Placeholder Image

Subtitles section Play video

  • JONAH JONES: Hey everyone.

  • Thanks for coming.

  • Today we're going to talk about the process of redesigning

  • Google Maps.

  • My name's Jonah.

  • I've been working as a designer at Google for seven years now.

  • ANNETTE LEONG: Hello everyone.

  • My name is Annette.

  • I'm also a designer on Google Maps.

  • I've been at Google for 2 and 1/2 years.

  • In the new Maps I designed the overall framework

  • and image browsing experience.

  • JONAH JONES: And I led the overall redesign.

  • So Google Maps was launched back in 2005.

  • And this time last year, we announced the new Google Maps.

  • And this was the biggest redesign we'd ever done.

  • Over the last year we've made a ton of changes.

  • We had a long preview period.

  • And now this is the new default experience for Maps.

  • So there are some pretty major changes that have taken place.

  • The map is now dynamic and full screen

  • and it adapts to every single one of your clicks.

  • And the imagery is also fully immersive in 3D.

  • But as you can imagine, redesigning a product used

  • by a billion people is quite a daunting task.

  • And we wanted to share a little bit of insight

  • into what our process was, so that you can design

  • your apps to be ready for a billion users as well.

  • And we thought about what this entailed.

  • And we realized that it basically

  • boils down to three key lessons.

  • The first is to think big.

  • So that you don't get trapped incrementalism.

  • The second is to question everything,

  • because sometimes your assumptions

  • can be holding you back.

  • And the third lesson is to listen to your users,

  • because even if you think big and question everything,

  • you're not going to be right about everything.

  • So let's jump into the first lesson, which is to think big.

  • We started with an audacious goal,

  • which is that we wanted to redefine expectations

  • of online maps.

  • And if you look at what we had with the previous maps,

  • we'd added so many features over the years

  • that the old user interface was starting

  • to get really, really cluttered.

  • And every new feature we added was

  • making it harder to find that new feature

  • and the old features too.

  • We'd added thousands of businesses.

  • We'd added thousands of miles of Street View.

  • And we'd also added further ways of transport,

  • so it's not just about driving.

  • It can be about walking or public transit.

  • On a technology side as well, Ajax

  • was the big innovation back in 2005

  • that let people smoothly scroll and zoom the map.

  • But nowadays, there are more recent technologies

  • like HTML5 and WebGL that let people do vector graphics

  • directly within a web browser.

  • And consumer expectations had changed too.

  • The arrival of mobile had brought new paradigms

  • for touch-based user interfaces.

  • But even more fundamentally than that,

  • we felt the existing mapping products were basically

  • just taking this old scanned map and sprinkling

  • a few pins on top and adding a search box at the top.

  • And we thought surely we can do better

  • with dynamic digital maps.

  • So we went back to the drawing board.

  • And we thought how would Maps be if we reinvented it today,

  • using all of the great data that we have as a foundation

  • and everything that we've learned over those last seven

  • years.

  • And we came up with this vision that we

  • called "The Future of Maps."

  • And we created this short video that we

  • shipped around the company.

  • And this was really, really good to get people

  • on the same page about where our vision was,

  • and where we were headed, and where our big thinking was

  • going.

  • And actually it's quite interesting.

  • Looking at this video, there's a lot of the concepts

  • that we had there made it through into the final product.

  • So here you see we have a beautiful new set

  • of rich dynamic bright map tiles.

  • And this is Lausanne in Switzerland.

  • And what happens now is we click on a transit station.

  • And suddenly the map shifts.

  • And dynamically becomes all about the transit network.

  • Later in the video we do a search.

  • And now we can see that the search results

  • are the primary focused thing.

  • And they're labeled too.

  • And then toward the end of the video, we zoom in further.

  • And amongst all of the regular points of interest,

  • we see the places that I've been to, and I've rated.

  • And the places that my friends have been to and rated as well.

  • So this is a map built for you.

  • Over to Annette.

  • ANNETTE LEONG: While the concept video focused on the map tiles,

  • it was a challenge to come up with a framework that

  • could really support and highlight the maps itself.

  • And yet be fully functional and easy to use.

  • We wanted to think big.

  • We wanted to be able to explore quickly and capture

  • all possibilities.

  • And wireframes and sketches were great for that.

  • We made hundreds of wireframes and sketches

  • to generate ideas and to [INAUDIBLE]

  • and also lay out variations.

  • This method let us go broad.

  • Then, from a subset of these wireframes and sketches,

  • we went steep.

  • We started filling in with the right typography, colors,

  • and visual elements to create higher fidelity marks.

  • The look and feel of the product started

  • to really come together.

  • As we generated more and more of these marks,

  • it can be hard to keep track of what's the latest version.

  • It's like an engineer constantly asked you

  • where can I find the latest file?

