H5: Social Proof

Hypothesis #5: Social Proof

Athletes or brands with higher levels of organic human discussion (e.g., Reddit, forums, reviews, LinkedIn) appear more prominently in LLM responses than those relying primarily on official marketing content.

Our prediction: Athletes (and brands) that get talked about by other humans show up more prominently in LLM responses than those relying on official marketing content alone.

Validation threshold: 3+ signals correlated with LLM visibility at p < 0.05

Athletes Tracked

Social Proof Signals

19,615

LLM Responses

28/33

Tests Significant

What We Found

Four Things the Data Proved (and One It Didn't)

28 of 33 statistical tests came back significant. But the most useful finding is the one that didn't.

Finding 01

The Last 7 Days on Wikipedia Beat Everything Else

Not 30-day views. Not 90-day views. Not Instagram followers. The strongest single predictor of whether an LLM mentions you is how much Wikipedia activity you had in the last week (ρ = 0.810). That is a stronger signal than total social reach across every platform combined. If you are planning a product launch, a rebrand, or a campaign, your Wikipedia page needs to be fresh before the moment hits, not after.

Finding 02

The Tier System Is Real, and the Boundaries Are Sharp

Mega-tier athletes averaged 14.3 LLM mentions. Major: 7.5. Mid: 4.2. Micro: 2.2. Statistically significant (p = 0.008). These are not soft categories. They are thresholds where AI behavior changes. If you are a Mid trying to show up like a Mega, the math is not on your side yet.

Finding 03

Surge Doesn't Equal Visibility. Period.

This is the finding that didn't work, and it is the most important one. Wikipedia surge percentage during the Olympics was not significant for any visibility metric (p = 0.11 to 0.51). Aicher surged +3,064% and got 2 mentions. Malinin surged +3,922% and got 15. Same Olympics, similar growth rate, 7x different outcome. The difference is entirely explained by where they started. You cannot event your way into AI relevance.

Finding 04

Crosby Hasn't Played a Single Game. AI Already Talks About Him.

Sidney Crosby does not have an Instagram, a TikTok, an X account, or a YouTube channel. Zero social followers across every platform. Men's hockey does not start until Feb 15, but he already earned 3 LLM mentions with a Position Score of 79.7, above average. His 183K monthly Wikipedia views and 231KB article are doing all the work. Social reach is the multiplier. Wikipedia authority is the foundation. You can have one without the other, but the ceiling is lower.

Primary Visualization

Bigger Footprint, More AI Mentions. Every. Single. Tier.

Each dot is an athlete. Hover for details. The trend line writes itself, but the outliers are where the strategy lives. Crosby hasn't played a single Olympic game yet and AI already talks about him. Glenn has 1.8M followers and barely registers. The correlation is real. The exceptions are more interesting.

Total Social Media Reach (log scale) x Estimated LLM Mentions · ρ = 0.759, p = 0.00001

10K

100K

TOTAL SOCIAL REACH →

LLM MENTIONS →

Shiffrin

Vonn

C. Kim

Malinin

Liu

Glenn

McDavid

Stolz

Aicher ★
+3,064% / 2 mentions

Crosby ★
Hasn't played / 3 mentions

Mega (1M+ reach)

Major (500K-1M)

Mid (100K-500K)

Micro (25K-100K)

Nano (<25K)

Signal Strength

Your Wikipedia Page Matters More Than Your Instagram

We tested 11 social proof signals against 3 LLM visibility metrics. Wikipedia recency, what happened in the last 7 days, beat everything else. Not total followers. Not lifetime views. What is fresh on your Wikipedia page right now.

Strongest Predictors (p < 0.001)

Wiki Views 7d x Presence

ρ = 0.810

Wiki Views 7d x Mentions

ρ = 0.792

Wiki Views 30d x Presence

ρ = 0.778

Total Social Reach x Mentions

ρ = 0.759

Instagram x Mentions

ρ = 0.728

Wiki Article Size x Position

ρ = 0.702

Not Significant (p > 0.05)

Wiki Surge % x Mentions

ρ = 0.312

p = 0.13

Wiki Surge % x Presence

ρ = 0.324

p = 0.11

Wiki Surge % x Position

ρ = 0.137

p = 0.51

The Olympics Effect

3,064% Growth. 2 Mentions.

Every athlete's Wikipedia exploded during the Games. Aicher's surge was nearly as large as Malinin's. She got 7x fewer mentions. The difference? Malinin started with 306K monthly wiki views. Aicher started with 6,800. Events are amplifiers, not creators.

