Fifteen years ago this week, Google Maps launched on the web. It wouldn’t be an understatement to say that Google Maps transformed both the way we think about maps and the way we move around in the world. Over time, the application has evolved from a fairly static online representation of an atlas to a GPS-powered navigation tool to a platform for reviews—and, of course, for ads.
Google Maps wasn’t the first online mapping service. The search giant had to make some key acquisitions in the early 2000s in order to get Google Maps off the ground. Over the years it’s faced technical challenges—and some societal ones too—in amassing enough data to ensure its mapping service wasn’t wholly reliant on licensed data. (If your memory is short and you don’t remember the earliest days of Google Maps and Street View cars, consider what Apple’s path has been since it launched its own Maps app in 2012.) But now Google Maps is at the point in some regions where it not only knows where the museum or the train station is, it can show you how to get around inside the building. It shows street-level images and a 3D layer of information on top of it. Perhaps most notably, algorithms are now capable of drawing up maps based on existing data sets.
Jen Fitzpatrick leads the Google Maps team, and also happens to be one of the earliest Googlers. She first joined the company as a software coding intern in 1999, and worked on Google’s search, advertising, and news products before finding her way to Maps. Fitzpatrick sat down with WIRED ahead of the app’s 15th birthday to talk about the evolution of digital maps, how she plans to keep advertising from being as confusing on maps as it is in search, and a future in which Google Maps is more than just a driving app. The interview has been condensed for length and clarity.
Lauren Goode: You’ve worked on a bunch of different products at Google. How did you end up working on Maps?
Jen Fitzpatrick: Well, from search I actually went and spent some time working on our ad products, and then came back to search and worked on everything from shopping to Google News to image search. I didn’t actually start working on the maps and local side of things until 2010. Local was a problem space that I had always been really interested in because I had gotten exposed to it through my work on search, but just had never had the chance to work on it [directly]. So when the chance came along to come over to Maps, it was a problem space that I really loved.
LG: When you say that was a problem space, what do you mean by that? Take people back to the period between 2005 and 2010 that you're describing, when maps are new and there are problems that exist around how people are using them. What was happening that you wanted to help solve?
JF: If you think back to 2005, digital maps existed but they really were very static and basic relative to what you would think about today. This was an era before mobile, so there was no notion of navigation. It was all about literally looking at a 2D cartographic map on a screen. Maybe if you were really lucky you were able to zoom in on that. I think if you think back to the moment when Google Maps came to the scene, the first thing that blew people's minds was that your web browser was all of a sudden able to act like a full-blown app. And you could pan and zoom and scroll, and you could do all these interactive things with the map in the context of a web browser. From a technical breakthrough standpoint that was really exciting, and I think it gave birth to a whole class of other, similar apps over time.
I think the second thing that it did was it really got Google thinking more broadly about a whole new class of information, right? Most of the information that Google had been dealing with up until that point was information that we were crawling from the web. When it comes to something like maps, you can't send a web crawler out to go learn about the real world. There was all this information that didn't really exist in digital form yet. And so we came up on this whole set of challenges around how do you actually find and build and learn and understand information about the real physical world, and bring that into digital form?
LG: What would you say is the source of information that really changed things for maps? Google talks about how it goes out and maps its own data. There’s Street View and other mechanisms for that. There’s contributor data. And then you've acquired data too, with acquisitions of companies like Waze. What’s been the most critical source of information as maps have evolved?
JF: There's not any one single data source or sort of silver bullet for how you map the world. I think a lot the magic lies in how we bring lots of different sources of information together, and combine those to find both the source of truth—because oftentimes data about the real world is noisy—but also to really have the right mix of accurate facts about places or real-time information as the world is changing.
LG: If you had to look back and say there was a moment when all of a sudden doing things with maps became infinitely more possible, what was the thing that really enabled that?
JF: Certainly Street View was huge, in the sense that it allowed us to really give people a sense that they could go visit—you know, “visit” in quotation marks—places that they otherwise might never be able to see. Over time what we've learned is that there's also great information in Street View that helps us understand where buildings are and what addresses are in particular locations. It's turned out to be an even richer source of information than we first recognized.