  • So we needed a better way to organize them.

  • As you may know, when engineers want

  • to show progress of their latest code,

  • they create a build to test how well the code in the biggest

  • system.

  • So we did the same for design.

  • We created a design build, where we put in all the latest

  • designs from individual designers into a single place.

  • To keep things simple and easy to digest,

  • we focused on a few key scenarios.

  • Like finding a restaurant, getting directions,

  • and browsing imagery.

  • This let the designers, PMs, engineers, and everyone

  • in the company to see where we were headed.

  • JONAH JONES: So it's really important to think big.

  • But if you want to think big, you

  • have to be willing to question everything.

  • So you could argue if it ain't broke, don't fix it.

  • This seems fairly reasonable.

  • But it turns out this isn't really

  • the case with technology.

  • As you saw before, adding more and more features

  • meant that we couldn't stay still anymore.

  • But also the world is changing.

  • So by staying still, you're basically moving backward.

  • So I want to give you a couple of examples

  • of where we were questioning things as we were redesigning

  • different parts of the system.

  • And the first is around the map tiles themselves.

  • So this is how the map tiles used to look,

  • when you were looking at a world view.

  • And you can see here that we've got

  • some countries and some states and obviously the land

  • masses and stuff.

  • But we realized that actually this

  • isn't answering all the questions

  • that our users might have.

  • Like why are people going to a map at this zoom level?

  • Well first of all, you can't really

  • differentiate between the cities and the states in the US there.

  • You can't really see where the mountains are.

  • You can't see where the forests are, and where the deserts are.

  • This map isn't really doing everything that it needs to.

  • So when we did the redesign, we answered all these questions.

  • Now you can clearly see the difference

  • between countries and states.

  • We've got elevation data, so that you

  • can see where the mountain ranges are.

  • And we have vegetation data too, so you

  • can see the lush green areas and the more barren orange areas.

  • And plus also, we used a new brighter, more vibrant color

  • palette, so this feels much more on brand

  • than how it did before.

  • The same is true at deeper zoom levels, too.

  • This is the map of how San Francisco used to look.

  • And you can see here our heritage, that basically we

  • just used to be a road map.

  • But that's really all you can see here.

  • But even the roads, they're kind of overwhelming.

  • They're really, really fat.

  • And this orange highlight color and the yellow

  • is being used way too much and not sparingly enough.

  • And you also can't see all of the other information

  • that people care about at this zoom level.

  • You can't see the neighborhoods.

  • You can't see public transit.

  • You can't see where the parks are.

  • So we really want to redesign this, too.

  • So this new design answers all of the questions

  • that people need to know about from a city level too.

  • You can still see the road hierarchy very clearly,

  • but it's cleaner.

  • You can now see the neighborhoods too.

  • You can see the businesses, the points of interest.

  • You can see the transit information.

  • You can see all of this stuff.

  • Now another example is around this concept

  • of a map of a place.

  • So let's imagine that tomorrow after I/O

  • we're all going to go for a picnic.

  • And we're going to meet in the Pullens Gardens.

  • And I share this map with you.

  • Now on first glance, this map looks fairly reasonable.

  • But if you look closer, this really

  • isn't the best map that I could have given you.

  • So for a start, there's a ton of labels on here

  • that you really don't need to know about.

  • And some of the labels you wish were there are missing.

  • So if you look at the Pullens Gardens in the middle,

  • you can see Penton Place bordering it.

  • But you can't see the names of those other two streets.

  • So surely we can do better than this.

  • So we took inspiration from the kind of map

  • that you might draw on the back of a napkin,

  • if you were telling your friend how to get to this place.

  • In that kind of map, people will only

  • show the most important information.

  • You can see here that all we're doing

  • is we're highlighting the major roads that are landmarks

  • and the roads that are nearby and bordering this place.

  • These are the roads that you need to know about

  • to get to this place, and the other ones

  • are much less important.

  • We even developed a pretty snazzy algorithm

  • to approximate this kind of data,

  • where we look at anonymized logs of all of the directions

  • that anyone has ever got to this place.

  • And we look at the most popular ones.

  • And if you're interested on how this works,

  • we get a Tech Talk last year at I/O that's available online,

  • that goes into more depth.

  • But so here we go.

  • We've got these roads.

  • We've got the ones that are important.

  • But now we have to decide how we're

  • going to render these on the map.

  • So the first idea we tried was we

  • thought, OK, let's take these important roads.

  • And we're going to visually lift these up.

  • And make them very prominent.

  • And all of the other roads we're going to let

  • sink back into the background.

  • And this kind of conceptually works.

  • You can see that it's a little bit

  • like that napkin map we had.

  • But there's one really big problem

  • with this, which is that we've lost

  • the sense of road hierarchy.

  • You can now know longer tell the difference

  • between the big major arterial roads and the more

  • minor, local roads.

