Tech by VICE

We Talked to the Creator of PokéVision About Losing 50 Million Users Overnight

The pokémon tracking site was hugely popular before 'Pokémon Go' developer Niantic took it down. Here's what happened.

by Victoria Turk
Aug 5 2016, 7:45pm

Image: inmedialv/Shutterstock

PokéVision creator Yang Liu is a big Pokémon Go fan. That much is obvious, from the impassioned open letter he penned to its creators earlier this week to the way he talks about his experience playing the game.

"I wasn't particularly looking out for Pokémon Go or anything, I wasn't like super hyped for it," Liu told me over Skype. "But when it came out and I tried it, I was like, 'Wow, this is really cool.' It brought back a lot of memories from when I was little."

Niantic has been less enthusiastic about Liu's contribution to the recent Pokémon frenzy.

PokéVision was a popular website that allowed Pokémon Go players to track pokémon in their area. PokéVision used the Pokémon Go API to display pokémon spawns on a map in real time, so people could see when pokémon appeared nearby and go and catch them. But last weekend, the site stopped working. At the request of Niantic CEO John Hanke, the three creators (Liu is the only one named, the others preferring to keep their identities hidden) pulled the project.

Screenshot: PokéVision

"John Hanke reached out to us and was like, 'Hey, would you mind please taking down the site,'" Liu recounted. He'd said previously that if Niantic asked them to take the site down they would. "So yeah, we did that," he said.

Liu wrote in his Medium post a few days later that within three weeks PokéVision had grown to almost 50 million unique users, or 11 million daily (which, if accurate, represents about half the number of Pokémon Go players). The site now only displays an error message. I reached out to Liu because I wanted to know more about how the site came about, and what it felt like to build something that gained millions of users in a matter of weeks, only to lose them all overnight.

Liu, a neuroscience graduate who now works in tech but wouldn't tell me where, said the idea for PokéVision was sparked when Pokémon Go's own tracking mechanism broke. The game originally showed players how close nearby pokémon were through a system of footprints, the number of which signified the rough distance needed to travel to catch a specific monster. This feature soon became much less useful, as a bug made it show three footprints for every pokémon. In a later update, the feature was completely removed, though Niantic has since suggested that it will return.

The PokéVision crew continued playing, waiting for a fix. "And then we saw that the game actually sends you, the client, the exact locations of the pokémon next to you," Liu explained. "Of course, that wasn't displayed on the interface, but it was sent on the backend—this was on Reddit and everywhere."

The way Liu tells it, PokéVision was "whipped up" as a temporary solution to share among friends. It took off on Twitter, and soon they had to quit playing Pokémon Go to maintain the site.

"The difficult part was definitely handling all of the users and making sure everything worked," he said.

They were surprised at the number of people who flooded to PokéVision. "If you used our site, it was basically down every hour—every hour of every day," he said.

They rushed to keep it working by scaling up their server size (they used Amazon Web Services) and optimizing their code as much as possible. Liu would not disclose how much it cost to keep PokéVision going, but that it was "quite a few thousand a day" towards the end, which was funded out of their "own pocket."

Why spend so much time and money? Liu said the main motivation was the popularity of the site, whose metrics offered what seemed to be a once-in-a-lifetime opportunity. But the cost was a worry, and if PokéVision hadn't come to an abrupt hiatus, the creators were considering how they could turn it into something sustainable.

"We were actually extremely concerned about the cost, because if we kept it up for even 30 days, it would have been well into six figures," he said. "So we were working things out with, like, Amazon—you know there's startup programs and stuff like that everywhere, right?"

They didn't get any further.

"Anything beyond what is actually in the game is, in the most technical terms, cheating"

Liu said they always aware there could be an issue, what with PokéVision clearly being based on someone else's intellectual property. They also knew Niantic might not like what they were doing.

"Anyone can see, this isn't pure white," he said. "It's not black-and-white, but it's definitely not pure white."

That's why they didn't want to monetize the service by charging to use it, or by making a PokéVision into an app, which would more obviously clash with Pokémon Go's territory.

Hanke did not give Liu a clear reason why he wanted the site taken down—and Liu concedes he doesn't have to—but in a blog post published on Friday he suggests that a main motivation for blocking third party "scrapers" is the strain they put on Pokémon Go servers, especially as the game is rolled out into new countries.

In the post, Hanke writes that these scrapers (he doesn't mention any by name) "hurt our ability to deliver the game to new and existing players" by using up server resources and taking developers' time away from working on other features.

He includes a graph which shows a drop in server resources consumed when Niantic "blocked scrapers." This drop occurs on August 3—several days after PokéVision shut down—and doesn't actually say much, as the y axis is not labelled, making it impossible to tell how much the server resources were affected.

The graph showing a drop in server resource consumption. Image: Niantic

I reached out to Niantic to clarify what the graph showed, but they said they had no additional details to share than what was in the blog post.

Liu wasn't too impressed with Hanke's post. "To be frank, it's an attempt to please the crowd," he wrote in a Skype message. "We all learned in grade school to label our graphs." He also suggested that bots would cause more server stress than trackers such as PokéVision.

Liu no longer plays Pokémon Go. He admitted that using PokéVision could, strictly speaking, be considered cheating—something Hanke has also insinuated. "Anything beyond what is actually in the game is, in the most technical terms, cheating," Liu said. "Say the game itself didn't have the footsteps stuff and then a site made the footsteps tracker again—that would technically be cheating, because it's not in the game any more."

He said that he reached Level 21 and his best pokémon was a Charizard, but without a way to track pokémon, the fun had gone for him.

"Because of how the game barred its own features and any other ability to really find pokémon reliably and catch them, it really turned me off," he said. "When you see a Snorlax, you're going to catch it whether you see it on PokéVision or in the game."

Augmented Reality
Pokemon Go
Mobile Gaming
motherboard show
John Hanke
Yang Liu