GA UXDI Project 2 Retrospective

I can’t believe 3 weeks have passed already; time is flying! We’ve reached the end of Project 2, our first group project, in General Assembly’s User Experience Design Immersive course. My fellow group members were Erika Friend, Pablo Papasidero, and Manan Shah.

Our initial assignment was to design a new feature for the Airbnb app. With this open-ended goal in mind, we sent out a screener survey to identify users of Airbnb, and ideally to pinpoint issues that arise from use of the service. We received 99 responses, and the next morning we began to plan to identify unmet user needs. We made a mind map to hone our questioning around issues that could affect primary users of the service (those actually booking on Airbnb) as well as secondary users (children, travel companions, neighbors, property owners, etc).

Drawing connections, brainstorming, and collaborating!

We soon received a project brief that ‘Airbnb stakeholders’ (our instructors) wanted us to focus our research efforts on a new feature to allow venue hosts and guests who want to rent space for events to connect with one another.

This new direction fueled our interviews, which we narrowed to survey respondents who had experience with Airbnb or similar services, and who had thrown events at venues. With our new goal in mind, the team devised a set of initial questions for interviews, with room for individualization. Users tend to provide the most interesting insights during comfortable conversation. Our 9 interviews were illuminating, and quotes and attributes of planning gleaned from our questioning helped to spur our next phase of research: synthesis.

Using affinity mapping, or grouping common themes we saw throughout our interviews on differently colored sticky notes, we were able to identify important user needs, pathways, likes, dislikes, and experiences that arose in the venue searching and booking process.

We went through several iterations of our affinity map, organizing and reorganizing until we were satisfied with our categories. This process helped us to determine which areas and problems were most important to focus on. After a while Pablo started to sprout sticky notes out of his arms, but luckily he recovered quickly.
The proud team reveals our (gorgeous!) final affinity map.

We were quickly able to discern that there were 2 primary user types: those planning professional events (conferences, meetings, job fairs), and those planning personal events (weddings, birthdays, parties). We began to trace user journeys for these proto-personas.

Our “professional” persona’s user journey, and a few preliminary bullet points about her.
The user journey for our “personal” persona was concerned with comparing venues based on feel, and contact with venue hosts.
It may have been 9:30 pm after a full day of work and class, but this team kept smiling.

We began to focus on specific traits, desires, and problems of our personas, and developed them fully. Every step of this process was informed by our interviews. It was important for our professional persona to be able to visualize a space without having to visit it, and our personal event persona wanted to be able to find the right vibe without hassle.

Although they would use Airbnb to book venues for different purposes, the user flows led us to a problem statement that applied to both Rachel and Nancy.

People planning events are busy and usually search for venues online or get recommendations from people they know. Rachel finds the event planning process time consuming and stressful. How might we help her to efficiently find and book event-appropriate venues?

Contemplating this opportunity, we executed 2 run-throughs of the Design Studio process for each persona, in which we compiled our best ideas based on what they could do for Rachel or Nancy.

Our group’s first round of sketches focusing on Rachel’s needs. We began by broadly imagining what a solution could look like, before focusing on our best received ideas in the next design studio round.
In our next round, we incorporated constructive feedback from our group, and made our good ideas better. My screen (far left) focused on a map feature, and also offered suggestions of places a user may like based on their searches.

We prioritized features to best suit user needs and business goals.

We agreed that media and events on the homepage were essential, and searching by attribute and a map would be important as well.

Once we had devised our paper prototype, we performed 2 usability tests. Our task scenario was as follows:

In this scenario you are Rachel Black, a Graduate Student Researcher trying to plan a student showcase, and you need to find a venue for the event.
You will have two tasks.
1.The first task is to use Airbnb to find an available venue you would be interested in and favorite it.
2. Now that you’ve favorited the venue, please go ahead and contact the host.
Our first iteration of paper wireframes. I drew three separate screens instead of one long scrollable screen (bottom three screens) for the venue page, which made users think that they could not scroll on this page, which was necessary to contact the host.

Users appreciated the straightforward design, but were confused by the map section and filtered search results.

In the next iteration before our digital sketches, we cleared up the usability issues we had identified. I drew the venue page as a single long screen, and I tried to make the graphs in the Filters section clearer. We drew strongly from Airbnb’s existing design scheme, which is beautifully simple.
Low-fi prototype testing.

We next translated this paper model into a mid-fidelity digital prototype, integrating insights from our interviews with business goals. Our feature brief specified that hosts and guests should be able to identify one another based on mutual interests, so we paid special attention to searchable “tags” and themes for each venue, allowing hosts to expand their repertoire of event types in their spaces, and guests to find the perfect venue for events.

This is one of the screens Pablo designed as a search result page, filtered by date.

In our first round of mid-fi usability testing, we presented users with the same task scenario as before. I acted as the note-taker and an observer for each usability test, Erika and Manan acted as testers, and Pablo observed user behavior. We discovered that users had trouble remembering their task, and did not understand the graphs on the filtering page. They learned that they could not scroll on early pages, so they felt they could not scroll on the subsequent venue page, which was necessary to access the “Contact Host” button at the bottom of the page. Users used the “Book Now” button located at the top of the page to complete this task, despite mistrusting it intensely. Users did not trust the sponsored venues that appeared atop the search results. I assembled user metrics, as pictured below.

Metrics from Usability Testing round 1. We identified some issues for users, such as the feeling that they could not scroll on the final venue page, and the need to click on the book now button despite not liking the idea of booking immediately and not being able to go back.

In our second round of testing, we split our first task into two to decrease confusion. We had much better success rates with scrolling, and users seemed to better understand the filters page. It seemed that users were satisfied to click on a venue without filtering, which was not what we intended. In the next iteration, we required users to filter to ensure an event-appropriate venue. We changed the name of advertised events from “sponsored” to “featured,” hoping that users would react more favorably. (Spoiler: it didn’t help).

Metrics for Usability Testing round 2. We had 100% scrolling on every page, and 75% using the “Contact Host” button! Only one user wanted to select a featured venue, which led us to believe the name change was not enough.

After addressing these issues with mostly minor redesigns, we developed our hi-fidelity prototype in Sketch and InVision, executed masterfully by Pablo. We strove to better achieve our business goal of allowing guests and hosts to be able to easily identify one another by adding an optional checkbox at the bottom of the filters page allowing hosts to contact guests within a limited time window, and by promoting venues with “stories,” a less pushy advertising strategy (both courtesy of Manan).

Check out our prototype below!

I would love for this feature to include more functionality, and to address more user needs that we had to de-prioritize for this iteration. I hope to conduct more tests to increase the efficacy of venue advertising, hitting the sweet spot between user desires and company bottom line.