  • So we had to come up with a solution that didn't interfere

  • with the fundamental road network.

  • So back to the drawing board.

  • We try another version.

  • And we thought what if all of the roads are grey.

  • And we can convey the network using widths instead.

  • And what we'll do is we'll light up these important roads.

  • And this actually looks pretty cool, even on the [INAUDIBLE],

  • you can kind of see the concept there.

  • But what happened when we built this thing is as you're

  • clicking around, the roads changing

  • was just far too visually jarring.

  • It created much too much prominence

  • for what the feature deserved.

  • So, threw that idea away as well.

  • And we came up with this other idea.

  • We thought, well, the way the algorithm works

  • is it's basically taking directions.

  • So why don't we just draw these tiny little arrows,

  • because it's all about getting directions

  • from the north, the south, and the east, and the west.

  • This seems like a cool idea, right?

  • But then we put this in front of people.

  • And this didn't work too well either.

  • We were too close to the algorithm.

  • And it turned out that the users were confused

  • between real sets of directions and these weird little blue

  • arrows we were drawing everywhere.

  • But the good thing about all of these explorations

  • was that we learned something.

  • So that let us finally come up with the end design

  • that we did.

  • This answers all of the problems.

  • Now the most important roads are brighter

  • and they're highlighted.

  • They're labeled as well.

  • You can see that those roads next to Pullens Gardens that

  • weren't labeled before, are-- Crampton Street,

  • Iliffe Street, Amelia Street.

  • And the change is actually quite subtle as well.

  • Because the other roads are still there,

  • but they've just sunk into the background a little bit

  • and the most important roads have lifted slightly,

  • this is not such a dramatic change

  • as you go and click around the map.

  • It subtly draws people's eye, without being

  • a massive headline feature.

  • To give you another example around search results.

  • So here, this is a search for bars in Sydney.

  • And there's something quite interesting

  • going on with this map too.

  • Which is that we have a lot of places

  • that are labeled on the map, but none of these

  • are the places that you've asked for.

  • These are all the generic things that were left on the map

  • from before.

  • Sushi Train, Oxford Vintage Cellars, Darlinghurst.

  • And the very thing that you're interested in-- the bars--

  • they're not labeled.

  • They're the only thing we don't label.

  • This is crazy.

  • So we've just got these red pins.

  • And there's a letter, and you have to cross reference it.

  • And part of this was because we didn't

  • use to be able to dynamically render labels,

  • but now we could.

  • So we thought OK, we can definitely do better than this.

  • So here's what we first tried.

  • We thought, well the search results,

  • they're conceptually the same as a landmark really.

  • They're just a thing that we're going to highlight on the map.

  • So why don't we just draw them the same as a landmark.

  • And we'll just draw a little red circle around them

  • to highlight them.

  • But this had problems, which is that it looked kind of messy

  • as you started zooming out.

  • But even more fundamentally, you couldn't tell the difference

  • between the most relevant search results and the search results.

  • So it could get a bit overwhelming.

  • So how are we going to solve that.

  • We thought OK, what we can do is we can put the star

  • ratings directly on the map.

  • This seems like a good idea.

  • But actually, here you can see this

  • isn't working so well, either.

  • The star ratings and the red dots

  • are kind of competing with each other,

  • and the whole thing looks a bit overwhelming.

  • And it also turns out that star ratings are not the best way

  • to differentiate between places.

  • Because all the top places tend to be about four stars anyway.

  • So what's a more human way of doing this?

  • We thought aha.

  • What we could do is we could take the top places.

  • And we could differentiate them with some human readable text.

  • So you can see here that Esan Thai

  • is known for its pineapple fried rice.

  • But Aroy Thai is known for the Pad Thai.

  • Now this seemed like a strong concept.

  • But even this version was lacking as well.

  • Because typographically, you couldn't really

  • tell the difference between the top and the secondary research

  • results.

  • And also we've lost some of the key information.

  • You can no longer now see what type of business each of these

  • are.

  • We don't have the knife and fork anymore.

  • So with the learnings from all of these different

  • explorations, we finally came up with something

  • that really worked.

  • We used a bright red color to indicate the search results.

  • And now they're conceptually very similar

  • than the landmarks, which are much more set

  • in the background.

  • You can tell the difference between the top

  • and the secondary results through size.

  • And we've still got this very useful text string, which

  • says herb marinated chicken, or whatever.

  • So we put this first in front of users.

  • And we were super excited to see how

  • their reaction was going to be.

  • We expected them to be blown away.

  • Wow.

  • You've labeled search results on the map.

  • This is awesome.

  • No one's ever done this before.

  • And what actually happened was nobody noticed.

  • And we realized that that's actually OK.

  • Because of course the search results are labeled.

  • Weren't you doing that before?

