A search engine for Bay Area theatres.
From idea to deployment.
Role
Role
Collaborator
Collaborator
Status
Status
Stack
Stack
End-to-end product designer
Vera Belova, founder
Live · phase 1
Framer · Airtable · Vercel · GitHub
Framer · Airtable ·
Vercel · GitHub
Background
Murmur started with a conversation. I met Vera Belova and complained to her that there were no good theatres around. She - a theatre advocate with 20+ years in the Bay Area - pushed back: there are plenty, you just can't find them. We were both right.
That exchange led us to a real gap: no platform in the Bay Area aggregates local theatre with searchable criteria for audiences. People were Googling, checking newsletters, following word-of-mouth. They didn't see the whole theatre picture. As it turned out, Vera had already been quietly solving this - she'd started collecting every local show into a single Airtable base. When I suggested we build something on top of it, she said it had been her plan all along.
That's how Murmur started.
From there: user interviews and personas, lo-fi prototypes, usability testing, and a fully deployed search system - built in Figma and Framer, wired to Airtable, and launched.
Murmur started with a conversation. I met Vera Belova and complained to her that there were no good theatres around. She - a theatre advocate with 20+ years in the Bay Area - pushed back: there are plenty, you just can't find them. We were both right.
That exchange led us to a real gap: no platform in the Bay Area aggregates local theatre with searchable criteria for audiences. People were Googling, checking newsletters, following word-of-mouth. They didn't see the whole theatre picture. As it turned out, Vera had already been quietly solving this - she'd started collecting every local show into a single Airtable base. When I suggested we build something on top of it, she said it had been her plan all along.
That's how Murmur started.
From there: user interviews and personas, lo-fi prototypes, usability testing, and a fully deployed search system - built in Figma and Framer, wired to Airtable, and launched.
What I did
User Research · 5 Interviews
User research · 5 Interviews
6 personas + journey maps
6 personas + journey maps
Wireframing
Wireframing
Lo-fi Prototype
Usability Testing · 5 users
Usability Testing · 5 Users
Front-end Development
Front-end Development
Built with
Front-end Development
Figma · Framer · Airtable · GitHub · Vercel · JSON / API
Front-end Development
Built with
Figma · Framer · Airtable · GitHub · Vercel · JSON / API
User Research and Analysis
Five interviews. Three universal blockers across every user type.
Five interviews. Three universal blockers across every user type.
1.
Nobody has one reliable source
1. Nobody has one reliable source
Everyone Googles or follows word-of-mouth. Eventbrite shows general listings, not curated local theatre. There is no specialist platform for Bay Area audiences.
Everyone Googles or follows word-of-mouth. Eventbrite shows general listings, not curated local theatre. There is no specialist platform for Bay Area audiences.
2.
Criteria matter but can't be filtered
2. Criteria matter but can't be filtered
Users want to filter by date, county, format, audience age, and production level - but no tool offers this for local theatre. The result: frequent disappointment and poor attendance.
Users want to filter by date, county, format, audience age, and production level - but no tool offers this for local theatre. The result: frequent disappointment and poor attendance.
3.
Trust is the deciding factor
3. Trust is the deciding factor
A credible, specialist resource would shift user confidence. All five participants said they'd trust a dedicated platform over a generic search result when choosing a play.
A credible, specialist resource would shift user confidence. All five participants said they'd trust a dedicated platform over a generic search result when choosing a play.
User Persona Creation
6 personas multicultural and multifaceted by design were created based on 5 interviews, theatre observation, and domain knowledge.
Two segments were identified: quality-seekers and family&social planners.
Quality seekers are interested in high production level, specific genre, SF-focused. Family&social planners are looking for some age-appropriate, nearby, date-driven shows.
6 personas were created.
Two segments were identified: quality-seekers and family&social planners
Composite archetypes built from 5 interviews, theatre observation, and domain knowledge - multicultural and multifaceted by design. Two primary segments shaped every search criterion in the product.
Maya Fridman ↓
Maya Fridman ↓
54 · SF · Cultural manager
54 · SF · Cultural manager
William Harris
61 · San Carlos · Retired professor
61 · San Carlos · Retired professor
Laila Moreno
22 · SF · Architecture student
22 · SF · Architecture student
Christin Koller ↓
39 · Mountain View · IT wife
39 · Mountain View · IT wife
Zoe Eglstone ↓
Zoe Eglstone ↓
36 · San Jose · Product manager
36 · San Jose · Product manager
Vikesh Nakhar
31 · San Jose · Cloud engineer
31 · San Jose · Cloud engineer
↓ expanded below — best represent the two primary segments that shaped the product's search criteria
↓ expanded below — best represent the two primary segments that shaped the product's search criteria
SF-cultural manager / Quality seeker
Maya Fridman
54 · San Francisco · Laptop
Subscriptions: SF Jazz, SF Opera, SF MOMA
GOAL
Experimental, high quality theatre by notable directors. SF venues only - unless something extraordinary goes elsewhere.
FRUSTRATION
"There are so many events in SF that I don't see any reason to go out to the Peninsula. Only if something truly extraordinary is happening."
