Skip to main content
zeracreators
How the score thinksUpdated continuously

Principles, not formulas.
Public philosophy. Private coefficients.

The Zera Score is proprietary. This page explains what we measure and why it's fair — without publishing weights, factor counts, or anything a competitor could reverse-engineer. You get comfort. Bad actors don't get a cheat sheet.

Seven principles

What the score rewards.

Framed from the creator's point of view, not as a list of weights. Each principle is enforced by the algorithm; we just don't publish the coefficient that does it.

Real humans, not raw follower counts

A million-follower account with a bot audience should score lower than a hundred-thousand-follower account with a real one. We detect inauthentic audiences and down-weight their contribution — on every platform we cover.

Audience authenticity signals

Engagement quality, not engagement theatre

Pods, coordinated comment rings, bursty-hour spikes, and copy-paste reply patterns are common in finance creator-land. We flag them and weight the score toward engagement we can trust. Depth beats volume.

Pattern-detection on public interaction

Rhythm brands can plan around

Sponsorships are operations. A creator who posts three times a week for six months is a better partner than one who goes viral every third month. We reward steady, predictable cadence across multiple dimensions of consistency.

Cadence & craft signals

Cross-platform portability

Single-platform creators are fragile — one algorithm change and the roster collapses. Creators who carry their audience across Instagram, TikTok, YouTube and X are structurally more durable and score accordingly.

Multi-platform coverage analysis

What the audience actually says

A thousand emoji comments matter less than a hundred real sentences. We read comment substance — length, questions, tone — because a replying audience is a buying audience.

Comment-substance signals

Sustainable growth, not bought spikes

Real growth has a shape — gradual, with a credible acquisition story. Sudden integrity-breaking spikes trigger a closer look and the score holds until we can explain them. We prefer creators whose growth we can defend to a brand.

Momentum & integrity checks

Legibly about one thing

Brands buy when a creator's audience is clearly about a topic. We measure topical focus, niche-adjusted engagement, and whether past sponsorships performed in line with organic content. Clear beats big.

Topic clarity & sponsor history

Hook craft, measured against a cross-niche library

Going viral is partly a hook problem. We classify a creator's last 50 posts against 12 proven hook archetypes (curiosity-gap, contrarian, listicle, transformation, authority, question, number-shock, negation, story, confession, pattern-interrupt, objection-handler) and produce a Hook Mastery score for both diversity (how many archetypes the creator can pull off) and strength (how each one performs vs. that creator's own baseline). The reference library that powers the comparison is auto-refreshed daily from a curated cross-niche watchlist — fitness, business, fashion, food — so trading creators benefit from hook patterns proven elsewhere.

12-archetype taxonomy · auto-refreshed reference library
What the score refuses to do

Guardrails, not just principles.

These are promises, not formulas. Publishable without giving anything away because they're a commitment — if we break them, the score stops being worth anything.

No pay-to-win

No creator can boost their score by paying us. Nothing that is sold moves a score, and nothing that moves a score is sold.

No per-creator hand-tuning

The same inputs produce the same score for every creator. No allowlists, no special cases, no 'friends of Zera' adjustments.

No private data

We only use public signals the creator has chosen to make visible. No DMs, no private analytics, no paid leaks from platforms.

Reproducible by audit

Every score is versioned. You can see what moved between any two refreshes and compare against the algorithm version it was computed under.

Data sources

Only what a public visitor could see.

Every input is something a creator has chosen to make public, or has explicitly opted to share with brands via Zera. Nothing scraped from private APIs, nothing bought from third parties.

Instagram
Public profile, post history, engagement metrics
TikTok
Public profile, video history, view & engagement counts
YouTube
Channel stats, video metadata, public captions + comments
X (Twitter)
Public tweets, reply threads, follower structure
Creator-declared
Bio contact channels, preferred email, site/linktree (creators opt in)
Brand mentions
Scraped from creator posts for the Affiliate OS / Brand intel layer
Brand Reputation Score

How we score the brands themselves.