  • People instantly forgot that we didn't used to do that.

  • And it's a reminder that sometimes the best design--

  • people won't even notice it.

  • It will just work and get out the way.

  • So Annette's going to give you a couple

  • more details around place details.

  • ANNETTE LEONG: Similar to the napkin map

  • design and the search results and signposts that Jonah gave,

  • we are still questioned a lot on the place details.

  • In the old maps, when you searched for bars,

  • there's a list of results in the left-hand panel

  • and also on the map.

  • When you click on the result, whether

  • from the left-hand panel or on the map,

  • there was this informational window

  • that appeared on the map.

  • We asked ourselves, why does this informational window

  • have to be that big and cover so much of the map?

  • And secondly, why is the information showing twice?

  • The same information is showing in the left-hand panel

  • and also in the informational window.

  • The more we question the UI, the more

  • we can get a better sense of how we can improve upon it.

  • We tried a version, what if we took away all the UI

  • and had the search box only appear

  • when you click on a place.

  • It didn't really work well, because you expect the search

  • box to be at a persistent location.

  • So you click on the place and it shows up

  • at a different location.

  • It's a style strange.

  • So we ditched the idea and tried another version.

  • What if we want to show more of the map?

  • So we show the text on the transparent background.

  • This also didn't really work well, because as you can see,

  • the text can be hard to read against different backgrounds.

  • And also, the actions that you may need,

  • like getting directions and saving a place,

  • are hidden with another click away.

  • What if we try a version that the text radiated

  • from the place you clicked.

  • Visually it looked cool and somewhat interesting.

  • But it also opened up another can of worms.

  • Like the space limitations and also legibility.

  • And the more we questioned all these explorations,

  • we learned from them.

  • So finally, we landed on this version where the text--

  • the information about the place is always

  • showing in the upper left corner in a predictable way.

  • And this information box can be minimized

  • when you want to see more of the map.

  • Sometimes the problems can be obvious,

  • but the right solution might not be.

  • It takes questioning everything, exploring everything,

  • to get to a much simpler design.

  • JONAH JONES: OK.

  • So the final lesson is that it's really important

  • to listen to your users.

  • And there's a bunch of different ways you can do this.

  • You could run usability studies.

  • You can study your user metrics.

  • You can directly solicit user feedback.

  • Or you can run an internal dogfood.

  • Dogfood is the idea that you eat your own dogfood.

  • So you test the app internally, before you release it

  • into the wild, to the general public.

  • But it's important that you do all of these

  • and don't just focus on one.

  • I'm going to give you a couple of examples

  • of why this is important.

  • So the first is around speed.

  • So when we first launched the new Maps,

  • we heard some users were giving us

  • feedback that we needed to make it faster.

  • And we thought, this is kind of interesting.

  • Because when we looked at our logs,

  • average speed was-- the new Maps was faster than the old Maps.

  • But it turns out that wasn't the whole picture.

  • It turns out the for people with fast computers

  • and fast network connections, the new Maps

  • was indeed significantly faster.

  • For people with average computers

  • and average connections, the new Maps was a bit faster or about

  • the same.

  • But for people with slow computers or slow internet

  • connections, the new Maps was slower.

  • That's why people were saying this.

  • This is a problem.

  • They were kind of seeing something like this.

  • They were seeing this extended loading screen.

  • And they were waiting for that to load.

  • Oh, man.

  • This is so slow, come on, come on, come on.

  • OK, it's loaded.

  • And so we sent our crack team of engineers on this.

  • And they worked night and day.

  • And they managed to get the loading speed

  • down to half of what it used to be.

  • And this was pretty amazing, but it's still not fast enough.

  • So this needed a bit of a more creative solution.

  • And we realized that people tend to do one of two things

  • when they first load up Google Maps.

  • Number one is that they just kind of stare

  • at the map tiles for one or two or three seconds.

  • Just to get a sense of orientation.

  • And number two was if they're not

  • doing that, they tend to go to the search box

  • and just start typing.

  • So we thought, what if we could be smarter.

  • What if-- is this the new loading

  • sequence or the old one.

  • It's the new one.

  • OK.

  • What if we could load the map tiles first, in a static way.

  • So they're not the normal 3D ones.

  • These are just a quick set of map tiles

  • that we can load immediately.

  • This gives all those people who just

  • want to look at the map something

  • to do much, much quicker.

  • And what if we could enable the search box,

  • so that people can type into that.

  • Now, in a much, much quicker amount of time,

  • we're showing the map and we're letting people type

  • into the search box.

  • And by the time they're done doing that stuff,

  • the rest of the map is loaded.

  • So this is a sign that real speed is very, very important.

  • But the perceived speed can be very important as well.

  • The number of people who are dissatisfied with the speed

  • went down dramatically after we did this change.

  • There's a second example too.

  • And this is what happens when you click on a place,

  • and then we load this new map that's

  • built for this place with the new roads.

  • And there's actually quite a few things

  • going on here that you'll notice.

  • So the first thing is that when we click,

  • we have to wait for 200 milliseconds

  • to differentiate between if this is a click, which

  • is to select something, or a double click to zoom in.

  • And there's nothing, really, you can do about this.

  • But if you just wait, it will feel slow.

  • So what we do is we immediately do a little ripple event that

  • responds to the first click.

  • And after that 200 milliseconds time out,

  • that's when we begin to lift the pen up in the air.

  • So this now feels immediately much more responsive.

  • The second thing that happens is that the pin jumps up

  • into the air and then lands.

  • And as it lands, the map responds and updates.

  • And this is done because as soon as we know that there's

  • been a click, we have to go back to the server,

  • figure out what all these roads are that we need to highlight,

  • come back again to the client, and then re-render

  • these new map tiles.

  • And had we waited to do that, before showing the pin,

  • that would feel laggy.

  • But lad we shown the pin immediately,

  • and then the map tiles had updated sometime later,

  • that would also feel laggy.

  • So this solution lets us begin the animation.

  • Everyone's watching that, they're distracted,

  • and then it lands.

  • And then the map tiles update.

  • And this not only hides the loading,

  • but it also adds personality and character to the map.

  • So we managed to kill two birds with one stone here.

  • Annette's going to give you some more

  • examples around Street View.

  • ANNETTE LEONG: Imagery was one of the major features

  • in the redesign.

  • We wanted to make discovering images much more easier.

  • In the old maps, there was this to little yellow guy

  • that we called Pegman, where you have to click the Pegman

  • and jerk him to the map, in order to see Street View.

  • And some people did not discover that.

  • So in the redesign, we wanted to put the images up front.

  • So we replaced the Pegman with a sequence of images

  • at the bottom.

  • We tested this design in our internal dogfood.

  • We were rather surprised that one of the comments we got

  • was that, where did you put Street View.

  • So in this design, we thought that we

  • had a clean and organized way to organize everything.

  • Information is at the top, and the images are at the bottom.

  • But that wasn't what users saw.

  • When the users saw the information at the top,

  • they were questioning where was Street View.

  • And they saw the images at the bottom,

  • and they questioned what those were.

  • In order to match the user's mental model better,

  • we changed our design.

  • We grouped the information and images together in one place.

  • And users can still see images about the neighborhood

  • by opening the imagery cursor at the bottom.

  • Finally, people are able to get into Street View more

  • intuitively.

  • Problem solved.

  • But there was another problem.

  • In our usability study, we noticed that once

  • people got into Street View they actually

  • did not move around much.

  • They actually were missing out on all the fun.

  • What a shame.

  • How could we fix it?

  • So we looked again, we looked into our old UI.

  • There was this circular highlight

  • that followed the cursor, supposed

  • to indicate that you could click and move.

  • But some people did not get that.

  • We tried a bunch of designs, and finally we

  • made two small but important changes.

  • One, we changed the cursor from an arrow

  • to a hand with a pointing finger.

  • Now people understand that they can click to proceed.

  • Secondly, we added an arrow inside the circle highlight.

  • Now people understand that when they click

  • they will also moved toward that direction.

  • This all leads to some really great news.

  • Our Street View usage was up three times,

  • according to our metrics.

  • In the meantime, something strange was happening.

  • People were still asking to bring back Pegman.

  • We were wondering why.

  • Our Street View usage was up three times.

  • Why are people still asking to bring Pegman?

  • So we looked into it further.

  • Turned out there's a group of users like Anita,

  • who thinks of a place first.

  • So she would search and see Street View, she is happy.

  • This explains why the new design was

  • successful to help Anita to discover images much easier.

  • Which also explains why our usage was up three times.

  • There's another group of users like Susannah,

  • who thinks of Street View first.

  • So she did not search.

  • And she's not seeing any images.

  • Nor the Pegman like in the old maps.

  • She is stuck.

  • So we decided to bring Pegman back.

  • And now Pegman is happier, because he's living closer

  • to the imagery carousel.

  • And our users like Anita and Susannah are happier too.

  • Sometimes when you design a product,

  • it may not be matching how your users use your product.

  • You may have to listen carefully and observe carefully, in order

  • to improve your design further.

  • Finally, it's really important to close the feedback loop.

  • During our preview period, we received tons of feedback

  • from our users.

  • As soon as we fixed an issue that users requested

  • or suggested, we wrote an email to thank them and let

  • them know that their issues have been fixed.

  • And people love this.

  • 98% of our users who received our email

  • would like to continue to receive communication

  • like this.

  • It really showed that to your users that you listened,

  • that you care, and that they have been heard.

  • JONAH JONES: OK.