I think the second big, big piece has been really getting our user community contributing content at scale. You know, if there's one thing that this job has given me a deep appreciation for, it's the fact that the real physical world changes fast, and it’s very big. Those sound like obvious statements, but I think when your job is to map the world you gain a visceral appreciation for just how true both of those things are. And our users are a critical piece of helping us have eyes and ears in places of the world that Street View cars can't go.
And I think the third pivotal shift for us has been, over the last several years, starting to introduce machine learning and lots more automation to how we make the maps themselves. It’s been a major accelerant both in terms of being able to bring lots more information to bear on the map much more quickly, but also to bring maps to many more places of the world where traditionally no good digital map has yet existed.
LG: When you think of this whole era, of tech that was introduced around 2005 and onward, it’s been transformative in a lot of ways, both in terms of how we use tech but also how tech is sort of using us. I’m wondering if in the early days the Maps team would ever get together and say, “OK, let’s envision the ways in which this platform could potentially be used in ways we’re not even thinking about right now.” Or, “What are the ways in which this technology we’re designing could ultimately be used for more nefarious things?”
JF: There are a couple of questions that we always try to think about. One is, what are the things that are possible now that maybe weren’t possible previously? And how do you use that to push the envelope of both the technology and the experience that you can provide? If I look in the more recent past, some of the work we've been doing around augmented reality, and really using the camera in ways that wouldn't have been possible … I think you can see moments like that throughout much of the history of Maps.
And then the second piece is focusing on what's going to be truly useful and helpful to the people who are using it. Whether that's thinking about accessibility and really getting an understanding of what are some of the challenges that might come up in the context of a disabled user, or making sure the information we're displaying on maps is fair and equitable in terms of accessing or covering full populations, not just covering the businesses in the super highly economically developed areas of a particular town. We're doing a lot of work right now in partnership with NGOs on bringing digital addressing to traditionally unaddressed populations which, turns out, is a billion or more people in various places around the world. I mean, just pause for a minute to imagine what it means not to have an address. It means emergency vehicles can't find their way to you. It means in many countries you probably can't get a bank account and you probably can't register to vote.
LG: What about something like misinformation, since people can contribute data to Google and to Google Maps in particular. How do you possibly fact-check it? How do you contain it?
JF: We spend a lot of time thinking about information quality both in the positive sense, of how to make sure the information we put on the map is as high quality and factual and accurate as possible, but also in that sense of how you fight misinformation or spam or abuse. Things that are inevitable anytime you're dealing with content at this scale and contributions at the scale that we are. We have humans who help us review and fact check and that can be anything from, you know, comparing information to ground truth sources of information. Like cross-checking with images that we might have, or in some cases we might call businesses and verify information that way. If you've got something that many people are telling you is true that's likely to be much more trustworthy than something where you've got a single data point. And we’ve definitely got basic algorithmic checking in place to try to detect patterns of bad behavior. We take down lots of content for those reasons.
So it’s not about any single one of those levers, but you put all those tools and techniques together. We know that we're not a hundred percent perfect. I don't think any dataset that's changing and evolving as fast as a map is is ever going to be 100 percent perfect. But we hold ourselves to a super high standard, and I think we're constantly trying to find the places where there might be misinformation creeping in. Our goal is to make sure that you know any low-quality information gets as few eyeballs on it as possible.
LG: How do you ensure that as Google Maps evolves, that the ads on the platform still provide a good experience for users? To me it seems like if someone is searching for something in Google search, it's a little bit more of a lean back experience, where you're sitting and browsing and are looking for information. Hopefully you're able to weed through the top several ad results, even though that's been more confusing recently. But if you're in transit, using Maps, you may be in a more need-it-now situation. So I wonder how you’ll keep that experience not totally cluttered with ads?
JF: We are hyper conscious of making sure that when we introduce ads into the maps experience we're doing in a way that is useful and that is additive to the experience and not a distraction or something that's going to get in your way. I think that's why we've been somewhat slow and deliberate in the ways that we've added ads into the experience. We’re conscious that there are moments where you're you're driving or you're walking or doing other sort of safety-critical things. I think we take that responsibility to be mindful of our users’ attention really seriously.