Two scores live on every prop-firm review page. The Zera Score (0-10) rates the product — profit split, regulation, account sizes, platforms. The Brand Reputation Score (0-100) rates the conduct — what creators say, how the brand behaves, how it's been judged in the wild.

v0.9 inputs
What the score reads today. Updated whenever creator content surfaces a new brand mention.
Creator-content sentiment
60 / 100
AI-summarised tone of every creator mention of the brand on TikTok, Instagram, YouTube, X. Recomputed weekly. Higher score when the dominant tone across recent mentions is positive and consistent.
Mention volume
15 / 100
Log-scaled count of analysed mentions. A brand with 10 mentions shouldn't rank like one with 1,000 — but the curve flattens fast so a single viral creator doesn't dominate the score.
Operating longevity
15 / 100
Years since founding. Past the 3-year mark new prop firms have cleared the high-failure window; past 5 years they're materially established. Unknown founding year = neutral baseline.
Source diversity
10 / 100
Number of distinct platforms / creators contributing mentions. Sentiment from one source can be one creator's relationship to the brand; diverse sources are a stronger trust signal.

Same dispute mechanism as the creator score (see below). A brand that believes a sentiment input is wrong files a contestation; we publish the resolution.

Recent updates

A public log of every methodology change.

Charter Promise 03: the methodology is public and contestable. Every change to the scoring algorithm, every guardrail revision, every product launch that touches the Score gets an entry here with a one-line rationale. Most recent first.

  1. v2.91·May 2026charter

    Pricing transparency added as Charter Promise 06

    Pricing tiers for the brand-side platform fee + add-on subscriptions are bound to the published table. No off-table rates, no enterprise sweetheart deals; tier promotion stays admin-gated and deposit-driven.

  2. v2.91·May 2026product

    Continuous Monitoring product launched

    $99/mo per creator daily-check subscriptions for brands. New alert kinds: SCORE_DROP/RISE, AUDIENCE_DROP/RISE (aggregate + per-platform), POST_SILENCE, ENGAGEMENT_DROP. Threshold tuner per-monitor.

  3. v1.38.7·Apr 2026scoring

    Activity-recency factor moved out of the capped dampener pool

    Previously the combined dampener product was floored at 0.90 (max −10% combined hit). With activity-recency now its own uncapped multiplier, a creator silent for 6+ months can drop further than the cap previously allowed — accurate to the Charter promise of "audit-grade" scoring.

  4. v1.38.4·Apr 2026scoring

    Activity-recency tier rework: 60d→0.95, 120d→0.65, 180d→0.40

    Old behaviour clamped at 0.95 even for a creator silent for a year — incompatible with brand-deal readiness. New tiered floor lets dormant accounts drop substantially while staying non-zero (severe-but-not-disqualifying floor at 0.40).

  5. v1.34.0·Mar 2026scoring

    Pod-uniformity penalty on engagement-authenticity

    When the like:comment ratio is unnaturally uniform across posts (CV very low) AND the median ratio sits in the pod-adjacent zone (200–500), an additional 0.05 dampening lands on top of the linear interpolation. Real audiences vary; pods produce flat reactions.

  6. v1.17.0·Feb 2026scoring

    Activity-recency dampener introduced

    Catches the case of a creator with strong 28d/60d averages whose most-recent post is old enough to make brand planning risky. Downward-only multiplier across the overall score.

Per the Charter, methodology changes are announced ≥30 days in advance. Changes proposed but not yet shipped land on the public roadmap before they appear here. Email methodology@zeracreators.com to propose or contest.

See a score you think is wrong?

Anyone — creator, brand, or curious bystander — can challenge a score. Mail dispute@zeracreators.com with the profile URL + what you think is off. We respond in 72h with either a revised score + the signals that moved it, or the reasoning behind the number.

Creators can also trigger a re-score on-demand from their dashboard — useful if a platform glitch caused a misread.

Know the philosophy.
Now get your number.

Free forever. 60-second setup. The same seven principles measured on your profile, the same guardrails applied to your score.