6.4 Towards More Actionable Tropical Storm Warnings, Google Public Alerts

Wednesday, 13 January 2016: 9:15 AM
Room 333-334 ( New Orleans Ernest N. Morial Convention Center)
Pete Giencke, Google, Inc., Mountain View, CA; and N. Savio, P. Patel, and S. Hakusa

People in the path of the next Hurricane Sandy can expect to receive news of a tropical storm watch or warning from myriad sources: the internet, social media, radio, TV, etc. Though these alerts often provide useful information, the general coarseness of the content (geographic and otherwise) can leave individual users wondering if they're actually at risk.

With over 80 million people[1] directly exposed to tropical storms worldwide, and the lead time associated with these events, tropical storms represent a major opportunity to improve public safety outcomes, at scale. Over the past year, Google has undertaken significant user and field research to better understand the needs of these at-risk people. The team learned not only of general alerting ambivalence ahead of major storms, but also of the cognitive overhead inherent to much of the early warning information currently being presented to users.

Key findings from this research include a) specific individual risks posed by a tropical storm is unclear to users, b) current tropical storm visualizations lead to fundamental misunderstandings of the nature and path of a storm, and c) the tone and presentation of alerting and preparedness content does not resonate with users, and these users then ignore this content, and d) probabilistic warning information is almost universally unintelligible to the lay user. Through its Public Alerts product, Google has built upon these and additional lessons learned to build a better tropical storm early warning content, optimized for comprehension and actionability.

This presentation will focus on the social science behind Google Public Alerts improved tropical storm warnings, and offer insights and recommendations from our user research and testing.

[1] http://www.preventionweb.net/english/hazards/statistics/risk.php?hid=58

Supplementary URL: http://www.google.org/publicalerts

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner