Title	UBS PLANET ART
Brand	UBS
Product/Service	PLANET ART APP
Category	A08. Data Insight
Entrant	RAZORFISH Berlin, GERMANY
Idea Creation	RAZORFISH Berlin, GERMANY

Name	Company	Position
Preethi Mariappan	Razorfish GmbH	CCO
Claudia Trippel	Razorfish GmbH	CD
Yasushi Zonno	Razorfish	CD
Kristin Mitschke	Razorfish	UX Lead
Sonia Lago, Wyndham Mead	Razorfish	Strategy
Alexander Czernay, Jeremiah Lewis	Razorfish GmbH	Business Intelligence
Etienne Vles, Isabella Nogueira	Razorfish GmbH	Art Director
Alexander Thiele	Razorfish GmbH	Technical Director
Till Hellwig	Razorfish GmbH	Backend Developer
Jens Malunat	Razorfish GmbH	Account Director
Petra Czechowski, Marina Herranz, Julia Böhme	Razorfish GmbH	Project Management

The Campaign

The Idea: Your Guide to the Art World Planet Art brings art news, reviews and inspiration together in one simple news reader app. It is the world’s definitive source of art intelligence for collectors, enthusiasts and influencers. In the world of art, thousands of sources, articles and opinions threaten to overwhelm both experts and beginners alike. It’s a complex and noisy place. A great deal of time and effort is required to embrace, understand, and make the most of the art scene. Our solution was inspired by this fragmentation. Through the creative use of available data, we were sure a more efficient, objective and intelligent distillation of contemporary art news and information could be achieved. Unlike popular newsreaders and content aggregators that offer thousands of general topics, Planet Art focuses only on news and information from the world of contemporary art. It uses computer-vision driven web scraping technology.

Creative Execution

Planet Art scours a broad range of sources, from mainstream media outlets to specialized websites and blogs. Utilizing state-of-the-art NLP and machine learning, an intelligent algorithm searches and matches relevant tags such as artists, galleries, events, locations and more. A data-led similarity engine identifies clusters connections, and accumulations between tags, and presents these in a qualified and objective ranked stream of articles. The algorithm constantly fine-tunes the ranking, meaning there is always an up-to-date, comprehensive and valuable overview of the most important trending topics in the art world. UBS Planet Art was launched in beta for iPad in December 2014 to an exclusive audience in the UBS VIP Lounge at Art Basel Miami Beach. Since then, Planet Art has been released on additional platforms and a wider audience at other Art Basel events and activated through paid media promotion. Global release timeline: iPad, December 2014 iPhone, March 2015 Android (Tablet and Smartphone), March 2016 The latest release to Android devices in March 2016 solidifies Planet Art’s hold of the Asian market. In the future, Planet Art will be expanded into an ecosystem of intelligent digital tools for the contemporary art world, with the mobile app as a centerpiece.

Planet Art has been a huge success for UBS. As of Septmber2016, Planet Art has approximately 42,000 downloads for iOS and Android, an extraordinary achievement for a niche product. Our dedicated user base are actively using Planet Art: -10% conversion from download to active feature use per quarter ->1% conversion from download to active feature use per week -36% conversion from download to unique user, per quarter -21% of unique users personalized the app with their favorite artists, location & institution tags -Average usage time more than three minutes Top App Store Ratings: • 4.0 stars in the Apple app store • 3.9 stars in the Google Play Store An immense database of art knowledge: • Keywords scanned: 210.572.561 • News sources: 78 • Tags identified: 936.023 • Articles indexed: 328.236 • Personal tags set: 32.452 High volume partnerships • 222,700 unique visitors total • 355,000 total engaged minutes

The Planet Art sources range from mainstream websites to niche specialty publications. Each of our sources were selected through a process that weighed their salience to the art world, the breadth and volume of coverage, the completeness of the data available and the unique neutral-observer angle which UBS sought for the app. For each of these sources, a scraper was adjusted to site-structure and article templates and fine-tuned to extract all of the highest quality content. For an especially important source, The New York Times, their content API was custom-built into the app. Each of these sources have their own styles and formats; to accommodate these differences, both manual tweaks as well as machine learning training were required to achieve optimal performance.