DRIVES THESE FILTERS
County → SF
Production level → Major / Regional
Format → Play / Immersive
IT wife / Family planner
Christin Koller
39 · Mountain View · Tablet
2 kids (elementary + middle school)
GOAL
Find an entertaining, spectacular show for the whole family - within 30–40 min drive, this weekend or next.
FRUSTRATION
"I value the time we spend as a family. Our routine feels repetitive and I want to brighten it, but I usually just pick popular musicals or events other parents recommend."
DRIVES THESE FILTERS
Audience → Family friendly
County → Santa Clara / San Mateo
Format → Musical
Working specialist / Social planner
Zoe Eglstone
36 · San Jose · Laptop
Product manager · plans spontaneously
GOAL
Find something interesting to do with friends on a free evening to experience together and discuss afterwards.
FRUSTRATION
"I'm not a theatre connoisseur, so I usually choose something on Google. About 80% of our choices end up disappointing us!"
DRIVES THESE FILTERS
Date → this weekend
Audience → Adult
Vera's resume → trust signal
SF-cultural manager / Quality seeker
Maya Fridman
Maya Fridman
54 · San Francisco · Laptop
Subscriptions: SF Jazz, SF Opera, SF MOMA
GOAL
Experimental, high quality theatre by notable directors. SF venues only - unless something extraordinary goes elsewhere.
FRUSTRATION
"There are so many events in SF that I don't see any reason to go out to the Peninsula. Only if something truly extraordinary is happening."
DRIVES THESE FILTERS
County → SF
Production level → Major / Regional
Format → Play / Immersive
IT wife / Family planner
IT wife
/ Family planner
Christin Koller
39 · Mountain View · Tablet
2 kids (elementary + middle school)
GOAL
Find an entertaining, spectacular show for the whole family - within 30–40 min drive, this weekend or next.
FRUSTRATION
"I value the time we spend as a family. Our routine feels repetitive and I want to brighten it, but I usually just pick popular musicals or events other parents recommend."
DRIVES THESE FILTERS
Audience → Family friendly
County → Santa Clara / San Mateo
Format → Musical
Working specialist / Social planner
Working specialist / Social planner
Zoe Eglstone
36 · San Jose · Laptop
Product manager · plans spontaneously
GOAL
Find something interesting to do with friends on a free evening to experience together and discuss afterwards.
FRUSTRATION
"I'm not a theatre connoisseur, so I usually choose something on Google. About 80% of our choices end up disappointing us!"
DRIVES THESE FILTERS
Date → this weekend
Audience → Adult
Vera's resume → trust signal
Key Insight - Across All Personas
Users filter first by date, county, and format - then need a confident "why" to commit: Vera's resume, rich description, and production quality tags.
A specialist, curated platform builds trust that generic Google results cannot - all five interview participants said they'd trust a dedicated resource over a general search.
Quality-seekers drove the production level, county, and format filters. Family & social planners drove the audience age, date, and location filters. Every criterion in the product has a persona behind it.
Users filter first by date, county, and format - then need a confident "why" to commit: Vera's resume, rich description, and production quality tags.
A specialist, curated platform builds trust that generic Google results cannot - all five interview participants said they'd trust a dedicated resource over a general search.
Quality-seekers drove the production level, county, and format filters. Family & social planners drove the audience age, date, and location filters. Every criterion in the product has a persona behind it.
Two Decisions That Shaped The Whole Product
DESIGN DECISION #1
Vera's Column as a trust engine. Research showed users trust specialist curation over algorithms. We turned Vera's expertise into a structural feature, not a sidebar afterthought.
-> Articles and recommendations appear alongside every search result
-> Vera writes a "watching next 3 months" column - personal, not ranked
-> No star ratings: every play deserves equal visibility and a fair chance, plays display randomly
Vera's Column as a trust engine. Research showed users trust specialist curation over algorithms. We turned Vera's expertise into a structural feature, not a sidebar afterthought.
-> Articles and recommendations appear alongside every search result
-> Vera writes a "watching next 3 months" column - personal, not ranked
-> No star ratings: every play deserves equal visibility and a fair chance, plays display randomly
DESIGN DECISION #2
No ticketing aggregation - intentionally. murmur directs users to buy tickets directly from each theatre. This was a deliberate ethical and product choice, not a limitation.
-> "Buy Tickets" links to the theatre's own page — keeping prices clean
-> Vera doesn't want to add a commission layer on top of theatre tickets
-> Supports local theatres directly rather than extracting margin from them
-> Keeps murmur's focus as a discovery platform, not a marketplace
No ticketing aggregation - intentionally. murmur directs users to buy tickets directly from each theatre. This was a deliberate ethical and product choice, not a limitation.
-> "Buy Tickets" links to the theatre's own page — keeping prices clean
-> Vera doesn't want to add a commission layer on top of theatre tickets
-> Supports local theatres directly rather than extracting margin from them
-> Keeps murmur's focus as a discovery platform, not a marketplace
From Wireframe to Lo-fi Prototype
Usability Testing of Lo-Fi Prototype
Five moderated usability interviews were conducted to evaluate how participants interacted with the prototype, identify friction points in the user journey, and better understand their decision-making process. The findings informed key improvements in the high-fidelity prototype.
