Influencer Marketing Glossary
From audience authenticity to unified scoring, this glossary covers the key terms and concepts you need to understand influencer vetting, brand safety, and creator scoring. Each definition links to deeper resources across our platform.
Audience Authenticity
A measure of how genuine a creator's follower base is, accounting for bot accounts, purchased followers, and inactive users. CreatorScore's Authenticity Agent blends content-level authenticity signals with follower-level fraud detection to produce a normalized 0–100 score. High audience authenticity means the creator's reach reflects real people who genuinely engage with their content.
Audience Quality Score
A composite metric that evaluates the health and engagement depth of a creator's community. It combines community health indicators (comment quality, reply rates) with engagement quality signals like average comment length and question ratio. The Audience Quality Agent weights these equally to produce a score that helps brands understand whether a creator's audience is actively invested or passively scrolling.
Bot Detection
The process of identifying automated or fake accounts within a creator's follower base and comment sections. CreatorScore analyzes comment patterns, posting cadence, account age, and behavioral anomalies to flag bot activity. When bot rates exceed 60% of a creator's followers, a knockout factor caps their score at 20/100, reflecting severe audience fraud. Learn more in our guide on how to detect fake followers.
Brand Safety Score
One of CreatorScore's seven scoring agents, weighted at 15% of the overall score. It evaluates three components: FTC disclosure practices (35%), controversy history (35%), and brand partnership patterns (30%). A web reputation signal is blended in at 15% when available. The Brand Safety Score helps marketers identify creators whose content and behavior align with brand values and regulatory standards. See our full scoring methodology for details.
Content Risk Score
The highest-weighted agent in CreatorScore's system at 20%, evaluating the safety of a creator's content across five components: hate speech (30%), NSFW content (25%), severity of flagged material (20%), visual analysis (15%), and profanity (10%). Content Risk scoring uses both NLP models and visual analysis to detect problematic material that could damage a brand's reputation. Read about brand safety red flags for practical examples.
Creator Score
The flagship metric produced by CreatorScore's platform — a single 1–100 number representing a creator's overall brand safety and partnership readiness. It is calculated as a weighted average of seven AI scoring agents: Content Risk, Brand Safety, Sentiment, Authenticity, Audience Quality, Community Trust, and ROI Prediction. Scores are categorized into tiers: Exceptional (90+), Excellent (80+), Good (70+), Fair (60+), and Poor (<60). See the methodology page for the full breakdown.
Cross-Platform Scoring
The ability to evaluate a creator's brand safety across multiple social platforms simultaneously. CreatorScore links a creator's YouTube, TikTok, Instagram, Twitter/X, and other accounts into a single identity group, then produces a follower-weighted unified score that reflects their total digital presence. This prevents creators from hiding risky content on lesser-known platforms.
Disclosure Compliance (FTC)
A measure of how consistently a creator follows Federal Trade Commission guidelines by clearly labeling sponsored content, affiliate links, and brand partnerships. CreatorScore's Community Trust Agent tracks disclosure rates across posts containing brand mentions. Creators with disclosure rates below 10% may trigger a knockout factor that caps their score at 35/100. Also see FTC Compliance.
Engagement Pods
Coordinated groups of creators or accounts that artificially inflate each other's engagement metrics by systematically liking, commenting on, and sharing each other's posts. CreatorScore detects pods through comment pattern analysis, timing anomalies, and network clustering. When pod participation exceeds 80%, a knockout factor caps the creator's score at 30/100. Engagement pods undermine the reliability of engagement rate as a performance metric.
Engagement Rate
The ratio of audience interactions (likes, comments, shares, saves) to a creator's total reach or follower count. While widely used in influencer marketing, raw engagement rate can be misleading without context — it must be evaluated alongside bot detection and pod detection to separate genuine audience interest from artificial inflation. CreatorScore uses tier-normalized benchmarks so micro-influencers are not unfairly compared to mega-creators.
Fake Followers
Accounts that follow a creator but are not real, engaged humans — including bots, purchased followers, and inactive or abandoned accounts. Fake followers inflate a creator's apparent reach without delivering real brand exposure. CreatorScore's Authenticity Agent detects fake followers through behavioral analysis, engagement anomalies, and follower growth pattern inspection. Our complete guide to detecting fake followers covers the key signals brands should watch for.
FTC Compliance
Adherence to the U.S. Federal Trade Commission's Endorsement Guides, which require creators to clearly and conspicuously disclose material connections with brands. This includes using labels like #ad, #sponsored, or platform-native disclosure tools. Non-compliance exposes both creators and brands to regulatory action. CreatorScore tracks FTC compliance as part of the Brand Safety and Community Trust agents.
