are fewer? (I.e., because they’re larger, perhaps fewer can fit?) But if the goal was to elevate their importance, Google seems to have accomplished the opposite in removing nearly a quarter of them.

Yet apart from this reduction in City Sub-Areas, the map’s overall character is almost identical to before:

What seems to have happened is that most of the 66 additions and subtractions we saw earlier have cancelled each other out. And apart from the 22% reduction in City Sub-Areas, the map’s profile is largely the same as before.

In fact, even when we tally all of the map’s different Places by category, we see the same general profile as before:

So from a content perspective, there’s been little change as to what’s on the map: the same kinds of things are on the map and in nearly the same amounts. And this also means that all of our earlier observations from our Comparison remain true.

 

*   *   *

 

It’s surprising that so little has changed. But then again, this is what we saw earlier: Google always makes slow, gradual changes—never hastily changing anything.

But even though the content is the same, the map clearly looks different now. And it looks as if there’s something new on it.

Take another look at the zooms below. Do you see those faint, copper-colored shapes on the new map?

A New Thing has been added…

 

TO BE CONTINUED

 


GET NOTIFIED WHEN NEW CHAPTERS ARE PUBLISHED

📡 RSS

🐣 TWITTER
Similar to RSS, but for Twitter.


🏠 Home  ·  ✍️ Essays

Copyright © 2010–2017 Justin O’Beirne