Improving listings' localization experience
Airbnb
CASE STUDY - Winter Release 2021
Airbnb
a cornerstone of the digital economy
Revolutionizing the travel industry by connecting travelers with unique accommodations worldwide. Through its platform, Airbnb facilitates the sharing of spaces in over 65,000 cities, empowering hosts and offering travelers authentic experiences, thus significantly impacting global travel and local economies.
The challenge
20% listings had
mixed languages
making the AI translation engine ineffective
We recommended a complete overhaul of the localization process, which boosted machine translation performance by 20%. Back then, Airbnb’s automatic translations couldn’t correctly differentiate between source and target languages, due to the mixed-language content in listings. By revamping the Localization Journey, we made it easier for hosts to control and refine their own translations, leading to clearer and more accurate listings for travelers worldwide.
Translated
Worldwide renowned trailblazers in machine translation
Translated was established in 1999 by a linguist and a computer scientist with an investment of around $100, which was used to purchase the internet domain and the first online advertising.
Professional translation services made easy. Crafted by expert humans, powered by technology, efficiently delivered.
Our solution
localization journey mapping
user testing & validation
Hosts have a better localization awareness
Improved UX
And there's a more inclusive tourism
Our main insights, discovered with research activities
A design team working with their design team
WHAT WE DID
(big) data driven recruitment powered research
I led a team of 3 designers in NOIS3, working together Airbnb User Research Lead and Data Scientists to select, recruit and interview Airbnb hosts, to learn if and how they would use the localization features in Airbnb to translate listings in different languages.
Â
Expert review
Secondary research
Usability Testing
Journey Mapping AS-IS and TO-BE
HOW THE PROJECT WORKED OUT
Activities and Deliverables
Fun fact
Recruiting hosts implied their Data Scientist running a query on their hosts' database with filtering criteria we would've spent days in searching.
Data-Driven Recruiting and Research set-up
"Sean please could you extract hosts matching this screener?"
"There you go, 543 hosts to select from."
This was a game-changer for the team: the first really Data-Driven Recruiting we faced;
accurate, fast, reliable.
Expert Review
Our product team worked in collaboration with Translated team, to isolate and spot the main usability issue during the whole listing publication process.
KEY TAKEAWAY
Even great design teams happen to overlook important parts of the journey and that's ok (if you fix it later).
Usability Testing
We validated our hypothesis observing recruited hosts in the process of creating a localized listing.
There was no clear guidance in the UX process that would lead hosts to create localizations for listings.
KEY TAKEAWAY
Hosts hacked the process, inserting mixed language messages, to achieve their goal: despite Airbnb's intentions they succeeded to deliver the message.
Redesign recommendations & prioritization
The joint product and Translated team, suggested to Airbnb Design Team the key interventions to improve hosts' localization experience.
KEY TAKEAWAY
(Re)mapping new insights in an established user journey leads to previously hidden gems and undisclosed opportunities.
Do you like what you see?
I can help
Made with
