Location-based Content Discovery
| Objective
Reduce the bounce rate of Discovery Channel’s web and mobile app visitors.
| Team Leads
| Insights
20-50 year old male audience that co-views with their family.
Discovery Channel’s programming strategy has evolved over the last 8 years to primarily focus on automotive, lifestyle, and competition shows.
The visitors to the site and apps are a younger demographic than the pay television audience.
The automotive, lifestyle, and competition shows promoted in the apps and website did not match expectations for the Discovery brand, which meant nature, science, and tech topics to visitors that failed to convert to viewers.
Nature, science, and technology shows received little to no marketing promotion.
When they found nature, science, or technology episodes few were enticing enough to watch because the show and/or its cast were not familiar to the visitors.
Location is a central element of the storyline for Discovery Channel’s nature, science, and technology shows. Many included the location of the production in the show name (e.g. Serengeti, Pacific Warriors).
| Hypothesis
Presenting videos by their location will increase relevancy, overcoming the low awareness of specific videos to better convert visitors to viewers before they bounce.
| User Experience
The Shark Explorer experience was developed by examining best practices from websites and apps that employed location to create prototypes. These were validated and evolved with weekly remote usability testing starting from the first prototype through production-ready software.
The key features of Shark Explorer were the near real-time shark tracking paired with Shark Week content presented as waypoints on an interactive world map. When tracking the sharks, visitors could find detailed information on the shark including species name, weight, length, gender, who tagged it and, its locations in the ocean. As visitors tracked a shark on its journey through the ocean they could quickly discover short, mid, and long form videos featuring it. In addition to waypoints for each tagged shark, there were also waypoints for the locations of other Shark Week videos.
| Engineering
To keep the development effort as small as possible the team selected the “Wizard of Oz” testing methodology, where members of the product team orchestrated the consumer experience manually.
The Shark API fetched shark location data from Argos (humane tagging partner), then combined it with data from a spreadsheet created by the product owner that described how the content should be presented in the user interface. The team traded off tooling with safeguards to reduce the development effort (e.g. confirm valid video URLs before publishing). Human error was mitigated with a manual peer review process before and after publishing.
The front ends integrated Google Maps with data from the Shark API. The team traded off creating more responsive native iOS and Android experiences, opting for a webview in the mobile apps that used the same responsive implementation from the website.
| Technologies
Location services
Golang / MongoDB / REST
MS Excel
HTML / CSS / JavaScript
Google Maps
| Experimentation
Shark Explorer was made generally available on Discovery.com and mobile apps for 3 weeks of experimentation in the summer of 2018 as part of the Shark Week marketing campaign. This provided a lab to test and learn whether nature content presented on an easily accessible map would get more visitors to watch videos.
The control for testing Shark Explorer’s performance was the homepage of the website and mobile apps that presented the same videos using a carousel and grid design pattern.
The variants tested included:
◦ Starting the user journey as a full home page takeover versus the user launching Shark Explorer from a UI control on the home screen.
◦ The default starting position and level of zoom for the map.
◦ Shark tracking waypoints versus show waypoints versus a mix of both.
◦ Short form video versus mid form versus long form versus a mix of content formats attached to each waypoint.
| Retrospective
◦ The project was delivered on time due to rigorous change management by the entire team, enabled by clear tenets and goals.
◦ Shark tracker waypoints converted better to viewing and generated longer sessions than waypoints about shows.
◦ Additional design solutions are needed to help viewers discover waypoints outside of the default map view.
◦ New metadata, QA, and operations automation capabilities are required to offer the experience at scale.
◦ The team had fun together, especially naming the sharks after friends and family.
| Key Results
Higher Streaming Starts
◦ 7% more visitors converted to viewers
Longer Sessions
◦ 11% more minutes watched per viewer
Happier Customers
◦ “I can’t believe how much time I spent playing with it. Fun.”
◦ “It made Shark Week interesting for my kids.”
Team Recognition
◦ Shark Explorer received a Shorty Award and was featured at a Google developers conference for its innovative experience