The document discusses lessons learned from building augmented reality games at large scale. It describes how Niantic built several applications like FieldTrip and Ingress to learn how to design a platform to support planet-scale AR. Key learnings include the need for interesting mapped locations, a dynamic and low-cost backend system, and engaging users to populate the map. The document stresses that success requires enabling many third-party experiences, not just Niantic's own apps, which means the platform must support external contributions.
3. Sought ways to motivate people to…
• Get outside and explore the world around them
• Get a little exercise along the way
• Enjoy the experience in the company of others
4. FieldTrip
• Discovery platform for geo-referenced, curated content
• Designed to unveil what’s hidden while walking around
• Leveraged AR for unobtrusive content delivery in 2013
5. • Engaging global game with global game play with SciFi story
• Mapped virtual game locations with physical objects
• Players have submitted and curated over 6M thus far
• With IoT can now create interactive objects
• Added FieldTrip content to foster exploration with missions
• Incredibly loyal fans
• Explored a sponsored location revenue model
Ingress
7. • Learned what it takes to build an engaging real world game
• Leveraging players to create the game board
• Fostering a sense of community
• Discovered what is needed for a successful platform
• Built a new platform designed to
• Increase efficiency to improve player experience
• Reduce operating cost by at least 10x
Ingress
9. • Leveraged key learnings from Ingress
• Cultural phenomenon, but still a hit game
• First title to be launched on the new platform
• Resolved many bottlenecks both internally
and externally
• Wild ride proved it would scale, exceeding
our goals
Pokémon GO
10. • A true lesson in scale!
• Pokémon GO Fest to Pikachu Outbreak to Safari Zone
• Responded to bottlenecks as they arose in real time
• Supported about 3.8M people in 2017
Pokémon GO Events
11.
12. • The primary metric we use is kilometers walked
• We feel pretty comfortable stating that our players have
walked more than those of all other AR titles combined
Measuring Success
13. • The primary metric we use is kilometers walked
• I feel comfortable saying that the players of one of our games
have walked more than those of all other AR titles combined
Measuring Success
> 340M KM
14. • The primary metric we use is kilometers walked
• I feel comfortable saying that the players of one of our games
have walked more than those of all other AR titles combined
Measuring Success
> 340M KM
Ingress
15. • The primary metric we use is kilometers walked
• I feel comfortable saying that the players of one of our games
have walked more than those of all other AR titles combined
Measuring Success
Ingress
> 340M KM
Pokémon GO
> 15B KM as of May 2017
16. • People talk about tech with little talk of the apps
• We started with our mission of finding ways to get people
outside, together, exploring, and getting a little exercise
• We built several apps to inform the platform design
• Similar in nature to how AWS and GCP were built
• With so few companies building AR applications at scale, I
thought I’d share some of what we’ve built thus far
Building a Platform
17.
18. • Maps, streamable to highly interactive clients, worldwide
• Having appropriate places to explore or play
• Single persistent world because the real world isn’t sharded
• Systems that allow people to communicate and coordinate
• Managing people traffic so infrastructure doesn’t collapse
• Highly dynamic yet low cost systems for many users
• Rapid scaling to accommodate fast changes in demand
• Ingesting and distributing many terabytes of data
A Platform for Planet Scale AR needs…
19. • Managing many petabytes of stored data
• Dealing with spoofers, bots, scraping
• Establishing customer service for a lot of users
• Coordinating with property owners
• Creating engaging events that support local communities
• Interesting real world locations to anchor the game
• In game location sponsorship
• We will be talking about some of the other challenges we are
working on in coming weeks.
A Platform for Planet Scale AR needs…
20. • Must include many map types: human and machine readable
• Making human readable maps useful takes time (months)
• Deriving machine maps is a whole different problem
• More than just where, must also provide the what and why
Maps
21. • FieldTrip content curated and submitted by
experts
• Places are curated and submitted by players
• Players are explorers of their community
• Submissions and updates ongoing since 2012
• Frequency continues to increase over time
• World is a dynamic place
Interesting Places
22. • Ingress showed that AR games can be expensive to operate
• Set the goal for our platform on a 10x cost reduction
• Achieved a 50x reduction at launch of Pokémon GO
• Without this focus Pokémon GO would not have survived
Dynamic and Low Cost System
23. • Technology is only a small part of the whole picture
• Success is dependent upon the users
• Too few users results in something akin to a game of
checkers with 90% of the squares missing
• Users are explorers who populate the map and share with
others what is interesting about where they live
Planet Scale AR - The Key Lesson
24. • We set out to encourage exploration in the great outdoors,
getting some exercise and having fun with other people
• To achieve this requires that we enable the creation of a
plethora of experiences, not just our own
• This requires many contributors
• Which means a platform is needed to make this a reality
Not a Platform for just Niantic Applications
25. • When successful, your ability to scale will matter… a lot
• A single consistent world enables social interaction
• AR depends upon people using your app, filling in the blanks
Things to Remember