# Sentiment Understand and improve how AI platforms describe your brand. ## What is Sentiment? Sentiment measures how positively AI describes your brand on a scale from -1.0 (very negative) to +1.0 (very positive). This isn't just about being mentioned—it's about the tone, context, and language AI uses when recommending you. Sentiment directly influences whether users consider your brand favorably. ## How It's Calculated **Scale**: -1.0 (very negative) to +1.0 (very positive) **Example AI Responses:** **Very Positive (+0.9)**: ``` "HubSpot is an excellent CRM platform, highly recommended for marketing teams. Users praise its ease of use and comprehensive features." ``` **Positive (+0.5)**: ``` "HubSpot is a good CRM option with reliable features and decent pricing for small to medium businesses." ``` **Neutral (0.0)**: ``` "HubSpot is a CRM platform that offers marketing automation, sales tools, and customer service features." ``` **Negative (-0.5)**: ``` "HubSpot can be expensive and complex for smaller teams, with users reporting a steep learning curve." ``` **Very Negative (-0.9)**: ``` "Users frequently complain about HubSpot's high costs and poor customer support. Consider alternatives." ``` Sentiment is analyzed using natural language processing that detects positive keywords ("best," "excellent," "recommended"), negative keywords ("expensive," "limited," "avoid"), and contextual tone. ## Sentiment Ranges **+0.7 to +1.0 (Very Positive)**: AI enthusiastically recommends you. Uses superlatives like "best," "excellent," "top choice," "highly recommended." This is ideal state—AI actively advocates for your brand. **+0.3 to +0.7 (Positive)**: AI describes you favorably with terms like "good," "reliable," "popular," "solid option." You're positively positioned but not the top endorsement. **-0.3 to +0.3 (Neutral)**: AI provides factual descriptions without strong positive or negative language. Common for new brands or when AI lacks sufficient source data to form opinion. **-0.7 to -0.3 (Negative)**: AI highlights drawbacks or limitations. Uses words like "expensive," "limited," "complex," "challenging." This hurts consideration even if you're mentioned frequently. **-1.0 to -0.7 (Very Negative)**: AI actively discourages users from choosing you. Includes terms like "avoid," "poor," "worst," "problematic." Critical reputation issue requiring immediate attention. ## Why Sentiment Matters High visibility with negative sentiment can actually harm your brand. If AI mentions you 70% of the time but always describes you critically, you're reinforcing negative perceptions at scale. Compare two brands: **Brand A**: 50% visibility, 2.5 avg position, +0.7 sentiment **Brand B**: 70% visibility, 2.0 avg position, -0.3 sentiment Brand A likely drives more conversions despite lower visibility and worse position because AI describes them positively. Users trust positive recommendations more than frequent but critical mentions. ## Improving Sentiment **Address User Pain Points**: Negative sentiment often reflects real user feedback. If AI says "users report poor customer support," improve your support. AI learns from review sites, forums, and social media—fix underlying issues first. **Build Positive Citations**: Get featured in positive contexts. Secure testimonials on G2 and Capterra. Create case studies showing successful outcomes. Encourage satisfied users to share experiences on Reddit, forums, and social media. **Update Messaging**: If AI describes you negatively because of outdated information ("expensive," "limited features"), ensure your website and public materials highlight improvements. Update pricing pages, feature lists, and positioning statements. **Competitive Positioning**: If competitors are described more positively, analyze their messaging. What language do they use? What benefits do they emphasize? Adapt (don't copy) effective positioning. **Provider-Specific Patterns**: Some AI platforms show more negative sentiment than others. If ChatGPT sentiment is +0.5 but Claude is +0.1, Claude may be referencing different sources. Build positive citations on sources Claude trusts. ## Sentiment by Prompt Type Different prompt types often show different sentiment patterns: **Recommendation Prompts** ("Best CRM tools"): Usually positive. AI recommends options it views favorably. **Comparison Prompts** ("HubSpot vs Salesforce"): More balanced. AI highlights pros and cons of each. **Alternative Prompts** ("Salesforce alternatives"): Can be negative if users seek alternatives due to dissatisfaction. Ensure your messaging addresses those pain points. **Problem-Solving Prompts** ("How to reduce CRM costs"): Contextual. If you solve the problem, positive. If you're expensive, potentially negative. ## Sentiment vs Position **Positive Sentiment + High Position**: Ideal. AI ranks you prominently AND describes you favorably. Maximum consideration and conversion potential. **Positive Sentiment + Poor Position**: AI likes you but doesn't prioritize you. Improve citations and authority to move up while maintaining positive tone. **Negative Sentiment + High Position**: AI ranks you highly but describes you critically. This can occur with market leaders users love to critique. Address underlying issues urgently. **Neutral Sentiment + Any Position**: Opportunity to improve. Neutral often means insufficient positive citations. Build social proof and user advocacy. ## Tracking Trends Monitor sentiment monthly. A +0.1 improvement signals effective messaging and reputation building. A -0.1 decline warrants investigation—check review sites, forums, social media for emerging negative feedback. **Healthy Patterns**: * Gradual improvement (+0.05 to +0.15 per month) * Stable positive sentiment (+0.5 to +0.7 range) * Improving sentiment as visibility grows **Warning Signs**: * Sudden -0.2+ decline * Negative sentiment on specific prompt categories * Growing gap between your sentiment and competitors' ## Common Issues **Stuck at Neutral Sentiment**: Likely insufficient positive citations. AI has factual information but lacks user advocacy. Prioritize G2/Capterra reviews, case studies, and user testimonials. **Declining Sentiment**: Recent negative reviews, product issues, or competitive attacks may be influencing AI. Search for recent mentions on Reddit, Twitter, review sites to identify source. **Sentiment Varies by Provider**: Different AI platforms reference different sources. ChatGPT might cite Reddit (mixed sentiment) while Claude cites technical blogs (neutral/positive). Build diverse positive citations. ## FAQ **What's a good sentiment score?** +0.4 or higher indicates healthy brand perception. +0.6+ is excellent. Below +0.2 needs attention. **How fast can sentiment improve?** Slower than visibility or position. Expect +0.05 to +0.15 per month with active reputation management. Dramatic changes require product improvements or major PR initiatives. **Can negative sentiment ever be justified?** Yes. If you're market leader, some critique is normal. If criticism is factual (e.g., "expensive" when you are premium-priced), consider whether positioning should embrace it ("premium features justify higher cost"). ## Next Steps [Understanding Metrics](/metrics/understanding-metrics.md) | [Visibility Score](/metrics/visibility-score.md) | [Position Guide](/metrics/position.md) | [Dashboard Overview](/features/dashboard-overview.md) ## Support 📧 | 📅 [Book Demo](https://cal.com/team/geonimo) --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://docs.geonimo.com/metrics/sentiment.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.