How a Google side project evolved into a $4B company

Niantic EC-1 Part 1: The Origin Story

How did Niantic happen? How did the company behind Pokémon GO and (soon) Harry Potter: Wizards Unite come to be?

When anyone talks about Niantic, they generally mention that it’s “a Google spinout” and move on. As if that’s something that just happens every day. That dozens of people within a massive company come together, build something… and then just leave and take their work with them to form a new, independent company that goes on to have a valuation of nearly 4 billion dollars.

So… how?

Over the last few weeks, I’ve interviewed dozens of people involved with Niantic’s story so far, including investors, executives, and employees past and present. I wanted to figure out the hows and whys of Niantic’s origins, what others might be able to learn from the company’s story so far, and where the company is going in the future.

The reading time for this article is 18 minutes (4,400 words)

The Keyhole into Google

“I started at Google with this idea that I’d be there for six months,” Niantic CEO John Hanke tells me.

We’re in a conference room at Niantic’s office, which takes up much of the second story of San Francisco’s Ferry Building. John’s wearing what I’ve come to realize is something of a daily signature for him: a T-shirt (often Niantic branded) beneath an unbuttoned blue dress shirt. His hair swoops forward and hangs just above his eyes. He’s laid back, but his words are very deliberate and still ring with the slightest hint of his Central Texas hometown.

“The whole time I was there,” he continues, “it felt like I’d be there for another six months. It just turned into 10 years.”

Keyhole’s John Hanke (left) and Chikai Ohazama (right) the day the Google acquisition closed in 2004. Photo by Brian McClendon

John started at Google in 2004 when it acquired his startup, Keyhole. Keyhole was a spin-out of sorts, too, growing its way out of a company called Intrinsic Graphics.

Intrinsic had set out to build a cross-platform video game engine – think Unity, but a decade and change too early. Along the way, the team at Intrinsic had built a demo that allowed users to zoom in and out of a wildly detailed view of the Earth. Inspired by the mind-bending documentary Powers of Ten (or maybe tech in Neal Stephenson’s novel Snow Crash, depending on which Keyhole co-founder you’re asking) the user could parachute from a sprawling view of the entire globe all the way down to a bird’s eye view of their home with the flick of a mouse wheel. It was just a demo, but it was a very, very good one. Everyone seemed to care more about that demo than anything else Intrinsic was building.

Keyhole’s Earth Viewer, circa 2002

Intrinsic hired John to build this demo into something more. When Intrinsic shut down operations in 2003, John and a handful of employees kept charging forward under the Keyhole name.

Keyhole never made a ton of money. A deal to use Keyhole imagery on CNN got them in front of an audience, and a partnership with NVIDIA got them distribution. An investment from the CIA-backed venture firm In-Q-Tel — who said they would use Keyhole’s tech to “support United States troops in Iraq” — kept them moving. But the ex-Keyhole folks I’ve talked to are pretty open about the company having danced on the edge of broke more than once before Google swooped in and bought it in 2004.

Keyhole’s earth viewer lives on to this day, of course. It’s just called Google Earth now.

John Hanke being lifted up by his colleagues at a Google party a few months post-acquisition. Photo by the late Andria Ruben McCool, used with permission from her family.

Quitting Google to join … Google?

Flash forward to October of 2010, just a few months before the team that would become Niantic started to form. John’s intended “six months” at Google had just clicked over to its sixth year. He had spent that time leading Google’s Geo division — that’s Google Earth, Google Maps, and just about anything else that had to do with location. If you were typing a street address into something Google-branded, it was probably in a product John oversaw.

At the beginning of October, word got out of a major, headline-dominating leadership change within Google: Marissa Mayer, who had been leading the company’s search division, was asked by then-CEO Eric Schmidt to instead focus solely on Geo, after what I’m told was a bit of a clash with the lead engineers on search.

Just months later, in January of 2011, John Hanke announced that he was leaving the Geo team to run a new, unnamed “entrepreneurial lab” within Google.

The timing wasn’t a coincidence.

“At the end of it,” Hanke says, “Marissa was going to be spending a lot of her time on Geo. And, well, I wasn’t sure if it was a big enough job for both of us.”

