Brand–Artist Matching: From Idea to Working System
The founding team had deep industry knowledge and identified a clear market problem.
I brought technical research expertise to turn their vision into a functional, scalable system.
Music Tech Venture · Pre-Investment·System Architecture·Research Translation
01 The Starting Point
A team with market insight and a technical gap
A small team came together to address a real market problem: matching brands with suitable
music artists for campaigns relies heavily on intuition and lacks transparency, resulting
in missed opportunities, even with substantial budgets.
The team brought deep domain expertise, including artist management experience, established
relationships with artists and brands, firsthand knowledge of process challenges, and product
and frontend engineering skills. However, they lacked the research depth and backend technical
infrastructure to evaluate viable matching logic, build the data layer, and clearly present
the system to investors. With a research background that includes a PhD in music recommendation,
I was a natural fit to join the founding team.
The team had a vision and industry knowledge, while I provided the technical research
foundation and designed the system architecture.
The team
Vision, industry relationships & product
The original idea, artist and brand relationships, business development, and frontend engineering. Firsthand knowledge of where the current matching process fails.
My contribution
Research, architecture & backend build
Feasibility assessment, matching logic, system architecture, backend and data infrastructure, data visualization design, and investor-facing technical documentation.
02 Why the Problem Is Difficult
A complex matching problem
Deriving a data-driven, algorithmic method to match artists with brand campaigns is a
complex problem. Solving it requires understanding the industry's intuitive decision-making
process and identifying measurable data signals. Quantitative factors such as audience location,
age, and online presence are easy to compare and align with campaign needs, but provide only a
partial view. A complete assessment must also include qualitative factors such as style,
reputation, and values, which cannot be directly measured.
Quantitative
Audience demographics
Age, gender, geography, and audience reach across streaming and social platforms. Data collected from publicly available APIs.
Qualitative
Brand tone compatibility
Alignment between an artist's persona and the emotional tone associated with the brand.
Qualitative
Visual language
Color palette, styling, aesthetics. Is the artist's output coherent with the brand's identity?
Qualitative
Cultural credibility
Is the brand artist pairing culturally aligned, or is it just a business deal? Audiences notice the difference.
Qualitative
Long-term positioning
Where is the artist's career heading? Strong partnerships are built on alignment that holds over time.
Quantitative
Audience interest overlap
Do the artist's fans show strong interest in the brand's category and related brands?
Signals also conflict, and none are decisive on their own:
An artist matches geography perfectly, but fails on cultural tone.
Aesthetic alignment is strong, but the audience isn't large enough to meet the campaign goals.
Quantitative signals look excellent, but career trajectory carries reputational risk.
The first challenge was to determine which factors to include, how to compare quantitative
and qualitative signals, and how to explain the reasoning behind a match.
03 My Contribution
Turning market insight into a functional system
My role was to bring the research depth and backend technical capability the team needed.
This involved five key areas of work.
Feasibility & prototype scope
Assessed what was realistic for the prototype given the available data sources and
technical constraints. We chose to build around a single, data-rich API and one
qualitative dimension — artist 'vibes'. I designed a lightweight method to determine
and incorporate vibes into the matching logic.
System architecture & backend build
Designed and built the backend architecture, including a data layer to normalize signals
from the API, a matching layer for filtering and ranking compatibility dimensions, as well
as the database, data collection pipeline, and deployment infrastructure.
Matching logic
Developed the core matching and ranking logic, drawing on music recommendation research
to define the key factors of brand-artist compatibility and weight them according to
campaign goals. The results are transparent and easy to interpret, enabling brand managers
to understand and explain recommendations with confidence, rather than relying on an
opaque scoring system.
Data visualization
Designed the first set of visualizations to help users understand the reasoning behind
compatibility scores. The goal was to break out each relevant signal and provide clear
explanations, so managers can easily see and assess the trade-offs involved in each recommendation.
Technical documentation
Authored the technical white paper for non-technical investors, explaining the system's
architecture and data-driven approach to matching brands and artists.
04 The Matching Framework
Matching system design
The diagram below illustrates the system design, divided into stages, beginning with data
collection and resulting in an explainable, ranked list of artist recommendations. The current
prototype demonstrates the platform operating on a simplified matching algorithm with a single
data source, while future development will incorporate multiple data sources and more
sophisticated recommendation methods.
05 Outcome
Where the work stands
A working prototype validating the core matching logic and providing explainable match
results. Active interest from management companies and investors. A technical white paper
providing a detailed explanation of the system design for fundraising conversations. The
venture is currently preparing to raise its first round of funding.
This project demonstrates how careful system design can begin to transform what has
traditionally been an intuitive, opaque process into one that is transparent and driven
by real-world data. Even at the prototype stage, the core matching logic shows that
brand-artist compatibility can be assessed in a structured, explainable way, giving
both brand managers and artists a clearer basis for partnership decisions.
06 What This Demonstrates
The pattern behind the work
What this work required
Applying research expertise to a practical challenge.
Building this system required an understanding of recommendation system subtleties,
beyond technical skills. It also required the judgment to identify which signals are
relevant and to design matching logic that is robust and transparent.
Making complex logic accessible for different stakeholders.
The matching framework needed to be understandable for engineers, practical for
brand managers, and credible for investors.
Translating qualitative concepts into data.
I developed the approach to identify and integrate a key qualitative factor, artist
vibes, into the matching logic, transforming an abstract idea into a functional
element of the system.
Writing technical documentation for investors.
I authored a white paper that explained the system's design and rationale,
demonstrating our thorough planning and domain expertise to investors in order
to strengthen our case for funding.