  • So back to the original question.

  • How can you design your apps for a billion people?

  • You think big, you question everything,

  • and you listen to your users.

  • Thank you very much.

  • [APPLAUSE]

  • ANNETTE LEONG: Thank you.

  • JONAH JONES: And we've got 17 minutes left for questions.

  • So if you've got any questions, there's a couple of mics there.

  • And we'll be very happy to answer them.

  • AUDIENCE: Hello.

  • Thank you for the talk.

  • And also thank you for redesigning Google Maps.

  • I think it's fantastic.

  • I had exactly the Pegman issue when Google Maps was resigned.

  • As a UX guy, I was like, it felt like a real misstep

  • for the design.

  • And when I eventually found him down in the bottom corner

  • there, I was a little bit confused

  • as to why you guys moved like from a purely UI design

  • perspective.

  • Why he's down there with the settings and also the tour.

  • But there are also some options at the top right

  • and you can also do stuff in the top left corner.

  • And there's also Street View in the bottom left corner as well.

  • It just seemed to me to not follow a hierarchy,

  • or like a functional hierarchy.

  • So can you talk a bit about that?

  • JONAH JONES: Sure.

  • I mean the original thinking behind taking away Pegman

  • actually was if you're going to redesign a product like Google

  • Maps, there's like thousands of features.

  • And we had to be really, really bold.

  • And we wanted to see how much we could take away and get away

  • with.

  • And like 90% of the stuff we took away,

  • we got away with, which is good.

  • So you managed to simplify, but we always

  • knew there was going to be somethings

  • that we had to bring back.

  • And the Pegman was one of them.

  • And it was because of these two mental models that people had.

  • A bit of insight is that conceptually we'd

  • always thought of Pegman as the deepest zoom of Google Maps.

  • And there used to be this behavior that you'd

  • zoom in, zoom in, zoom in, zoom in,

  • and then suddenly you'd hit Street View.

  • And that was why Pegman used to be

  • on top of the zoom slider over there.

  • But we would constantly see people zooming in and then

  • accidentally getting into Street View, and like--

  • and then coming back out again.

  • So it turns out that it didn't really

  • make sense, conceptually, to our users to have it that way.

  • And that's why when we brought the Pegman back,

  • we brought him back in the imagery area

  • down on the bottom right.

  • Because that's where you summon all of the other imagery.

  • If you pull that up, that's where the Street Views live,

  • that's where the panorama photos live and everything else.

  • So it felt like that was Pegman's natural home,

  • even though it was a different place than where

  • he'd originally been living.

  • So that's why we put him there.

  • AUDIENCE: Hi.

  • When do you make the decision to do such a large redesign?

  • I mean for applications with so many users and things

  • like that, at what point do you stop saying, oh we

  • should do that.

  • And say all right, we're starting now.

  • JONAH JONES: We'd kind of reached-

  • I alluded to this at the beginning,

  • but we kind of reached the stage where most people kind of felt

  • like they were happy with Google Maps.

  • And that was fine.

  • But we realized that first of all, there

  • was a big opportunity cost.

  • Because it wasn't doing all these things

  • we wanted it to do.

  • But second of all, it was kind of running out of space,

  • adding all these new features.

  • And we were adding features.

  • And all the discussions were always around discoverability,

  • because everyone wants their feature to be big and bold.

  • But it means all the other ones can't be discovered.

  • And so we were basically running out of space

  • to fit new features.

  • So we wanted to rethink it and simplify it and also

  • build a frameworks to let us add more and more new stuff

  • in a contextual way.

  • So it was kind of like 50% necessity, just

  • looking into the future of where we were going.

  • And 50% that we wanted to try something new.

  • And as we mentioned before, just kind

  • of redefine expectations of what Maps would be, rather than just

  • keep plodding away and adding features to the same thing.

  • AUDIENCE: Hi.

  • So I notice now when you search for cities,

  • you get a nice city border.

  • But when you zoom in enough to figure out whether something's

  • in one city or the other, it disappears.

  • And so there's this frustration of zooming in enough

  • to see what you're looking for, and then you lose your border.

  • I wonder if you could explain the rationale for why

  • those disappear.

  • JONAH JONES: We actually have quite a lot

  • of heated arguments about this internally as well.

  • Because it turns out that when people are searching for a city

  • or anything, like a neighborhood or a zip code

  • or any of this stuff, you might want different things.

  • Like some people are really interested in the exact border

  • of the city.

  • In which case maybe we should really highlight that,

  • get rid of everything else.

  • Just show all these stats about the city and everything.

  • But then some other people, maybe they just

  • want to send to the viewport there.

  • Then in that case, you don't want

  • to draw too much visual attention to it.

  • And you want to show the context too.

  • So we try to achieve a balance between highlighting the thing

  • that you've searched for, but not in such a way

  • that it's at the detriment of anything else.