“I like Marissa,” he adds. “I have a huge amount of respect for her, and she was a great boss. But it just didn’t… you know. Her other responsibilities got collapsed, and she was going to be doing mostly what I’d been doing.”

Add in a traffic-heavy commute from the East Bay that had been grating at him for years, and John Hanke decided his “six months” were finally up.

So he quit. Or, at least, he tried to quit. He sent a letter to Eric Schmidt, Larry Page and Sergey Brin announcing his intent to resign and “do something new.”

Like most companies, Google doesn’t like when employees leave. Especially employees who ran key parts of the company for years. Leaving means competition. Leaving means potential opportunities lost.

John eventually sat down with Larry Page to figure out what it’d take to keep him within Google. They talked about John’s interest in augmented reality. They talked about a book called Freedom by David Suarez, which centers around an out-of-control AI that taps a network of real-world operatives to control the world (the earliest hints of Niantic’s first game, Ingress, already sneaking in here years before it’s made.)

John wanted to take his interest in AR and his background in maps and gaming and mash them all up and see what it could look like. Larry wanted it to happen within Google.

Meanwhile, David Lawee (who ran Google’s M&A department, and now runs Google Capital) and Salar Kamangar (a creator of AdWords, and previously the CEO of YouTube) had been pushing for ways for people to start new endeavors within the company, rather than leaving to start on their own.

“They recognized that this was going on” says Hanke. “That people wanted to do new stuff and that all of it would happen outside of Google, and potentially there was this exodus of talent that they might rather keep in the company.”

John’s “lab” would be one of the first iterations of a new concept that would come to be known as an “autonomous unit.” They wrote a contract. It’d be a startup, of sorts, within Google. It’d get a budget and dedicated head count. He’d have 24 months to figure out… whatever this thing was.

They’d later extend the contract by over a year-and-a-half — but along the way, John made sure the contract included one key thing: when time was ultimately up, the fate of the company was his to decide. If he wanted to take the company outside of Google, he (theoretically) could. They’d just have to figure out how that would work when the time came.

“I was like, well… if I’m going to do this inside of Google, I want to know this isn’t all going to get taken away in a quarter where somebody decides they want that headcount for some other project.”

This, of course, ended up being pivotal.

The Autonomous Unit

The autonomous unit started out small.

Hanke found an unclaimed desk in Google’s San Francisco office, so he wouldn’t have to drive to Mountain View anymore. He cracked open a Python book and started brushing up and tinkering with ideas. He knew he wanted to build something with AR, something with maps, and something to get people to look up off their phones. John has three kids, so he understood the siren song of the screen all too well. Could they build something that used those screens to get people up and out, instead?

As word spread of what he wanted to build, the desks around him started filling up.

Some of those first employees had been with John since before the Keyhole acquisition, or otherwise followed him from the Geo division. People like Phil Keslin, now Niantic’s CTO, and Lenette Posada Howard, its VP of Operations.

Others came from all areas of Google. As part of the mission to keep people at Google, Google made it relatively easy to hop from team to team internally, and Niantic was, at this point, just another team.

Dennis Hwang, the designer of the original Gmail logo and many of the first Google Doodles, joined and still serves as the company’s head of Visual/Interaction design. Paul Martini, who would years later lead UX/UI design on Pokémon GO, was a receptionist at Google whose sketches John spotted one day. Others started by contributing during their “20% time” — a Google policy which encouraged employees to work on side projects with a fifth of their on-the-clock hours.

But the growing team still didn’t know exactly what they were going to build. They threw around ideas, doodled, and tossed them out.

Niantic’s first prototype (and its first cheater)

At some point in the first few weeks, I’m told, a Google employee posted a message on the company’s internal message boards, offering up a massive collection of board games for cheap. John and at least one other employee went and picked them up, and the team sat around playing round after round of RISK and other board games, tearing apart their mechanics for inspiration.

The first result of this tinkering, or, at least, the first that anyone inside of Niantic remembers, was a prototype called BattleSF.

Illustration by Bryce Durbin of TechCrunch, 2019. Sadly, no screenshots of BattleSF seem to exist anymore.

