Culture Predicts
Revenue.
We Proved It.
The first large-sample evidence that structured cultural measurement predicts quarterly revenue growth. 847 public brands. 16 quarters.
72.9%
Mid-quarter accuracy in Consumer Discretionary
~2.5 months before earnings are publicly reported
The Problem
Every tool measures how much attention.
None measure what kind.
Brand Trackers
Measures: Awareness, consideration, NPS
Expensive, slow (quarterly), limited coverage. Measures what consumers say, not how they behave. No structural insight into where or why perception is changing.
Social Listening
Measures: Mentions, sentiment, share of voice
Requires active brand discussion. Sentiment analysis unreliable at scale. Platform-dependent and API-constrained. Cannot map structural cultural position.
Google Trends
Measures: Aggregate search interest over time
Treats all search activity as equivalent. A product recall spike and genuine cultural embedding register identically. Cannot distinguish productive attention from noise.
All three share the same blind spot: they treat attention as a scalar quantity. Volume goes up or down. But a brand embedded across wellness, sustainability, and lifestyle arenas is in a fundamentally different position than one spiking from controversy. The structure of attention is where the economically significant information lives.
Key Findings
The Numbers
72.9%
Mid-Quarter Accuracy
Depth Velocity at Month 2 in Consumer Discretionary. ~2.5 months of lead time before earnings.
847
Public Brands Studied
All 11 GICS sectors. Q1 2022 through Q4 2025. Revenue validated against SEC EDGAR filings.
+13.9pp
Arena vs. Google Trends
Communication Services: Cultural Velocity 63.9% vs. search volume 50.0%. The largest gap in the study.
5/10
Signals Survive Validation
Five-stage hardened pipeline: walk-forward holdout, block bootstrap, Fama-MacBeth factor controls.
$110M
Revenue per 1 SD
Per quarter, for the median Consumer Discretionary brand. Controlled for firm size and momentum.
847 Brands Studied
Same Search Spike.
Different Story.
Three Consumer Discretionary brands. All showed double-digit search interest increases. Google Trends treated them identically. Cultural Velocity did not.
Cultural noise, not signal
“Tesla robotaxi” surged +3,491% and “tesla pi phone” spiked +2,917%, while core product queries declined. “Tesla model 3” fell 42%. Arena-relative momentum: bottom 8th percentile.
Genuine cultural embedding
“Chipotle near me” grew +89%, “chipotle rewards” surged +335%. Deepening across dine-out culture and fast-casual arenas. Competitors declining. Revenue grew +8% YoY.
Trailing the category it created
“Lululemon belt bag” fell 82% while “lululemon stock” surged +339%. Its arenas grew +24–27 points, but the brand sat at the median, trailing the movement it pioneered.
Cultural Velocity correctly identified Chipotle as heading toward revenue growth, Tesla as culturally hollow noise, and Lululemon as losing depth relative to the movements it once led.
Sector Gradient
Where Cultural Signals
Work Best
The signal works where the mechanism operates: consumer cultural engagement drives purchase. It does not work where revenue is driven by contracts and commodity pricing. This is expected, and is itself a validation.
Revenue Economics
Commercially
Significant
$110M
Additional quarterly revenue per 1 SD of Cultural Velocity
Median Consumer Discretionary brand (~$1.1B quarterly revenue). Fama-MacBeth controlled for firm size, revenue momentum, and prior-quarter reversal.
Bottom Quintile
+3.0%
YoY revenue growth
51.6% probability of positive growth
Top Quintile
+28.8%
YoY revenue growth
74.5% probability of positive growth
25.7 percentage-point spread (p = 0.002). 204 Consumer Discretionary brands. The effect concentrates at the extremes: the top 20% pulls dramatically away.
Download
Get the
Full Paper
50+ pages of methodology, validation results, sector deep dives, and the complete five-stage hardened pipeline. Everything behind the headline numbers.
Format: PDF, delivered to your inbox
Length: Full working paper with appendices
Data: 847 brands, 16 quarters, SEC EDGAR revenue
Inside the Paper
What You'll Learn
Why 5 of 10 signal-timing combinations survive rigorous five-stage validation and why the other 5 fail
How arena-decomposed signals outperform Google Trends by 5–14 percentage points in consumer-facing sectors
The Tesla/Chipotle/Lululemon contrast: three brands, same search spike, completely different revenue outcomes
Sector-specific signal regimes: why depth drives discretionary, persistence drives staples, and volume predicts nothing for media
Revenue economics: the $110M/quarter finding, quintile spreads, and Fama-MacBeth factor-controlled coefficients
The complete stock prediction null result, published in full, because transparency is the strongest defense against selective reporting
PDF delivered to your inbox.