OVERVIEW
0-1 Startup concept of an AI Multimedia Prompting Website that optimizes music selection for e-commerce advertising campaigns.
MY ROLES
Led end to end UIUX design including identifying problem, feature prioritization, design system and branding, wireframe and usability testing.
My previous experience in the advertising industry also helped me develop a deeper understanding of music in enhancing brand image and visual storytelling, which sets the foundation for the user research and design iterations.
​
GOAL
Increase advertisement engagement rate.
Reduce music discovery and licensing time.​
TEAM - Cornell Tech Startup
Mengying -- Designer
Aiden -- Project Manager
Jonathon -- Business Strategist
Branden -- Engineer
Daria -- Researcher​
DURATION
Jan - May 2024
TOOLS
Figma, Open AI API
Context
The advertising landscape is evolving rapidly, driven by multimedia innovations and expanding consumer expectations.
Finding the right on-brand music becomes crucial.
Understanding the Problem
Sifting through vast and generic music libraries is time-consuming for advertisers, further compounded by a lack of data-driven insights. This manual process makes securing appropriate licensing rights difficult and ultimately hinders innovation in ad production.
Our Opportunity
How might we utilize music to increase advertising efficiency and engagement within the sync licensing space?
Design Direction
Leverage AI-powered multi-media analysis to streamline music discovery and create emotionally resonant brand experiences that drive audience engagement and ROI.
Design Highlights
Multi-media prompting music discovery with brand alignment
Challenge: Finding on-brand music manually in a vast music library is time-consuming.
Solution: By leveraging multimodal prompting technology to recommend music that reflects brand identity and campaign objectives.
​
After capturing key brand attributes through an initial onboarding survey, users can input campaign details, including visuals, text descriptions, and filters, to generate highly targeted music recommendations. This approach ensures that every recommended song aligns seamlessly with the brand's identity and campaign objectives.
Design Highlights
Shortlist music and sync license with the right artists base on data
Challenge: Lack of artist data insights and inefficient licensing negotiations.
Solution: By providing comprehensive artist profiles, including follower demographics data, artists background info, and direct contact information to simplify music selection and licensing. ​
​
​Check the artist profile page to understand the artist’s vibes. Users can also curate personalized collections to streamline the decision-making process and collaborate with team members, ensuring the selected music perfectly aligns with their brand.
Process
How we solve the problem
Competitive Analysis
The competitive landscape reveals a variety of solutions catering to specific needs within the music sync industry.
​
Our analysis identified market gap and opportunities:
1. Streamlining the music search and licensing processes as it currently characterized by manual workflows.
2. Underserved market segment encompassing brand focus music discovery.
1. Market Research
2. User Research
Interviews
To validate our assumption from market research and learn more about the user pain points, we interviewed 18 stakeholders in the music licensing and advertising industry through our professional network.
Pain Points
“I spent hours and days looking through royalty free music, or music which maybe be licensed. And had video on the other monitor, press play on music site. Trying to look at if that worked or not.”
Rebecca, Social Media Associate at an e-commerce brand
Journey Map
To synthesis previous research findings, we mapped the two types of users journey here. Licensor and licensee' goals are different in the beginning, but eventually they come together. The pain points are happening when they come together in the negotiation and matchmaking stages which has a lot of intermediaries and it’s unautomated.
3. Ideation
Brainstorm and Assumption Mapping
As a team, we brainstormed ideas based on users problems, product development and market innovation.
To determine what to tackle on first, we created assumption maps across Market, Product, Model and Channel.
​
After rounds of team discussion and in-depth interviews with music sync coordinator and advertiser as well as my own working experience with different brands, we decided to focus on licensors for advertising purpose(startup to mid-size brands, especially in fashion, entertainment and catering industry) as main users and start with copyright free music. Building up artist side slowly starting with less-known artists to provide market place for them.
Feature Priorization
We drafted a North Star Metric, # of Monthly Campaigns Completed and pushed out, and prioritized product features started from listing user basic, articulated and unarticulated needs. And then mapped them based on importance and satisfaction level.
Key Features​
4. Design Experiment
Storyboard -- How it works
Based on prioritized features, I created end-to-end storyboard to present our product vision and user key actions to potential VC and customers.
Wireframe
To validate our idea through user testing, I designed the wireframes based on user flow which served as prototypes.
5. Design Iteration
User Test
To evaluate user experience and validate our ides, I utilized the interactive wireframe as a prototype for usability testing. This prototype was presented to 6 potential users with experience in music for advertising purposes. The user feedback received following this testing session informed modifications made to the user interface.
Iteration 1 ---- Enhanced Prompting Experience:
​​
1. Condense Filters Categories Displayed:
Based on user feedback of confusing about filter-heavy options, I prioritized a streamlined prompting page, which encourage users to leverage the power of free text and media prompts. This approach aligns with Syncro's core strength: interpreting multi media prompts to deliver music recommendations.
​​​
2. Add AI-powered Genre & Style Recommendations:
User feedback indicated a need for more AI-powered guidance. I implemented a feature that generates genre and style references based on multimedia prompts. This functionality translates users' creative vision into clear musical direction, empowering them to achieve the desired brand storytelling goals.​​
Iteration 2 ----- Artist Profile Page Optimization:
​
A key finding from user testing was the desire for more comprehensive artist information on the artist profile page. Users emphasized the importance of this information in determining stylistic alignment with their brand. To address this need, the following enhancements are proposed:​​​
1. Integration of Social Media Performance Data:
Including data on artist performance metrics within social media platforms will provide valuable insights into audience engagement and reach.
2. Incorporation of Online Portfolio References:
Providing access to the artist's online portfolio and visual style references will allow users to gain a deeper understanding of the artist's creative aesthetic.
6. Next Step
Product Roadmap
As a team, we came up with the potential roadmap for our startup idea to come to life.
Takeaway
1. Connect the dots from advertising to product development
As a startup project, we spent lots of time to figure out the right problem area. With previous internship experiences in advertising industry, I found out the connection between visuals and sound in ads and the needs to find the on-brand music. I also learned a lot about the consideration of choosing music for ads through the interviews with my previous colleagues. Those insights contributed a lot to our product direction.
​
2. The importance of both objective and subjective data to make informed music decision
During the user test with advertisers, I found that fashion brands would prioritize subjective data(artist vibe) of the music for brand alignment while catering brands would prioritize objective data(artist performance in social media) of the music for better engagement. It’s important to take reference of both data reference to make more informed music decision.
