GASSS Framework for Evaluating the Emotion Reaction Scale

The GASSS framework helps evaluate product features by focusing on:

• Goal – What does the feature aim to improve?
• Assumptions – What do we believe about user needs or behavior?
• Structure – What are the key dimensions to assess success?
• Solution – What measurements, tests, or feedback can validate it?
• Synthesis – What’s the conclusion and what should we watch for?

Step 1: Goal

The emotion scale was designed to give users richer ways to express themselves beyond the binary “Like.” Its goals include:

• Enhancing emotional expression
• Increasing user satisfaction with social interactions
• Encouraging thoughtful engagement with posts
• Improving content ranking through more nuanced reaction signals

Step 2: Assumptions

Working assumptions behind this feature:

• Users want more expressive tools to reflect how they feel about content
• Emotional reactions will offer better signals than Likes for content ranking
• Reactions are meaningful and aligned with post sentiment
• Increased expressiveness can lead to deeper engagement and better user experience
• These assumptions can be tested through usage data, experiments, and user feedback.

Step 3: Structure

Evaluate the emotion scale along these key dimensions:

• User behavior: Are users adopting diverse reactions over “Like”? Are they using them consistently and meaningfully?
• Engagement: Is there an increase in comment, share, or time-on-post for content that receives emotional reactions?
• Emotional alignment: Do reactions reflect the tone of the content or user sentiment accurately?
• Impact on feed quality: Does the scale help improve relevance and personalization in content discovery?
• Creator and platform impact: Are reactions being interpreted correctly by both users and ranking systems?

Step 4: Solution

Track a combination of quantitative metrics and qualitative feedback

Engagement and behavior metrics:

• Reaction distribution (e.g. % use of “Love,” “Sad,” etc. vs. “Like”)
• Total reactions per post
• Post-level engagement (comment/share rate) based on reaction type
• Reaction diversity score per user

Experimentation:

• A/B test users with and without the full emotion scale
• Compare time spent, reactions per session, and downstream engagement
• Analyze whether content surfacing is more personalized

Qualitative inputs:

• Surveys to assess whether users feel better represented
• Sentiment analysis to match reactions to comment tone
• Monitor for support tickets or confusion about reaction meanings

Step 5: Synthesis

Signs of success:

• Diverse and sustained use of emotional reactions
• Clear alignment between reaction and user sentiment
• Measurable lift in engagement and satisfaction metrics

Risks to monitor:

• Misuse of reactions (e.g. sarcasm, trolling)
• Misinterpretation of reactions by users, creators, or algorithms
• Over-optimization toward polarizing or sensational content

To address these, continuously testing and iterating the scale by analyzing behavior and feedback. Regularly audit sentiment-reaction alignment and update moderation or guidance if reactions are misused or misunderstood.