Wikipedia Daily Avg (before → during Olympics)

Surge

LLM

Malinin

369,618/d

+3,922%

mentions

Chock

146,192/d

+3,278%

mentions

Sakamoto

24,101/d

+3,264%

mentions

Aicher ★

4,259/d

+3,064%

mentions

Glenn

65,473/d

+2,431%

mentions

102,298/d

+1,400%

mentions

Vonn

418,272/d

+892%

mentions

Crosby

10K/d

+60%

mentions

The Two Pathways

Some Athletes ARE the Answer. Others Get Recommended.

Only 8.1% of prompts were discovery-type questions: "athletes to watch," "medal favorites," "rising stars." But they reveal two completely different ways LLMs decide to talk about you. And they require different strategies.

Authority Pathway

They ARE the answer

These athletes appear in 0 of 5 discovery tags. LLMs never "recommend" them because they do not need to. They dominate substantive queries: biographical, predictive, event-specific. They are cited as facts, not suggestions.

Shiffrin (#1, 18 mentions) · Vonn (#2, 16) · Malinin (#3, 15) · Stolz (#5, 13) · McDavid (#12, 4) · Crosby (#16, 3)

Discovery Pathway

LLMs introduce them

These athletes appear in 1 to 3 discovery tags. LLMs actively recommend them in curated lists. "Athletes to watch," "medal favorites," "rising stars." They are being positioned as someone you should know about.

Chloe Kim (#4, 15 mentions, 3 tags) · Eileen Gu (#6, 8, 1 tag) · Alysa Liu (#7, 7, 2 tags) · Diggins (#10, 5, 2) · Glenn (#11, 4, 2) · Hall (#14, 4, 3) · Brignone (#17, 3, 1) · Aicher (#18, 2, 1)

Three Stories

Same Olympics. Same Surge Window. Three Different Realities.

The same Games, the same window of attention, three completely different outcomes, and each one maps directly to a different GEO strategy.

The Amplifier

Ilia Malinin

Figure Skating · Major Tier

Wiki Views 30d306,831

IG Followers515,000

Olympics Surge+3,922%

LLM Mentions15

Strong foundation + event surge = maximum amplification. He already had the signals. The Olympics just turned up the volume. This is the model: build first, then let events compound.

The Cautionary Tale

Emma Aicher

Alpine Skiing · Micro Tier

Wiki Views 30d6,853

IG Followers28,000

Olympics Surge+3,064%

LLM Mentions2

Massive surge, almost no AI visibility. 3,064% growth means nothing when your starting point is 135 daily views. You cannot event your way into AI relevance without a foundation.

The Exception

Sidney Crosby

Ice Hockey · Nano Tier

Wiki Views 30d182,955

Social Followers0

Wiki Article Size231 KB

LLM Mentions3

Zero social presence. Hasn't played a single Olympic game yet. AI still talks about him. Men's hockey does not start until Feb 15, but Crosby already has 3 LLM mentions. Wikipedia authority alone, a massive article with consistent traffic, sustains mid-range AI visibility. Social is the multiplier, not the prerequisite. Reputation built over decades creates a signal no surge can replicate.

H5 Verdict

Hypothesis Assessment

H5: SUPPORTED / 10 OF 11 SIGNALS, 28 OF 33 TESTS SIGNIFICANT

Your digital footprint doesn't just correlate with AI visibility. It predicts it.

We required 3+ social proof signals correlated with LLM visibility at p < 0.05. We found 10 of 11 signals significant, with 28 of 33 individual tests passing. The strongest predictor was Wikipedia recency (ρ = 0.810), followed by total social reach (ρ = 0.759) and Instagram followers (ρ = 0.728).

The one signal that did not predict visibility: Wikipedia surge percentage. This is perhaps the most actionable finding of all. Events amplify existing signals but do not create visibility from nothing. The Malinin/Aicher contrast proves it: similar surge percentages, 7x different outcomes, entirely explained by baseline signal strength.

Additionally, the discovery tag analysis revealed two distinct pathways to AI visibility: authority (dominating substantive queries) and discovery (appearing in curated recommendation lists), each requiring different optimization strategies.

Strategic Implications

What to Actually Do About This

Practice 01

Update Your Wikipedia Before the Moment, Not After

If you have a product launch, earnings call, or campaign coming, your Wikipedia page needs fresh edits, current stats, and recent citations in the 7 days before, not during or after. Recency beat cumulative authority as the strongest predictor of AI visibility. A well-timed Wikipedia update may matter more than three months of social content.

