Sparrow Explore Tab Case Study

Project Brief • Solution • Challenges • Guiding Decision Making • Defining Fun • Creating Useful Tags • Activity Dimensions • Overfiltering Prevention • Afterword

PROJECT BRIEF:

Help young adults find personalized opportunities to go out and interact with other people.

Investigating Key Terms:

1. FIND: How should Sparrow users search for and discover experiences?

2. PERSONALIZED: Why do some people enjoy certain experiences and not others?

3. OPPORTUNITIES: What makes experiences similar?

SOLUTION:

How to Discover Experiences?

By Interest (art)? By Action (paint)? By Location (museum)?

The Sparrow Users Take

1. Activity vs. Atmosphere

   To get the ball rolling, users choose between these two options. Are they looking for things to do or a vibe? This keeps the start of the decision-making process as easy as possible.

2. Describe the activity

   Search parametrically by selecting tags that represent the activities they want to engage in.

3. Or atmosphere

   And/or select some to describe the environment they want to visit.

4. Refine the results with the four features listed below

   UI updates based on the chosen tags to train the users to associate these tags with these axes. Users can adjust the dimensions as needed.

Why Do People Prefer Certain Experiences?

Personality. Exhaustion. Willingness to engage in some, if any, activities. Accessibility and tolerance for certain environments.

What Makes Experiences (Activities) Similar?

For activities to be recommended to users, they must first be comparable. To make these comparisons, I deconstructed what makes an activity enjoyable or draining into four dimensions:

1. Immersion

2. Social Interaction Level

3. Mental Challenge

4. Physical Challenge

Impact:

This dimensional approach allowed us to construct the feature vectors for experiences for the AI recommendation system. With them, we had a framework for defining and comparing activities⁠⁠.

CHALLENGES:

1. Create confident satisficers.

   • Limit aimless scrolling that’s commonplace with social media apps.

   • Avoid choice overload.

   • Limit decision fatigue.

2. Creating an exhaustive list of tags that:

   • Captures what users find fun.

   • Can be used to filter the available experiences.

   • Isn't excessively long.

3. Crafting unbiased copy that captures both ends of each spectrum without negative connotations.

   • E.g., Silly (not demeaning like stupid or braindead) to Academic (not too pretentious or remote) for Mental Effort.

GUIDING DECISION-MAKING:

I designed the first iteration to create a structured, step-by-step process for finding fun activities. The goal of this step was to limit scrolling and help users make quick decisions. The longer they take to decide on something to do, the less likely they'll even leave home.

As users select filters, new ones appear, posed as questions. With each decision, each question becomes more refined and cognitively difficult.

User Decisions:

1. When and Where

   • Easy logistical questions

2. New or Familiar Experiences

   • Two choices: Explore or Check History

3. "Chill" or "High Energy" Experiences

   • More complex and subjective

4. Fun Tags as breadcrumbs

   • Filter what's left based on what users find fun.

   • Like Reddit’s onboarding flow, selecting interests reveals a new set of related interests (E.g., Sporty → Basketball, Soccer, Hiking)

     
     

Other steps considered

• Social battery

• Something popular/hidden gem (social proof vs. individuality)

DEFINING FUN:

Coming from a game development background, I realized I was assuming what "fun" meant to users. While Sparrow focuses on activities outside the house, a ludic perspective on defining fun proved helpful.

Using the ludic taxonomy:

Marc LeBlanc’s Eight Kinds of Fun:

Sensation, Fantasy, Narrative, Challenge, Fellowship, Discovery, Expression, Submission

• Fellowship inspired the Social tag

• Expression → Creative tag

• Discovery/Submission influenced the inclusion of the New vs. Familiar Experiences step in the first iteration.

Roger Caillois’ Play Categories:

Agon (competitive), Alea (luck), Mimicry (roleplay), Illnx (sensation)

• Agon → Compete and Collaborate activity tags

Takeaways:

New initial tags for describing events (a sporty event; a creative meetup)

CREATING USEFUL TAGS:

Why Are Tags "Activities" rather than "Interests" (Fun Tags)?

Other apps use interest-based tags because the activity is already set:

• Reddit - read/write about (games, the news)

• Spotify - listen to (pop, rock, hip hop)

• Sparrow - do/experience… experiences

Because Sparrow's main goal is to get users to do things, users are encouraged to search for experiences based on activities they want to participate in.

Someone interested in art might want to visit a museum or paint. But, someone who’s interested in painting might not want to visit a museum.

The Search Bar vs. Filters:

However, because people might want to search by interests, they can, but it's discouraged. Using the search bar, they can search by interest, location, activity or atmosphere, but the Quicktype suggestions are set to recommend activities related to the search term.

It’s at the bottom of the screen rather than the top to implicitly guide users to focus on the leading filter questions and top bar first.

Why do users select atmosphere tags?

One interview that stuck out to me was with an introverted homebody. When asked about what he finds fun on vacation, he spoke more about the atmosphere of locations and who they might interact with.

Based on LeBlanc’s eight types of fun, this is sensory pleasure. Take two users who enjoy listening to music. One user might love events like concerts and nightclubs, whereas the other might find these events overstimulating, even if they like the music being played.

Considering this, I created a separate group of tags to describe experiences divided into nine categories:

ACTIVITY DIMENSIONS:

What Makes an Activity Draining?

Immersion/participation level

• Does this require participation?

• Is the user looking for a place to relax or engage with their interests?

Social interaction level

• Does this require talking?

• Is the user looking for an independent space with like-minded individuals, or are they feeling chatty?

Mental challenge

• Does this require thinking or creativity?

• Is the user looking to goof around or looking to challenge themselves intellectually?

Physical challenge/activity

• Does this require movement?

• Is the user exhausted after work or looking for a workout?

Atmosphere/vibe

• Who's there?

• What's it like to be there?

Relevance to Interests

• Is this related to the user’s interests?

This list of features came after a few qualitative interviews, card-sorting exercises and personas constructed based on these conversations. This research helped determine how an activity might be draining, tolerable or enticing.

Immersion As an Axis:

Immersion refers to how much participation an activity requires. When users set immersion high, the search prioritizes activity tags. If they prefer more passive experiences and set immersion low, atmosphere tags carry more weight in recommendations.

OVERFILTERING PREVENTION:

What Happens If There Are No Results?

Users are presented with four options for how to proceed and then a list of events that almost fit their query.

Find something random

• Based on the “I’m Feeling Lucky” button.

• Returns the best-fit recommendation based on broader terms. This ensures users receive relevant results, even if their original query was too specific.
 E.g., A Paint tag collapses into Create; Dance → Exercise)

Edit filters

• Users can review and remove any of their selected filters.

Create their own event

• Users can create a new event based on the filters they selected. Sparrow isn’t just a tool for finding experiences. It’s also for shaping them, which adds long-term engagement.

View their recent searches

• For users who tend to explore and tweak settings a lot. Sometimes they don’t realize they already searched for something close to what they want.

• Helps people quickly retrace their steps instead of starting over.

These options are also available at the end of each page of search results. They’re always present to keep users mindful of their browsing time (like Netflix’s Are You Still Watching?) and the other options available to them.

AFTERWORD:

Working on this feature, I drew from a broad spectrum of knowledge and interests. I already mentioned how game design played a role in this project, but developing the dimensions for describing experiences also reminded me of the various personality type inventories I’ve explored in the past. It was also directly inspired by how Spotify structures its personalization—particularly its use of perceptual audio features like energy, danceability and valence to group and recommend music.