Case Study

The Relevance Index: Scoring 1,200 Brands on Cultural Relevance

No objective measure of cultural relevance existed. So one strategy director built one -- automated, data-driven, and updated every week.

By Mike Litman • Cultural Capital Labs • 2025-2026
Wikipedia API
Reddit API
GPT-4o-mini
yfinance
GitHub Actions
1,203
Brands Scored
5
Scoring Domains
Weekly
Automated Updates
260
With Stock Data

The Problem

Everyone talks about cultural relevance. Brands want it. Agencies promise it. But nobody can actually measure it. There is no Bloomberg terminal for culture -- no objective, data-driven way to track which brands are culturally relevant and which are just commercially successful.

The gap is significant. Commercial performance and cultural relevance are not the same thing. A brand can sell billions and have zero cultural cachet. Another can dominate conversation without selling a single product. Understanding that distinction -- and tracking it over time -- is valuable for anyone in the business of brands.

The Approach

The Relevance Index uses a hybrid scoring model that combines real data signals with AI-powered qualitative assessment. Every Wednesday at 3am UTC, a GitHub Actions pipeline runs the entire scoring process automatically:

Wikipedia pageviews
Reddit activity
3x GPT-4o-mini calls
yfinance stock data
Score calculation
1,200+ brand pages
Deploy

Each brand is scored across five dimensions, each contributing to an overall score out of 100:

Attention
Is the brand being looked at? Wikipedia pageviews, search interest, media mentions.
Conversation
Is the brand being talked about? Reddit activity, social discussion, community engagement.
Creation
Is the brand inspiring people to make things? User-generated content, fan art, remixes.
Desire
Does the brand generate want? Aspirational pull, status association, emotional connection.
Zeitgeist
Does the brand capture the moment? Alignment with cultural currents, timeliness, relevance to now.

The scoring uses three independent GPT-4o-mini calls per batch (not one) to reduce bias and increase reliability. Wikipedia and Reddit provide the quantitative signal. The LLM provides qualitative assessment. The combination creates a more robust picture than either approach alone.

For 260 publicly traded brands, stock market data (market cap and 30-day price change) is layered in, enabling comparison between cultural relevance and financial performance.

How It Was Built

1
Brand Curation
Curated 1,203 brands across 15 categories -- from luxury fashion to fast food, social platforms to automotive. Every brand hand-selected for cultural significance.
2
Hybrid Scoring Model
Built a hybrid scoring engine combining Wikipedia pageviews and Reddit API data (quantitative signal) with three independent GPT-4o-mini calls per batch (qualitative assessment) to reduce bias.
3
Financial Enrichment
Added yfinance stock data enrichment for 260 publicly traded brands -- market cap and 30-day price change -- enabling cultural vs. commercial comparison.
4
Brand Pages & OG Images
Generated 1,203 individual brand pages with sparkline charts, radar comparisons (Chart.js), AI insights (Claude Sonnet), and brand-specific OG images (Pillow, 1200x630 PNG).
5
Weekly Automation
GitHub Actions pipeline runs every Wednesday at 3am UTC -- scoring all 1,203 brands, enriching financial data, generating insights, building pages, and deploying to Netlify. Fully autonomous.
Technical Architecture
DATA SOURCES SCORING ENRICHMENT OUTPUT [Wikipedia API] ---> [Score Engine] ---> [Claude Sonnet] --> [1,203 Pages] [Reddit API] ---> | | | [3x GPT-4o-mini] -> | AI insights Brand pages 5 dimensions OG images 0-100 score | Category pages | [yfinance] | | 260 stocks [Netlify Deploy] | Market cap | GitHub Actions 30d change Email digest Wed 3am UTC (Buttondown)

The Result

1,203
Brands Scored
5
Dimensions
260
With Stock Data
Weekly
Updates

The Relevance Index now tracks 1,203 brands with individual brand pages, category pages, AI-generated insights (via Claude Sonnet), sparkline trend charts, radar comparison charts, and embeddable badge widgets. Each brand has its own OG image for social sharing.

A weekly email digest (via Buttondown) summarises the biggest movers -- brands gaining or losing cultural relevance. The entire system runs autonomously, requiring no manual intervention.

The most surprising finding has been the consistent divergence between cultural relevance and stock price. Some of the most culturally relevant brands are financially underperforming, while some commercial giants barely register culturally. That tension is where the story lives.

The Key Insight

Cultural relevance and business performance often diverge -- that tension is where the story is.

A brand like Patagonia scores exceptionally on cultural relevance -- talked about, admired, referenced -- but its commercial scale is modest compared to, say, Zara. Conversely, a brand like Oracle is commercially massive but culturally invisible. The gap between cultural weight and commercial weight is the most interesting thing The Relevance Index reveals.

For strategists, marketers, and anyone working in brands, this gap is the opportunity. Understanding where your brand sits on the cultural-commercial axis is the starting point for any meaningful strategy. The Relevance Index makes that visible for the first time.

Lesson Learned
1,203 brands scored weekly by a pipeline that runs while I sleep. The question is not whether AI can do the work. It is whether you have the taste to direct it.
Visit The Relevance Index Back to All Projects