{"id":25381,"date":"2026-02-08T21:06:27","date_gmt":"2026-02-09T03:06:27","guid":{"rendered":"https:\/\/www.designandexecute.com\/designs\/?p=25381"},"modified":"2026-02-25T07:46:57","modified_gmt":"2026-02-25T13:46:57","slug":"sentiment-is-the-most-abused-field-in-data-models","status":"publish","type":"post","link":"https:\/\/www.designandexecute.com\/designs\/sentiment-is-the-most-abused-field-in-data-models\/","title":{"rendered":"Sentiment is the most abused field in data models"},"content":{"rendered":"\n

1. What sentiment_score<\/code> really measures<\/h2>\n\n\n\n
\n

Sentiment score = emotional valence<\/em><\/strong>
Positive \u2194 Neutral \u2194 Negative<\/p>\n<\/blockquote>\n\n\n\n

It answers:<\/p>\n\n\n\n

\n

\u201cHow does the person feel<\/strong> about this topic right now?\u201d<\/em><\/p>\n<\/blockquote>\n\n\n\n

Not:<\/p>\n\n\n\n

    \n
  • How important it is<\/li>\n\n\n\n
  • How much effort they\u2019re putting in<\/li>\n\n\n\n
  • Whether they\u2019ll act<\/li>\n<\/ul>\n\n\n\n

    You can have:<\/p>\n\n\n\n

      \n
    • High engagement + negative sentiment<\/strong><\/li>\n\n\n\n
    • Low engagement + positive sentiment<\/strong><\/li>\n<\/ul>\n\n\n\n

      That\u2019s why this field matters.<\/p>\n\n\n\n

      \n\n\n\n

      2. Recommended scale (keep it simple)<\/h2>\n\n\n\n

      Option A \u2014 Discrete scale (best for explainability)<\/h3>\n\n\n\n
      Score<\/th>Label<\/th>Meaning<\/th><\/tr><\/thead>
      -2<\/strong><\/td>Strongly Negative<\/td>Frustration, anger, resentment<\/td><\/tr>
      -1<\/strong><\/td>Negative<\/td>Dislike, concern, reluctance<\/td><\/tr>
      0<\/strong><\/td>Neutral \/ Mixed<\/td>Informational, balanced<\/td><\/tr>
      +1<\/strong><\/td>Positive<\/td>Enjoyment, optimism<\/td><\/tr>
      +2<\/strong><\/td>Strongly Positive<\/td>Excitement, pride, love<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

      This is human-legible and stable over time.<\/p>\n\n\n\n

      \n\n\n\n

      Option B \u2014 Normalized (-1.0 \u2192 +1.0)<\/h3>\n\n\n\n

      Use only if you\u2019re piping NLP outputs directly.<\/p>\n\n\n\n

      Range<\/th>Interpretation<\/th><\/tr><\/thead>
      -1.0 to -0.6<\/td>Strong negative<\/td><\/tr>
      -0.59 to -0.2<\/td>Mild negative<\/td><\/tr>
      -0.19 to +0.19<\/td>Neutral<\/td><\/tr>
      +0.2 to +0.59<\/td>Mild positive<\/td><\/tr>
      +0.6 to +1.0<\/td>Strong positive<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

      You can always bucket it later.<\/p>\n\n\n\n

      \n\n\n\n

      3. Examples (this is the important part)<\/h2>\n\n\n\n

      A. Same topic, different sentiment<\/h3>\n\n\n\n

      Topic: Work \/ Career<\/strong><\/h4>\n\n\n\n
      Utterance<\/th>sentiment_score<\/th>Why<\/th><\/tr><\/thead>
      \u201cWork is fine.\u201d<\/td>0<\/td>Flat, informational<\/td><\/tr>
      \u201cI like my job.\u201d<\/td>+1<\/td>Mild positive<\/td><\/tr>
      \u201cI love what I\u2019m building.\u201d<\/td>+2<\/td>Strong positive<\/td><\/tr>
      \u201cI\u2019m stuck at work.\u201d<\/td>-1<\/td>Frustration<\/td><\/tr>
      \u201cI hate my job.\u201d<\/td>-2<\/td>Strong negative<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

      Engagement could be high in all of these.<\/p>\n\n\n\n

      \n\n\n\n

      B. Same sentiment, different engagement<\/h3>\n\n\n\n

      Topic: Health & Gyms<\/strong><\/h4>\n\n\n\n
      Utterance<\/th>Sentiment<\/th>Engagement<\/th><\/tr><\/thead>
      \u201cGyms are great.\u201d<\/td>+1<\/td>1<\/td><\/tr>
      \u201cI train 5x a week and track PRs.\u201d<\/td>+1<\/td>5<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