Growth Trajectory
The rate and pattern of a creator's audience growth over time, measured across 30-day, 60-day, and 90-day windows. Healthy, organic growth tends to be steady and correlated with content output, while sudden spikes may indicate purchased followers or viral but unsustainable attention. Growth trajectory is weighted at 35% of the ROI Prediction Agent, making it a key factor in forecasting a creator's future partnership value.
Hate Speech Detection
Automated identification of content that promotes hatred, discrimination, or violence against individuals or groups based on protected characteristics. CreatorScore uses NLP models to analyze captions, transcripts, and comments, weighting hate speech at 30% of the Content Risk Score. When hate speech signals exceed 90%, a knockout factor caps the overall score at 35/100. Commentary, news, and educational content niches receive adjusted thresholds to reduce false positives.
Identity Group (Cross-Platform Identity)
A unified profile that links all of a creator's social media accounts (YouTube, TikTok, Instagram, Twitter/X, etc.) into a single entity. Identity groups enable cross-platform scoring by ensuring that a creator's behavior on every platform contributes to their overall unified score. This prevents creators from maintaining a clean image on one platform while posting risky content on another.
Knockout Factors
Hard score caps applied when severe issues are detected, overriding the normal weighted average. Knockout factors exist for the most critical brand safety risks: bot followers exceeding 60% (cap 20), engagement pod participation above 80% (cap 30), hate speech above 90% (cap 35), NSFW content above 95% (cap 35), and FTC disclosure rates below 10% (cap 35). These ensure that no amount of strength in other dimensions can mask a fundamental brand safety problem. Learn more on the methodology page.
Micro-Influencer
A creator with a follower count typically between 10,000 and 100,000 across their primary platform. Micro-influencers often deliver higher engagement rates and more niche audience targeting compared to larger creators. CreatorScore uses tier normalization to benchmark micro-influencers against peers of similar size rather than against mega-creators, ensuring fair and accurate scoring. See our ROI benchmarks for tier-specific performance data.
Macro-Influencer
A creator with a follower count typically between 100,000 and 1 million. Macro-influencers offer broad reach but may have lower per-follower engagement compared to micro-influencers. CreatorScore evaluates macro-influencers with tier-appropriate benchmarks, recognizing that reply rates, comment depth, and engagement patterns differ at scale. The 7 dimensions of creator quality apply across all tiers.
NSFW Detection
Automated identification of "not safe for work" content including nudity, sexually explicit material, and graphic imagery. CreatorScore employs both visual analysis models (for images and video frames) and text-based NLP to detect NSFW content across a creator's posts. NSFW content is weighted at 25% of the Content Risk Score, and extreme levels (>95%) trigger a knockout factor capping the score at 35/100.
ROI Prediction
CreatorScore's forward-looking agent that estimates the likely return on investment from partnering with a creator. Weighted at 10% of the overall score, it combines engagement quality (40%), community health (25%), and growth trajectory (35%). ROI Prediction incorporates shares and saves data where available and uses blended velocity metrics across 30, 60, and 90-day windows. See our 2026 ROI benchmarks for industry context.
Sentiment Analysis
The use of natural language processing to classify the emotional tone of comments, captions, and audience interactions as positive, negative, or neutral. CreatorScore's Sentiment Agent combines sentiment stability (how consistent the tone is over time) with audience sentiment (how the community feels about the creator) in a 50/50 blend. The system uses LLM-based reclassification to correct false negatives from automated models, particularly for sarcasm, humor, and context-dependent language.
SHAP Explainability
A machine learning interpretability technique (SHapley Additive exPlanations) used by CreatorScore to explain why a creator received a particular score. SHAP values quantify the contribution of each scoring component — both positive drivers and areas of concern — making the scoring process transparent and auditable. Each component appears as either a positive or negative driver, never both, ensuring clear and non-contradictory explanations for brands reviewing a creator's profile.
Tier Normalization
The process of adjusting scoring benchmarks based on a creator's audience size tier (nano, micro, mid, macro, mega). A micro-influencer with a 5% engagement rate is performing well for their tier, while a mega-creator with the same rate would be exceptional. Tier normalization ensures fair comparisons by evaluating creators against peers of similar scale. It also adjusts penalties — for example, mega creators receive a maximum 25% generic reply penalty versus 100% for smaller creators.
Unified Score
A single score that represents a creator's brand safety across all their linked social media platforms, calculated as a follower-weighted average of per-platform scores within an identity group. The unified score gives brands a holistic view of a creator's risk profile rather than requiring separate evaluation of each platform. Knockout factors are applied at the per-platform level before unification, ensuring severe issues on any single platform are reflected in the final number.
See these concepts in action
CreatorScore applies every concept in this glossary to produce a single, transparent brand safety score for any creator. Explore our methodology or start scoring today.