Five moderated usability interviews were conducted to evaluate how participants interacted with the prototype, identify friction points in the user journey, and better understand their decision-making process. The findings informed key improvements in the high-fidelity prototype.
"First thing I would check is location. Second - date. I prefer to see is the list of events fitting these two criterias and nothing else."
Mood - that's a new one. That's interesting.
I would like to know more about local theatres. What theatres do I have here in Palo Alto besides Stanford Live?
Usability Test Implications
1. Improve clarity of information organization → Redesigned the 3-column layout: search filters on the left, results in the center, Vera's column on the right - each with a clear role.
2. Keep filtering simple and avoid drop-off during selection → Filters apply in real time as users check boxes - no "apply" button, no separate filter page.
3. Prioritize date and location as the most important criteria → Date and County filters are placed first in the sidebar, above genre and mood.
4. Preserve Vera's column to strengthen trust through expert guidance → Vera's recommendations appear alongside every search result, not buried in a separate section.
5. Highlight the Mood criterion as a distinctive discovery feature → Mood filter (Uplifting, Emotionally Intense, Thought-Provoking) is surfaced as a unique differentiator no other platform offers.
1. Improve clarity of information organization → Redesigned the 3-column layout: search filters on the left, results in the center, Vera's column on the right - each with a clear role.
2. Keep filtering simple and avoid drop-off during selection → Filters apply in real time as users check boxes - no "apply" button, no separate filter page.
3. Prioritize date and location as the most important criteria → Date and County filters are placed first in the sidebar, above genre and mood.
4. Preserve Vera's column to strengthen trust through expert guidance → Vera's recommendations appear alongside every search result, not buried in a separate section.
5. Highlight the Mood criterion as a distinctive discovery feature → Mood filter (Uplifting, Emotionally Intense, Thought-Provoking) is surfaced as a unique differentiator no other platform offers.
Usability Test Implications
1. Improve clarity of information organization → Redesigned the 3-column layout: search filters on the left, results in the center, Vera's column on the right - each with a clear role.
2. Keep filtering simple and avoid drop-off during selection → Filters apply in real time as users check boxes - no "apply" button, no separate filter page.
3. Prioritize date and location as the most important criteria → Date and County filters are placed first in the sidebar, above genre and mood.
4. Preserve Vera's column to strengthen trust through expert guidance → Vera's recommendations appear alongside every search result, not buried in a separate section.
5. Highlight the Mood criterion as a distinctive discovery feature → Mood filter (Uplifting, Emotionally Intense, Thought-Provoking) is surfaced as a unique differentiator no other platform offers.
Designed it. Then built it.
With 250 records and no need for a custom backend, the right architecture was a JSON API bridge from Airtable through Vercel - wiring a Framer front-end to live data without writing a full server.
With about 250 records and no need for a custom backend, the right architecture was a JSON API bridge from Airtable through Vercel - wiring a Framer front-end to live data without writing a full server.
With 250 records and no need for a custom backend, the right architecture was a JSON API bridge from Airtable through Vercel - wiring a Framer front-end to live data without writing a full server.
Responsive desktop + mobile
Full responsive build in Framer - the 3-column layout adapts intelligently across screen sizes. The search experience is fully functional on smartphone.
Responsive desktop + mobile
Full responsive build in Framer - the 3-column layout adapts intelligently across screen sizes. The search experience is fully functional on smartphone.
Airtable → JSON → Vercel API
Wrote a JSON script to pull all play data from Airtable. Deployed through GitHub to Vercel. Connected to the Framer front-end via API - no custom backend needed for 250 records.
Airtable → JSON → Vercel API
Wrote a JSON script to pull all play data from Airtable. Deployed through GitHub to Vercel. Connected to the Framer front-end via API - no custom backend needed for 250 records.
Coded search filter logic
Custom code components for each filter criterion: date, county, format, audience, mood, genre, and production level. Filters combine dynamically to show only matching plays in real time.
Coded search filter logic
Custom code components for each filter criterion: date, county, format, audience, mood, genre, and production level. Filters combine dynamically to show only matching plays in real time.
65-gradient theatre identity system
No play posters or photos were available across 65 Bay Area theatres. Designed 65 unique two-colour gradient combinations - one per theatre - so every card is visually distinct and branded without imagery.
65-gradient theatre identity system
No play posters or photos were available across 65 Bay Area theatres. Designed 65 unique two-colour gradient combinations - one per theatre - so every card is visually distinct and branded without imagery.
Researched.
Designed.
Shipped.
Researched.
Designed.
Shipped.
murmur is the only search system of its kind for Bay Area theatre audiences. It was built from zero: user research, six personas, usability testing, hi-fi design, and a fully custom front-end with live Airtable integration - all in Framer.
This project is an end-to-end product design ownership - from empathy research to deployed, working software.
murmur is the only search system of its kind for Bay Area theatre audiences. It was built from zero: user research, six personas, usability testing, hi-fi design, and a fully custom front-end with live Airtable integration - all in Framer.
This project is an end-to-end product design ownership - from empathy research to deployed, working software.