  • So it's just about trying to balance what everybody wants.

  • And you'll probably see this has changed in the past,

  • you'll probably see it change again,

  • as we're trying to tweak this to best answer

  • all of our users needs.

  • AUDIENCE: Hi.

  • I'm wondering if you have plans to include

  • the embed code somewhere in the new UI?

  • Because I can't find it.

  • And I always have to switch back to Classic Maps to get it.

  • JONAH JONES: I believe the embed code is down in the--

  • ANNETTE LEONG: Yeah, so it is actually in the Gear menu.

  • It's not the most intuitive way at all.

  • We're still looking into finding a better place for that.

  • If you click into the Gear menu, you

  • should be able to see share and embed map.

  • AUDIENCE: Thank you.

  • AUDIENCE: Hi.

  • I have a question about 3D buildings

  • and mapping and the 3D new imagery system.

  • You've just abrogated it last year.

  • And just-- there were users who were able to build and upload

  • by those [INAUDIBLE] using SketchUp, using Building Maker.

  • And it seems maybe two or three months ago, they

  • were not able to just upload everything, because it

  • was canceled.

  • And there were more than 10,000 modelers.

  • They had made millions of buildings since you started.

  • And why did you decide to cancel this project,

  • and what's the main reason?

  • JONAH JONES: So I'm probably not the best person

  • to talk about this.

  • But I can give you like a more of a design perspective.

  • My understanding on this basically

  • is that we try to kind of automatically

  • generate a lot of the 3D buildings we have

  • as we've been flying over cities and stuff and building them up.

  • And we do have some amount of buildings

  • that have been specifically hand-modeled.

  • But I'm not 100% sure on the exact policies

  • of what gets included and what not.

  • So if you follow out with me afterwards,

  • we can follow up with [INAUDIBLE].

  • And he'll be able to answer those questions.

  • So come and chat to us afterwards

  • and I'll give you more detail.

  • Thanks.

  • AUDIENCE: Hi.

  • I was just curious.

  • To what degree was the redesign informed

  • by the needs of low-end mobile users versus desktop users?

  • JONAH JONES: So that's a really interesting question.

  • So the vision of the future of Google Maps that we talked

  • about was very, very platform agnostic.

  • It was about this thing that's personalized to you.

  • It's showing the data that you care about.

  • It can dynamically change for every single place.

  • And we wanted to achieve this on all platforms.

  • And what you saw was that I guess the first hint of that

  • was with the new iOS app that we built with the new design

  • language.

  • And then a lot of the new functionality and concepts

  • came into the desktop version.

  • And now they're starting to find their way back

  • into the mobile version as well.

  • So all of the platforms are kind of-- they're

  • almost like in a race.

  • Some are readier earlier than others.

  • But we're all heading toward this unified future.

  • And low-end devices are going to be very much

  • an important part of that as well.

  • It's not something that we've been focusing heavily on

  • to begin with, but definitely we're

  • going to be focusing very much on speed and devices

  • with low connectivity and low CPU and that kind of thing.

  • It's really important to get those things working,

  • especially with the Android 1 phones coming out

  • and everything.

  • Actually, I haven't tried one of those yet,

  • but I'm really curious to see how it works.

  • AUDIENCE: More generally, about the process.

  • When you do decide to do a redesign,

  • the three lessons are sort of like three steps.

  • Is there a particular order that you guys followed?

  • Did you listen to users first, or did you

  • try to think big on your own?

  • Or is it all at the same time?

  • JONAH JONES: It's kind of all at the same time, actually.

  • Because the benefit we had of working on this product

  • for seven years, is that we've run repeated usability studies.

  • We're getting feedback all the time.

  • So we know what our users' pain points are.

  • And we know what areas of the products

  • are being underutilized.

  • So we knew, for example, that we had all this great imagery,

  • and people weren't finding it.

  • So some of them are explicit issues.

  • And other things are just opportunity areas.

  • So we kind of just have this idea of oh man,

  • this could be so much better.

  • So that was just informing things.

  • And then obviously, if you want to change something and do

  • something audacious, you've got to question everything.

  • You've got to think big.

  • It's all kind of part and parcel.

  • It's not really a step one, two, three thing.

  • ANNETTE LEONG: Similarly, like when you build a product,

  • you want to focus on a few key use cases.

  • And then constantly, you're looking back

  • at what you designed to see if you are matching those use

  • cases.

  • And each time you are not, then you

  • go back to the sketching board, or talk to the researchers

  • how we can improve the products further.

  • AUDIENCE: Hi.

  • I was hoping you could share a little more about how

  • you conduct the user research.

  • Both before, during, and after this process

  • for around the world.

  • It seems like a pretty difficult problem.

  • I'm wondering what the main steps were.

