Get our product and content more discoverable for people who look for travel inspirations online, e.g. where to eat, things to do, etc.
One year ago, we launched Airbnb guidebook for 16 popular cities, with 9 categories of different types of popular places which are recommended by Airbnb local hosts.
In order to keep the quality of guidebooks, we don’t have plans to a) increase the number of categories (levels) for long-tail keywords, or b) launch guidebooks for other cities with fewer categories.
Therefore, we want to create a product which makes use of the all the host recommendations in our dataset and cover as many travel-related keywords for users looking for travel inspirations as possible, a.k.a. Top 10 [keyword] in [city name]. By doing this, we could also drive more traffic to Airbnb and higher brand awareness by ranking more pages.
The first thing I wanted to lock down is the user types and their needs. Given the situation that we are targeting at people who search for travel inspirations on Google, this project shares the same types of users, trip planners and on-the-go travellers, with existing guidebooks.
The typical scenario is like, Katie is planning a summer trip in San Francisco, she wants to buy some local gifts for her friends on her last day. Her flight is scheduled in the afternoon, so she has the entire morning to visit local souvenir shops. She needs to find a couple of souvenir shops which are close to the airport and open in the morning. So she searches “best souvenir shops in San Francisco” in Google, and hopefully an Airbnb page comes up first.
I chatted with the researcher and read through the usability testing report on existing guidebooks, trying to figure out what is doing well, and what could be improved.
Given the complexity of the keywords from host recommendation, I tried different search queries in Google looked into the pages that ranked high, such as “best hotels in NYC” on booking, “best korean BBQ in San Francisco” on Yelp, and “Best museums in NYC” on TripAdvisor, etc.
Based off of what I have found from the previous sessions, I started sketching rough ideas on paper. Here is a couple of questions that I thought about during the process.
We are providing local and authentic recommendations
I decided to highlight that the reviews are coming from local hosts, which helps Airbnb to be different from other UGC products e.g. Yelp & TripAdvisor.
How to simplify the page?
I saw the guidebooks are showing three host reviews per place, which causes information overload on that page. I brought this question to the PM and asked if we could only display one review, but got declined, because we are so late in the game that we couldn’t sacrifice the SEO ranking for a simpler design.
How to showcase hosts’ tastes and lifestyle?
I tried to add a short description for each host next to their reviews, but it makes the page even busier. So I decided to use a rollover widget to show host tastes and lifestyles, inspired by Medium.
Bringing the design to our weekly critique meeting, I put them up on a whiteboard and invited other EPIC peers on the team to put themselves into the shoes of a real user, walk through the whole process, and use post-its to comment.
“What if I want to explore other keywords, is there a page for all of them?”
Initially I was thinking about creating a sitemap for all the keywords, but after a conversation with an engineer who pulled all the keywords, which is about 2k, I realized that it is impossible to create an user-friendly page with 2k keywords in there and people still use it.
So I decided to add a “related keywords” section to the bottom of the page after users browse through all the recommended places. On top of that, there is also a section for popular keywords in the city that users are searching for.
“What if I want to see more reviews?”
Each place in the host recommendation dataset has a product details page, a.k.a place PDP. In v1, I was planning to direct users to place PDP if they want to see more reviews. However, the fact is that place PDP doesn’t have any extra information compared to the main page except more host reviews. In this case, we decided to temporarily popup a modal to display all the host reviews for the sake of efficiency, and link back to place PDP when Trips team finishes updating place PDP in the near future.
In addition to the design feedbacks that I mentioned above, I also received a couple of comments on the content. An online marketing manager pulled out a report of frequent search terms that are SEO friendly, and recommended us to optimize the content, so I set up a working session with the content strategist and the OMG manager to make sure the copies are both user friendly and SEO valuable.
We are expecting to launch the MVP version in November, and will keep iterating on it while we keep track of the data and measure the impact.