It was a game, meant to be played in the browser on your phone. The goal? Build your army, take over SF. The in-game city was divided block by block. You’d physically walk to a block — even with this first prototype, the company had its focus locked on getting people outside and moving — and press a button to attack it. If your army was strong enough, you’d take over that block. They would later add a feature called “donuts” which, depending on who’s telling the story, would motivate your army or convince enemy soldiers to defect.

“It was just this shitty Python code,” John tells me. “But it was really fun, and we all started playing it. In a sense, it led to Ingress.”

Hilariously, this is also when Niantic busted its first spoofer. Players “spoofing” (tricking a phone’s GPS into thinking it’s somewhere it’s not to warp around the game world and effectively cheat) is something that Niantic battles with to this day. But their first spoofer? That was an employee.

“We actually learned about spoofing because of [him.]” Niantic CTO Phil Keslin tells me. “He had this habit of going out and doing things in the game in the middle of the night, when his wife and kids were asleep. Except it’s like, 1:30 in the morning. John’s like, ‘Wow! He’s really diligent! He gets up and conquers San Francisco!’ But no he wasn’t! He was doing it from his bed! He figured out that he could enter the latitude and longitude in the address bar and warp around.”

Before they even really had their first game, they had cheaters.

Alas, no known screenshots of BattleSF exist. I can find very, very few traces or mentions of it on the Internet before this, and John Hanke says it probably all got destroyed.

“We lost a bunch of stuff when we spun out” he notes. “They let us take some emails and documents, but they ran this algorithm over it to strip out anything sensitive. There was a lot of loss in the process.”

But before there was Ingress, before there was Pokémon GO, there was BattleSF. It was RISK, but on top of the real world. And with Donuts.

Taking a Field Trip

Whereas BattleSF never saw the light of day, one of the team’s other early concepts, Field Trip, would be the first thing it launched publicly. It launched in September 2012.

It … was not a huge success. You can still download it to this day, and the website is still around — complete with “Niantic Labs at Google” copyright and all. Just don’t expect updates any time soon.

Field Trip was built to tell you about the world around you. Like the Niantic games that would come later, it was heavily focused on moving around the real world. When you walked near a location with an interesting history, a notification would pop up prompting you to learn more about it.

“From the beginning, it was this kind of grand search for ‘How can we channel someone’s commute route?'” Dennis Hwang tells me. He’s wonderfully humble for a guy whose logos and designs have been seen by pretty much anyone with a computer or smartphone. In about an hour of chatting with him, he never once mentions that he’s the artist behind the Gmail logo. If I were him I’d probably put the logo on a t-shirt and try to get everyone to ask me about it.

“We had this hypothetical person in mind; someone who would commute from work, to home, to work, to home, for say, 20 years,” he continues, “Same routine, every morning. If we could just change that person’s route, even by a block, just because our app said ‘Hey, there’s something interesting!’ and break that habit, we’d have been high-fiving each other.”

The first challenge? Finding interesting locations to share and getting them all in a database. Niantic worked with Arcadia, a San Francisco publishing house with books focused on local interest and history. Phil Keslin tried building a system to parse the text of these books to automatically extract information and geotag it to the right real world spot, but that proved harder than expected. They called in for insight from the Google Books team (something Niantic could do, being a part of Google), but they still couldn’t crack it. After about two months, they switched entirely to manual, human input. That might work for a few cities here and there, but it doesn’t scale very well.

It was in one of these local history books, I’m told, that the until-then-nameless team learned about a ship called The Niantic that had run aground in San Francisco. Over time, the stranded ship was converted into a hotel, repeatedly burned to the ground, and rebuilt as a different building and business time and time again. According to Atlas Obscura, the last remnants of the Niantic’s bow are still buried beneath the San Francisco Maritime Museum, just a mile or two from Niantic’s SF office.

Perhaps most damningly for its future, Field Trip was originally conceived for Google Glass — which, as anyone who has a Glass unit gathering dust in some closet could tell you, limited its reach. The team wised up and shifted focus to smartphones before Field Trip actually launched, but even still, it never quite caught on.

A simulated view of what Field Trip looked like through Google Glass. Photo credit: Google

“We didn’t quite nail the user experience. It was a hard balance to strike, especially without Google Glass really taking off,” says Dennis. “On the phone… it was just a constant interruption when you needed to get somewhere. You’d turn off the app, and forget to turn it back on. So we pivoted.”

