What’s in a Word?
Using location context to build better search tools for travel
Valerie, age 28, was searching for a place to stay with her sister on their first trip to New York City. She describes her ideal location as, “Fun, close to restaurants, lots of energy and easy to get to public transportation.” She wanted to explore the sites, eat at trendy restaurants and feel like a New Yorker for the 5 days she would spend there.
We know from research that 43% of our Airbnb travelers, like Valerie, fit in a category we call “explorers.” For these travelers, the neighborhood vibe and personality of their accommodations are a top priority. While she carefully read host descriptions of Airbnb listings, Valerie ultimately needed to visit other sites and learn more about New York City neighborhoods before making a decision about where to stay.
Janet, on the other hand, is a 56 year old experienced traveler, who chose an Airbnb in London while pursuing a tour of European museums. “My Airbnb was lovely,” she says, “but I wanted to be somewhere more quiet and quaint. The area I was staying in was hip and loud– great for someone else, but not for me.”
Location: More than just a pin on a map
After interviewing hundreds of travelers and surveying thousands more, a resounding theme emerged: Understanding location context is crucial to an easy search and positive stay. Guests want to know more than a listing’s amenities– they want to know about the location itself. We brought participants like Valerie and Janet into our research lab for interviews, and observed them using early prototypes to help our designers, engineers, content writers and data scientists learn from their feedback. We found that for many travelers location is so much more than an address; it’s about the qualities of a neighborhood and how they fit with each traveler’s needs and personality.
This challenge of understanding neighborhoods is especially unique to travelers on Airbnb. Prior to Airbnb, most travelers booked their accommodations in city centers closer to downtown neighborhoods. Today, 74% of Airbnbs in a dozen major cities around the world are located outside of downtown neighborhoods, which means that guests are able to stay in places not usually included on tourist maps. In this map of San Francisco, there is no Haight Ashbury, Mission or Outer Sunset– all places loved by locals. Because of this shift, there’s a need for travelers to better understand these neighborhoods when they’re looking for places to stay. With this in mind, we had two questions: How can we incorporate neighborhood content into our search flow, and what neighborhood attributes are most important to guests?
Luckily for us, members of our content team had already written a wealth of descriptions for neighborhoods in 23 cities. The content currently lives on the Airbnb neighborhood page but had never been fully integrated within our search flow. In other words: We had all of this great neighborhood information on our site, but there was no direct way to see it while searching for a place to stay, which made it difficult to use. Utilizing this existing content was no easy feat, though. First we needed to re-imagine what it means to search on Airbnb.
Asking upfront for the best results
Our designers and researchers led countless brainstorms and design proposals before landing on a final solution: Including a question that asks travelers about their neighborhood preferences upfront before showing them their results. In this way travellers are immediately able to select neighborhoods that match their preferences, without having to sift through neighborhoods that don’t. Now guests can explore and compare multiple neighborhoods using quick descriptions to help them make decisions about how to narrow their searches.
What’s in a word?
Matching neighborhood recommendations in the results meant that we needed to define and collect neighborhood preferences up front. Feedback from our initial interviews showed that current search terms were too generic, and led us to conduct a survey where we asked travelers searching on Airbnb to define their ideal neighborhood in a single word. After receiving over a thousand responses, we categorized the words into themes. We found that the most common words used to describe ideal neighborhoods fell into themes focused on location (e.g. “convenient,” “central,” “walkable”), energy level (e.g. “quiet,” “vibrant,” “fun”), and the type of traveler (e.g. “family-friendly”).
Participants preferred neighborhood descriptors that were polarizing, because those words helped them make clear-cut decisions over vibes that were generic.
Over the course of three months we held nine rounds of one-on-one interviews with over 50 trip planners. These interview sessions were iterative, meaning that we were able to relay the feedback from each round of interviews to the design team, who then incorporated it into future prototypes. In addition to research interviews in San Francisco, we also traveled to Denver in order to get feedback from a broad range of travelers outside of Silicon Valley. It was through these interviews with many different types of travelers that we gained an additional learning: Travelers don’t want to choose a neighborhood based on words they consider generic.
For example, “trendy” was considered a word that could apply to almost all neighborhoods to some extent, and didn’t tell people enough about a particular neighborhood. Because of this, participants preferred neighborhood descriptors that were polarizing, because those words helped them make clear-cut decisions over vibes that were generic. For example, “quiet” is a useful word that describes a particular quality that guests either want or don’t want, whereas terms like “trendy” and “artsy” give guests less specific information about an area.
Some neighborhood descriptors were also missing location context – does the word “dining” mean that a neighborhood is close to many restaurant options, or that there are simply a few restaurant options within a few miles? Can you see a pattern here? Location is really important to travelers.
With this feedback, we added location context to otherwise ambiguous search tags, such as “close to restaurants” and “close to nightlife.” We also added pairs of words that describe specific opposites, like “lively” and “quiet”, to make it easier for travelers to choose one or the other rather than generic terms that might apply to a wide variety of neighborhoods.
Next steps and challenges
Although our team has undergone multiple design and content iterations, incorporating user feedback along the way, this doesn’t mean we’re done iterating. This implementation of searching via neighborhood preferences upfront is an early attempt to help travelers like Valerie and Janet search more easily and establish more accurate expectations regarding location. As we learn more from the data we will be able to continue improving this feature.
Finding the right neighborhood for you
After many iterations, we have settled on a final set of neighborhood descriptors which will be presented to travelers upfront, before they see their search results. Now, guests like Valerie who want to stay in locations that are, “fun, close to restaurants, lots of energy and easy to get to public transportation,” will see listings located in specific neighborhoods based on their preferences on mobile. Janet, and travelers looking for more quiet locations, will have more clear neighborhood expectations when they arrive at their destinations. We hope the ability to become familiar with a neighborhood from the start will instill a greater sense of belonging. Staying at an Airbnb is not just about having a place to sleep– it allows guests to feel like they live there. When the place you chose to live in is in a neighborhood you enjoy, the sense of belonging during a trip is even greater. We hope these neighborhood tags enable a better “Live There” experience for travelers around the world.