As a designer at an architecture firm based in the Bay Area, my passion for crafting unique spaces stems not only from my work but also the belief that how a space is designed significantly influences the emotions and experiences of those within it. Inspired by my architectural background, I redesigned Airbnb's mobile app to optimize the chaotic, time-consuming decision making process guests often experience when planning a trip. I streamlined the various steps involved in trip planning and integrated AI/ML to tailor recommendations for guests.
Guests have to do too much due diligence to evaluate a listing’s desirability—they have to find nearby attractions on Google Maps, coordinate with their travel group over external platforms like iMessage, and constantly revisit the filters to balance tradeoffs. A complex and lengthy planning process often leads guests to miss out on desirable listings and in some cases, abandon the booking process altogether.
Providing more booking context about location insights will allow guests to engage in faster and more informed decision-making. We can also improve the relevancy of listing results by including other preference factors, such as the guest's desired activities for their trip and their mode of travel. When guests decide to travel in a group, providing collaboration tools will help groups come to a consensus quicker and more efficiently.
User research indicates that a bleak and lengthy trip planning process often leads guests to go to another competitor like Expedia or abandon the booking process altogether. If Airbnb guests can quickly find tailored listings and balance the tradeoffs, they are more likely to proceed with their bookings leading to higher conversion rates and increased revenue for Airbnb. When guests don't have to resort to external tools to research a listing, they have more time to explore additional Airbnb Experiences and activities surrounding their trip. This creates more opportunities for the company to upsell and increase the average booking value.
Not inclusive of guest preferences like desired activity and mode of travel, resulting in broad search results
An AI-driven, guided search journey helps you customize your trip by showing relevant listings and their trade-offs based on your preferences.
Difficult to pinpoint & evaluate a listing's location, requires map navigation & external research
Listings are equipped with a Location Card that provides approximate location & insights about Safety, Convenience & Things to Do. Guests can access location-based reviews and explore nearby Airbnb Experiences.
From the Location Card ratings, guests can access more specific location insights on the Safety, Convenience and Things to Do Page. I prioritized ethical design by creating a 'Safety' page featuring safety tips and relevant, safety-focused reviews, consciously avoiding the use of crime data to eliminate the biases often associated with public safety metrics.
Passive and fragmented, requires external sharing + collaboration
Ability to see at a glance which travel group members preferred an Airbnb Listing or Experience
I conducted 5 usability tests to evaluate the existing trip planning process. Guests navigate Airbnb's limitations by opening multiple listings in separate web tabs for comparison, or by transferring key details to Google Docs or Sheets for group sharing.
The primary planner shoulders the bulk of trip planning and group coordination. Recognizing this, I chose not to dilute their role, considering the risk of inviting potential group conflict. Instead, to prioritize expediting the decision-making process, I focused on simplifying the organizer's tasks.
While designing the 'Safety' page to demonstrate how safe a listing's location is, I initially considered including crime statistics from a reliable third-party vendor. Subsequent research into redlining and racially restrictive covenants indicated a greater risk of reinforcing future bias, leading me to opt for host-provided safety recommendations & relevant safety-focused reviews.
I developed strategies to simplify trip planning by refining the search process, integrating AI/ML in various ways to provide smarter search results and contextual insights. To enhance decision-making in travel groups, I experimented with levels of authority tailored to the dynamics of trip planning.
After conducting two rounds of usability testing with five participants, it became clear that users highly value location insights provided for each listing, particularly how these insights inform their decision-making. To enhance this experience in the next iteration, I reinforced the rating categories with dedicated landing pages for Safety, Convenience and Things to Do. Users reported that the detailed insights diminished their need to use external tools such as Google Maps and Yelp, making it easier to evaluate a listing's location directly within our platform.
The first round of usability testing indicated that users valued personalizing their search but found the effect of their choices on the search results unclear. To remedy this, I introduced a personalized breakdown for each listing, clearly illustrating the tradeoffs based on their selections. This improvement in transparency greatly increased users' confidence in Trip Tailor's capacity to fine-tune search results according to their preferences.
The final design is implemented in stages to balance user/business impact and engineering work, considering AI/ML integrations:
1. Provide context about listing's locale
User research indicates that lack of context about a listing's location is the primary concern.
2. Add more depth to location insights
If guests are actively engaging with the Location Card, we will provide more context through the Things to Do, Convenience and Safety pages.
3. Implement Likes in shared wishlists
To ease the burden on the primary planner, the Like feature aims to encourage efficient collaboration and minimize group conflict.
4. Optimize search with Trip Tailor (AI/ML)
If previous milestone metrics indicate the need for more context, we can address the issue from a wider angle by using additional criteria to provide smarter, personalized search results.
5. Refine prompts and balance tradeoffs
Finally, we can leverage user engagement with the Trip Tailor feature to further refine prompts and integrate smart feedback in search results.
Future AI/ML integrations not only allow us to provide guests with an intelligent, guided search experience but also to anticipate their desires ahead of time. By analyzing guest preferences, search activity and previous stays we can curate the ideal experience for a particular guest, eliminating the need for them to search at all. Airbnb is committed to evolving into an all-encompassing travel platform, focused on simplifying each aspect of trip planning. The ultimate vision is to offer an experience akin to consulting a travel agent, where every travel need and inquiry is met effortlessly in one unified, accessible destination.
I discovered that crafting a successful product involves recognizing current norms and progressively steering them towards the ideal state through incremental steps. Although it was difficult at first to identify the root cause of user frustrations, I found myself improving as a designer by asking deeper questions to users to illustrate the bigger picture. Throughout the course of this case study, I realized the importance of tying my design decisions back to my user research, especially when articulating and defending my ideas during group discussions. My biggest struggle was devising UX copy, specifically in crafting clear and concise messages that effectively conveyed a new feature to users. By relying on generative tools like ChatGPT and drawing inspiration from popular applications, I was able to balance clear comprehension while maintaining simplicity in my design.