Ingress sets the course

Users seemed to like the core idea of Field Trip enough to check it out — they just weren’t coming back. Perhaps the answer was some sort of gaming mechanic? The team began re-imagining Field Trip as a game, mashing it up with what they’d learned from those early days where it was just a handful of employees tinkering with battleSF.

Keep the map, and the mechanic of encouraging users to move around the real world. Take the focus on real world points of interest from Field Trip, and the idea of battling for “control” of an area from battleSF. Throw in heaping fistfuls of Sci-Fi backstory, swirl ’em together, and…

You get Ingress.

Ingress, circa 2014

Ingress is a tale of two opposing factions: the Enlightened (Green) and the Resistance (Blue). Research at (the in-game, fictional version of) CERN leads to the discovery of Exotic Matter (or “XM”). The two factions stand in staunch opposition as to how XM should be handled. The Enlightened see it as a driving force behind mankind’s progress; the Resistance see it as a threat from outsiders trying to influence the course of humanity.

After it’s discovered that XM tends to accumulate around “Portals” — real world points of interest, like statues, murals, fountains, etc — the two teams set out for control. Players attack and “hack” these locations to claim it as theirs, then “link” multiple portals to conquer everything in between for their respective faction. With proper coordination, factions can take over absolutely massive swaths of land. As in entire countries.

There’s more to it than that, with a plot that expands over the years but, well, the rabbit hole goes a bit too deep for this post. There are conspiracies. And spies. And murder. There’s a dude named Hank Johnson (not to be confused with John Hanke) and — harkening back to those early days of geeking out about Sci-Fi books with Larry Page — artificial intelligence trying its damndest to steer the world.

Ingress launched at the end of 2013. It was Android only, at first, because Google. (The team’s desire to be on both iOS and Android would later play a part in the decision to spin out.)

Niantic’s sandbox

Despite the shared DNA, Ingress has never really taken off in the way that Pokémon GO would later in Niantic’s story. It’s seen roughly 25 million downloads in the five years since launch, whereas Pokémon GO would go on to hit a billion downloads in under three.

Years after its initial debut, Niantic would overhaul and re-release Ingress as “Ingress Prime”. They’d rebuild it on the Unity game engine, and polish the tutorial experience in hopes of making it more accessible to new players. The success of this effort, I’m told, has been “mixed”. They kept the original Ingress app around under “Scanner [REDACTED],” knowing that version one of this newly rebuilt app couldn’t cover everything that longtime players might expect.

“It’s not attained that level of market fit or buy in from the players. There’s a lot of work that still needs to happen on that product.” said one Niantic employee. “The team is working hard to ensure that we can get to parity, and we can start creating all these other features. We launched it and people were excited for it, but my sense is that a lot of people still play the original.”

But none of that means Ingress isn’t important. While Ingress might not have as huge of an impact on Niantic’s bottom line as other titles, it serves a key role: it’s the company’s sandbox.

Much of what the company does with its games, it battle tests with Ingress first. Things like point of interest submission, events, and many of the core game mechanics — they all grew out of Ingress.

With Field Trip, Niantic discovered just how hard it was to build a database of interesting locations.

With Ingress, the company finally cracked it.  They started to allow higher level players — those less likely to game the system, as they’ve got more to lose if caught — to submit locations they think might work as potential “Portals”. They’d still need people to manually moderate these submissions for accuracy, but it was a whole lot more feasible than doing it all themselves. These user-submitted portals later doubled as Pokéstops and Gyms.

Unlike the Pokémon and Harry Potter titles that follow it down the road, Ingress belongs entirely to Niantic. If they want to try out some wild idea, they can; they don’t need any third party IP holder’s approval. If it doesn’t work? Whoops, roll it back, or tweak it into something else. If it’s a hit? Maybe they can find a way to make it fit in their other games.

Owning the whole thing also lets Niantic have more fun with the story and the universe surrounding it. Multiple books have been written that take place within the world of Ingress, and the company launched an Ingress anime series in Japan in October of 2018. They wrapped recording on the english voice acting back in early March, and it’s slated to launch on Netflix in the US on April 30th.