  • JONAH JONES: I wish we had a researcher with us

  • to talk about this.

  • But we basically run a variety of different research studies.

  • We can do super small scale stuff, even within the company,

  • where you just go into a cafe, and you get someone

  • to click on stuff.

  • We can bring people in and have them test things out.

  • We can go to people's homes and see

  • how they use things in the real life.

  • There's diary studies, where you can

  • get people to document the things that they need

  • every day, and what they're using.

  • And that can be across different products.

  • Obviously, there's the feedback that you get from users

  • when you get to send feedback from within the product.

  • You can look at the metrics.

  • You can see what's getting clicked and what isn't.

  • So we kind of used like a dozen or so different ways

  • of getting this, because they're all

  • flawed in their own different ways,

  • so you're trying to get around it.

  • AUDIENCE: And is it something that if you look at that data

  • and say it's not really making sense

  • for a particular region of the world.

  • You'll go out there and really attentively study that.

  • JONAH JONES: You can certainly notice

  • that there are regional differences in how people use

  • things, the kind of things that they search for,

  • the kind of things that they need.

  • And of course the data is very different as well.

  • The data in the Bay Area is great.

  • You go somewhere else, maybe the data is a lot lower quality.

  • So things vary a lot.

  • It's really interesting.

  • ANNETTE LEONG: And sometimes when you design a feature,

  • you may need multiple rounds of research and usability testing

  • for that.

  • And then we had a long period of time with internal dogfood,

  • where we collected a lot of feedback from internal users.

  • So we actually during that round we

  • made the feedback button super prominent.

  • And people wherever they use and run into any problem,

  • they can just click there and that

  • would allow them to easily file a request or bug.

  • JONAH JONES: It's also worth pointing out

  • that we're a very diverse team.

  • Like you can hear that neither of us are American here.

  • We have teams all over the world.

  • And so I think that actually helps

  • in bringing a very global perspective

  • in building a product like this.

  • That we're not all just focused on one place.

  • We've got people from different point of views.

  • That helps too.

  • AUDIENCE: That was my additional question.

  • Sorry for interrupting.

  • But I wanted to know how localized the maps are.

  • And how different they are in Asian countries, or let's

  • say in the Middle East.

  • JONAH JONES: So we try and achieve

  • a balance between localization where it makes sense,

  • because we want people to feel familiar.

  • But we also want Google Maps to be a consistent experience,

  • so that you recognize it as you travel around.

  • So you'll see some elements of the design,

  • like we have localized public transit icons, for example.

  • But the things like-- what else was a good example-- things

  • like restaurant icons or something

  • like that are kind of the same everywhere.

  • So we try and achieve a balance of localization

  • where it makes sense to be different.

  • But otherwise we try and have a consistent style

  • so that everyone can predict how to use the product.

  • ANNETTE LEONG: And in terms of the UI,

  • we also created the framework that

  • supports multiple languages.

  • Like some language like you have to read from right to left.

  • So we had to make sure that the framework could

  • allow a format like that.

  • AUDIENCE: I was wondering if you could talk more

  • about the differences between the mobile app versus the web

  • UI, which you were showcasing more in this talk.

  • In particular, I run a UX design agency in our office.

  • We were talking about how much we really

  • enjoy the updates in the web app.

  • And in the mobile UI, we kept getting

  • lost in how you even start a navigation session.

  • And it looked like something's transferred well,

  • and other things didn't.

  • And maybe somethings you're working on.

  • So I was just wondering if you could shed more light on that.

  • JONAH JONES: I would characterize that

  • as saying that we have this vision of the future.

  • And then the mobile and the desktop versions

  • were being developed concurrently.

  • And so they're almost like slightly different

  • interpretations of that same future.

  • So the teams are sitting right next to each other.

  • So the desktop team were borrowing ideas

  • from the mobile team and vice versa.

  • So it's mostly the same.

  • There are some things that are deliberately different.

  • There are a few things that are kind of unintentionally

  • different, that we want to fix.

  • But I think gradually you're going

  • to see these things converging increasingly.

  • And especially with the new material design staff.

  • We really want to have an experience where

  • it's very, very consistent as you go from desktop to tablet

  • to mobile to watch to Glass, whatever.

  • AUDIENCE: So for new features like highlighting roads

  • next to points of interest, are there

  • any plans to make those accessible via the API

  • or any of the new features that you guys talked about today?

  • JONAH JONES: That's a really good question.

  • They do show up on the API embed.

  • Follow up with us afterwards.

  • We've got the API guy here so we can find out more about that.

  • I believe that's everything.

  • So thanks everyone for your time.

  • ANNETTE LEONG: Thank you.

  • [APPLAUSE]

JONAH JONES: Hey everyone.

Subtitles and vocabulary

Click the word to look it up Click the word to find further inforamtion about it