Having said that, we also know that a lot of what people are increasingly doing in the context of maps are things that are inherently commercial activities, right? [They’re] deciding what businesses to go visit or transact with next. Or looking for a local service business that can come help them with a given task. I could give you more examples, but you get the idea. We do see that there is potential to, in a thoughtful utility-focused way, have ads continue to play in the broader maps experience. But we’re very mindful of doing that in a way that is, like I said, both respectful of users attention and actually adding to the experience.
LG: How do you feel Apple is doing in maps right now?
JF: I don't feel like I'm in the best position to comment on how they're doing. I think we focus on a couple things. One is building a maps experience that works great for everyone regardless of the platform they’re on, so we want to make sure that our map product works well across Android and iOS and any place else that our users are. I can't think of a point in time in history when there's only been one canonical map of the world. So I think it’s not a bad thing to have multiple representations of the world out there. What we try to focus on is, how do we build the best possible map, the best possible experience, and then how do we make sure that that's available for anyone, anywhere, who wants to use it.
LG: Is there anything you would say you think Apple is doing well in maps right now?
JF: You know, I think they're in the process of making some updates to their maps and getting those rolled out more broadly, and I think time will tell how that does.
LG: So, augmented reality. You said a couple of years ago in an interview with Fast Company that you feel that the mental process of walking and navigating is still a problem, that it’s something that needs to be fixed. And you felt as though AR was going to help in some way. How are you feeling about the AR product in Google Maps? Has it done this?
JF: I think we had a hypothesis going into the AR-based walking navigation effort that it would be most helpful for people when they were in very unfamiliar settings, in a travel kind of scenario. And I think we've seen that to be the case. Talking within the team, a lot of us have anecdotal experience of saying, “Oh my gosh, I went to city X or city Y and this really helped me handle those moments of just getting out of the subway or coming out of a hotel and seeing how I get to where I'm going.” But I think if you look at the units or early data that we have, it lines up with that. We're seeing heavy usage for people who are sort of far from home.
I think we're still in early days in terms of AR. We had to do a lot of pretty groundbreaking technical work just to get an experience like that to even work, and to work at the scale that it does. And I also think that experience helps with a portion of the problem, right? We're also thinking about other ways that walking can fundamentally be improved.
LG: What does a day of using Google Maps look like five years in the future?
JF: I can't profess to tell you exactly what the app is going to look like in five years, but I think if I look at some of the themes that we've been pushing on and some of the opportunities that we see as we look ahead … I think on the navigation side, whereas Google Maps covers lots of different modes of transportation today, it's still primarily a driving app. And yet we know that the ways people get around the world are changing and evolving fast. I think you'll see Google Maps evolve to be an application that's much, much better at getting you from here or there no matter what mode of transportation you're using, even if you’re chaining together multiple modes of transportation. That's a clear opportunity.
On the exploration side, we added the Explorer tab to Maps back to 2018 and I think that’s an arc we’re still in the early stages of. There are lots of fun ideas on the horizon around making it easier for people to understand what's around them right now and what's interesting and relevant. And then when it comes to getting things done, we know that many of the reasons people go out and about in the real world is that they're trying to get something done, whether that's meet up with a friend or find a business they want to transact at or a place where they want to do a particular activity. When we think about something like messaging a business in the context of maps, the goal is not to turn Maps into a messaging app. The goal is, if you're in the context of looking up information about a business and you have a quick question, like let's make it super easy to do that without a lot of context switching or apps switching.
LG: I appreciate how you make clear that it's not another messaging app from Google, since Google has had a lot of messaging apps. But another area where Google hasn't had a huge amount of success is in social networking, and what you're describing sounds a little bit to me like a loosely connected social network through the lens of Maps. Will Maps become more social in the future?
JF: We don't think about it as trying to make maps into a social network. I would say far from it. We think about it as trying to help you get the things done that you're trying to get done. So if what you're trying to get done is to share with another person your location so that you can meet up more quickly and easily, you don't need to know your entire social network to solve that specific use. If what you're trying to do is communicate with the business to find out if they're open late tonight or whether they have the product in store that you're trying to find, let's make it super easy for you to do that. So I think it's much more about taking focused use cases and making them work really, really well and really, really easily. It’s less about trying to solve some sort of generic social network type of problem.