Early work and playtesting on Ingress also led to a concept that would quickly become a core element of Niantic’s operation: events. Across its games, Niantic hosts real-world events meant to bring players together in the same physical space. Lots, and lots of players. One meet-up (or “anomaly”, as it’s known in Ingress parlance) saw 15,000+ players take over a massive square in Tokyo, Japan.

Niantic’s Archit Bhargava in October of 2013 at one of the first big Ingress gatherings. Photo credit: Niantic

But the first inklings of these events were a bit more humble.

“We had this travel policy. If anyone was going on vacation, and they were okay with giving up one evening, Niantic would fund, basically, a meetup. You could go to any bar, any cafe, and you’d just buy beverages and snacks for Ingress players,” says Archit Bhargava, Global Marketing Lead at Niantic. “When John was in Italy on vacation in 2013, he met up with a bunch of players. He bought them a round of beers at this bar in, I think, Rome. He sent us all these notes after, like ‘here’s what I’m hearing, they love all this’… That’s how the events strategy evolved.”

Out of time

It’s the end of 2014. Ingress has been out for about a year, and has a growing fanbase. Niantic had grown to a team of about 75 people. Pokémon GO was secretly already in the works — we’ll go deep on that in part two, but at this point The Pokémon Company is onboard to build something with this little team inside of Google and development on the game is underway.

Niantic’s original contract had been extended because things were working well enough. But that couldn’t last forever, and time was up.

Google was prepping for another massive executive change. The company’s then-CFO, Patrick Pichette, was preparing to step down, with a new CFO lined up to take the lead. Every department was going back and cleaning up their books. It was, as many have told me, a period of “belt tightening”.

Niantic Labs wasn’t making any significant amount of money at this point, and not everyone at Google was onboard with it. It took employees from other projects. Its main attraction, Ingress, seemed like this weird little game that had to do with “hacking” and murder and (maybe evil?) AI. It wasn’t exactly something that Google’s marketing teams liked explaining.

Even those who did get it largely wanted Niantic to be integrated into another team within Google.

“We started having these discussions,” John Hanke tells me. “And, you know, the internal options weren’t very attractive.”

Maybe it’d work as part of the Android team? They could integrate the work they’d done with Field Trip into Google’s homescreen— but that’d mean throwing out Ingress, or at least splitting up the team.

Maybe it could become a part of the Google Play team? That wasn’t ideal either.

“If you’re going to make stuff for consumers, you have to be on iOS and Android.” says John. “You can’t have a hit product that’s only for one.”

But, as you might recall, John had left himself one other option in the contract: spinning out. But it wasn’t as simple as packing up the office and heading for the door.

“We had filed a bunch of patents.” says John. “We used a bunch of Google technology. We had invented a bunch of technology during this period as an autonomous unit. So we had to figure out how all that stuff could even legally leave Google.”

That process — just figuring out who owned what, and what Niantic could take with them — would stretch until October of 2015.

Simultaneously, John was trying to raise the venture capital needed to keep the company going. Google would give them a few million in exchange for a stake in the company — but beyond that, Niantic would be its own operation, with its own employees to pay. The effectively unlimited resources that came with being under the Google umbrella would be gone. Niantic needed more money.

“It was a hard deal to sell,” John says. “VCs were looking at this and saying ‘Well, this is weird. You’re spinning out of Google? We don’t really understand that. Google’s going to own this chunk, so we get less?”

One of Niantic’s earliest investors, Cyan Banister, had the same experience when trying to help the company raise. Cyan was an early Ingress player, and had reached out personally to John Hanke to invest as an Angel. But when she tried to get other investors onboard, she says “it was like I was handing them a tasty cake and they didn’t know what to do with it.”

And even if Niantic could raise enough to pay everyone’s salary, how many employees would choose to set sail with Niantic? Google would be happy to hang on to everyone who’d stay. It had stability, perks, and massive annual bonuses. It had dozens of teams you could transfer to when a project got boring or failed. Niantic had none of that.

But time was almost up. They needed more money to do this right, but few VCs were willing to jump in. And either way, they were probably about to lose a good chunk of their team.

Oh, and they still had to tell The Pokémon Company about the whole spin-out thing.


Niantic EC-1 Table of Contents

Also check out other EC-1s on Extra Crunch.