Evidence: Wiki Views 7d x Presence %, ρ = 0.810, p = 0.000001

Practice 02

Be on More Platforms, Not Just Bigger on One

Athletes on 4 platforms averaged 14.3 LLM mentions vs. 2.2 for those on 1 to 2 platforms. LLMs appear to weight multi-signal consistency. Showing up in multiple places tells a different story than a large following in one. Do not concentrate all effort on a single channel.

Evidence: Total Reach x Mentions, ρ = 0.759; Platform Count x Mentions, ρ = 0.633

Practice 03

Make Your Wikipedia Article Bigger and Better-Sourced

Article size correlated most strongly with Position Score: where you rank when you are cited. Crosby's 231KB article drives a Position Score of 79.7 despite zero social presence. Detailed sections, sourced claims, structured data. Size predicts WHERE you appear, not just IF.

Evidence: Wiki Article Size x Position Score, ρ = 0.702, p = 0.00009

Practice 04

Stop Expecting Events to Create Visibility From Nothing

Aicher's +3,064% Wikipedia surge produced 2 LLM mentions. Malinin's +3,922% produced 15. Same event window, similar growth curves, completely different outcomes, because Malinin had 306K monthly views to amplify. Build the foundation before the campaign. The event is the amplifier, not the source.

Evidence: Wiki Surge % x Mentions, ρ = 0.312, p = 0.13 (NOT significant)

Practice 05

Know Your Tier and Set Targets Accordingly

Mega = 14.3 avg mentions. Major = 7.5. Mid = 4.2. Micro = 2.2. These boundaries are statistically real. Optimization strategies should aim to cross the next tier threshold, not skip two. A Micro-tier brand targeting Mega-tier AI visibility is budgeting for disappointment.

Evidence: Kruskal-Wallis H = 13.80, p = 0.008

Methodology

Data Considerations

Data Sources & Collection

LLM visibility data: 19,615 responses collected via Scrunch across 5 LLMs (Gemini, Google AI Mode, AI Overviews, Meta AI, ChatGPT) from February 6 to 10, 2026 (Olympic days 1 through 5). Metrics: Estimated Mentions, Presence %, Position Score.

Social proof signals (11): Wikipedia views (90d, 30d, 7d), Wikipedia edits (30d), Wikipedia article size (bytes), Instagram followers, total social reach, platform count, Wikipedia views during Olympics (4d), Wikipedia surge %, Wikipedia edits during Olympics.

Wikipedia baseline: Snapshot date February 5, 2026. Olympics surge data: February 6 to 9, 2026 vs. January 29 to February 5, 2026 baseline.

Discovery tag data: 1,586 prompts across 5 discovery tags (Athletes to Watch, Country Favorites, Dominance, Medal Favorites, Rising Stars), 8 athletes surfaced across tags.

Statistical Methods

Primary test: Spearman rank correlation (non-parametric, appropriate for non-normal distributions). 33 tests: 11 signals x 3 visibility metrics. Significance threshold: p < 0.05.

Group comparisons: Kruskal-Wallis H-test across 5 social tiers. Mann-Whitney U test comparing top 5 vs. bottom 5 athletes by visibility.

Result: 28 of 33 tests significant. 10 of 11 signals significant on at least one visibility metric. Only Wikipedia surge % failed to reach significance on any metric.

Limitations

Sample size: n = 25 athletes. Sufficient for Spearman correlation detection at moderate effect sizes, but limits ability to run multivariate regression. Results should be validated with larger samples.

Aggregated LLM data: Scrunch export aggregated across all 5 platforms. Per-platform breakdowns were not available for H5 analysis. Different LLMs may weight social proof signals differently.

Temporal scope: Social proof signals captured at a single baseline snapshot (Feb 5). LLM responses collected over 5 days. Both are snapshots, not longitudinal measures.

Correlation ≠ causation: Strong correlations do not prove that social proof signals cause LLM visibility. Both could be driven by underlying fame or relevance. The surge % finding (not significant) provides the strongest causal hint: that pre-existing authority matters more than dynamic change.

Hypothesis 5 | Social Proof

Emma Aicher Grew 3,064% on Wikipedia and AI Didn't Care

What's happening in the test of the Olympic games tests?

Four Things the Data Proved (and One It Didn't)

Bigger Footprint, More AI Mentions. Every. Single. Tier.

Your Wikipedia Page Matters More Than Your Instagram

3,064% Growth. 2 Mentions.

Some Athletes ARE the Answer. Others Get Recommended.

Same Olympics. Same Surge Window. Three Different Realities.

Hypothesis Assessment

What to Actually Do About This

Data Considerations