Two weeks ago, Pokémon Go launched in 26 countries to mixed reviews and wild popularity. "Your game launches as a viral mega-hit," is a pretty good problem to have, all things considered, and both Nintendo and Niantic's share prices confirm this fact.
Yet the launch also came with its share of server troubles, which provides an interesting window into the magnitude of the problem faced by mostly ex-Googler-staffed Niantic, the developer of Pokémon Go. Even the CTO of Amazon, Werner Vogels, took a shot:
What can we learn from such a massive and publicized cloud-based app roll-out?
1. You never know how big the wave will be
Even though Pokémon Go was backed by the Google Cloud Platform – no stranger to heavy workloads – the sheer volume of downloads, account creation, and player interactions overwhelmed whatever resiliency and failover mechanisms Niantic had in place. It's worth mentioning that fellow game developer, Blizzard, no stranger to big launches themselves, had similar scale issues with the release of Diablo 3. The key lesson here is to be prepared to adapt as conditions on the ground change.
Thank you for your patience. We have been working to fix the server issues. We will continue rolling out #PokemonGO to new countries soon.— Pokémon GO (@PokemonGoApp) July 8, 2016
2. Slow your roll (out)
While it could be that the largest app launch in history was just something nobody could be prepared for, it has been suggested by many that a grueling launch schedule enforced by Nintendo was partially to blame. Niantic's decision to segment their roll out across the globe, and to delay further launches as conditions demanded was probably a good one. Not every app can expect this level of retention.
3. Lock down your endpoints
This is kind of obvious advice and yet it is often overlooked. In less than two weeks, the Pokémon hacking community researched and implemented large swaths of the Pokémon Go API to build third-party tools that allow (among other things):
- Automated Pokémon Capture
- Automated Spoofing (GPS modification)
- Automated alerts for Pokémon in area.
These tools are almost all violations of the terms of service for the game, yet you can bet they will be expanded as time goes on. Niantic and Nintendo stand to lose some in-app purchase revenue through such cheating. The tools also generate undue load on the game servers. What if this app were handing more serious transactions? Clearly the ability of the community to adapt to lax security checks is something that needs to be carefully considered at launch time.
So, to recap, the three important lessons to be learned from the launch of the smash-hit gaming app, Pokémon Go, are: You never know how big the wave will be, Slow your roll (out), and Lock down your endpoints. What are your thoughts on this new cloud-based app? Comment below and check out some of our other posts on the cloud here.