August 9, 2016
What you can learn from Pokémon Go: Be prepared for massive success so you can scale to match customer demand.
John Hanke, creator of Pokémon Go, is a technology veteran who is no stranger to the startup world. He launched three gaming companies before landing at Google through an acquisition. So, it’s not a fluke that the overnight rise of Pokémon Go is nothing short of spectacular. Within a week of the game’s launch, over 10% of Android users had installed it and 6% of the user base was active. Moreover, the daily usage average of 43 minutes exceeds the more popular messaging apps.
That growth is unheard of and with it comes challenges in scaling up to match demand. But, Hanke must have anticipated the game’s unprecedented popularity. At least he was hopeful because he baked in a plan before the game’s launch to scale the game to accommodate record-high usage rates.
Even though Niantic, the company behind Pokémon Go, leverages Google’s cloud capabilities, the massive scale of its user base (>20M DAU) experiencing hyper growth of 4-5M users daily can be taxing. The company’s primary provider and investor struggled to keep pace with the sheer data volume and computational demands that the growth posed. Users were experiencing trouble logging in.
Given the rapid growth, Niantic had to pause its international rollout, but the suspension didn’t last long. It resulted in interim outages, and with avid users finding unofficial workarounds to play the game in locations where it was not available. There were news articles on the potential demise of the popularity given the short attention span of its users. Well before millions of users started scrambling to nab Pokémon characters, Hanke had to have designed the game to scale across servers beyond the incumbent cloud provider. When users started experiencing difficulty, Niantic was able to leverage a second cloud provider to correct the problem within 24 hours. Though there are users who may be disappointed by the bumpiness of the experience, the price of scaling up was capped through foresight and diligent planning.
With Hanke’s forward thinking, Niantic mitigated the impact of such issues on longer-term retention rates and engagement. Reliable, instant gratification is key to making a game go viral. Without Hanke planning to place the game on a backup cloud, Pokémon Go’s global growth would have been a No Go.
In an ever-connected world, where user experience and engagement is essential to growth, organizations cannot take infrastructure and availability for granted. For large-scale deployments, some of the lessons include
- Leverage cloud for business models where there is uncertainty around the uptake and thus demand
- Use instrumentation on the end points (e.g., mobile app) to aid in the detection of anomalies and understanding end-user engagement. It providers early warning of issues and allows for mitigating actions to be taken
- Have two or more cloud providers in case one of them has challenges scaling as fast as needed (admittedly this would be the case of massive user base and/or hyper growth on the scale of Pokemon Go)
The Pokémon Go game is operating at a level of capacity that most companies won’t ever see or need to plan for. But, cloud capacity issues will increasingly arise as companies grow more sophisticated in their use of more advanced technologies like augmented reality, virtual reality and the Internet of Things that require muscular data crunching power. The sky is not the limit when it comes to cloud computing. If you want to scale, you have to plan accordingly.