      Sentiment = feeling
      Engagement = energy<\/p>\n\n\n\n

      \n\n\n\n

      C. Business context examples<\/h3>\n\n\n\n

      Topic: Project Management<\/strong><\/h4>\n\n\n\n
      Utterance<\/th>sentiment_score<\/th><\/tr><\/thead>
      \u201cWe\u2019re on track.\u201d<\/td>+1<\/td><\/tr>
      \u201cThis is harder than expected.\u201d<\/td>-1<\/td><\/tr>
      \u201cThis project is a mess.\u201d<\/td>-2<\/td><\/tr>
      \u201cChallenging, but solvable.\u201d<\/td>0<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
      \n\n\n\n

      Topic: Negotiations & Proposals<\/strong><\/h4>\n\n\n\n
      Utterance<\/th>sentiment<\/th><\/tr><\/thead>
      \u201cLet\u2019s explore options.\u201d<\/td>0<\/td><\/tr>
      \u201cI\u2019m excited to move forward.\u201d<\/td>+2<\/td><\/tr>
      \u201cThis pricing doesn\u2019t work.\u201d<\/td>-1<\/td><\/tr>
      \u201cThis is unacceptable.\u201d<\/td>-2<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
      \n\n\n\n

      D. Personal & emotional topics<\/h3>\n\n\n\n

      Topic: Emotions \/ Love<\/strong><\/h4>\n\n\n\n
      Utterance<\/th>sentiment_score<\/th><\/tr><\/thead>
      \u201cI\u2019m okay.\u201d<\/td>0<\/td><\/tr>
      \u201cI feel close to her.\u201d<\/td>+2<\/td><\/tr>
      \u201cI\u2019m confused.\u201d<\/td>-1<\/td><\/tr>
      \u201cI\u2019m heartbroken.\u201d<\/td>-2<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

      This is where sentiment is most valuable<\/em>.<\/p>\n\n\n\n

      \n\n\n\n

      4. Common modeling mistakes (avoid these)<\/h2>\n\n\n\n

      \u274c Mistake 1: Using sentiment as importance<\/h3>\n\n\n\n
      \n

      \u201cThis is a big issue\u201d \u2260 negative sentiment<\/p>\n<\/blockquote>\n\n\n\n

      \u274c Mistake 2: Using sentiment as an outcome<\/h3>\n\n\n\n
      \n

      \u201cWe closed the deal\u201d \u2260 positive sentiment<\/p>\n<\/blockquote>\n\n\n\n

      \u274c Mistake 3: Averaging sentiment blindly<\/h3>\n\n\n\n
      \n

      A -2 followed by +2 is not \u201cneutral\u201d \u2014 it\u2019s volatility<\/strong>.<\/p>\n<\/blockquote>\n\n\n\n

      \n\n\n\n

      5. How sentiment works with your existing fields<\/h2>\n\n\n\n
      Field<\/th>Question it answers<\/th><\/tr><\/thead>
      Topic<\/td>What are we talking about<\/td><\/tr>
      Intent<\/td>Why now<\/td><\/tr>
      Engagement score<\/td>How invested<\/td><\/tr>
      Sentiment score<\/strong><\/td>How it feels<\/td><\/tr>
      Outcome<\/td>What happened<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

      This separation is why your model is actually strong.<\/p>\n\n\n\n

      \n\n\n\n

      6. Design recommendation (this is subtle but important)<\/h2>\n\n\n\n

      Store:<\/p>\n\n\n\n

        \n
      • sentiment_score<\/code><\/li>\n\n\n\n
      • sentiment_confidence<\/code> (low \/ medium \/ high)<\/li>\n<\/ul>\n\n\n\n

        Why?<\/p>\n\n\n\n

        \n

        Some statements are emotionally clear. Some are ambiguous.<\/p>\n<\/blockquote>\n\n\n\n

        Example:<\/p>\n\n\n\n

          \n
        • \u201cInteresting.\u201d \u2192 sentiment unclear \u2192 low confidence<\/li>\n<\/ul>\n\n\n\n
          \n\n\n\n

          Summary<\/h2>\n\n\n\n
            \n
          • Sentiment = emotional direction<\/strong><\/li>\n\n\n\n
          • Use -2 to +2<\/strong><\/li>\n\n\n\n
          • Keep it independent of engagement & outcome<\/li>\n\n\n\n
          • Anchor it to language, not interpretation<\/li>\n<\/ul>\n\n\n\n

            \/\/todo<\/p>\n\n\n\n

              \n
            • Combine sentiment \u00d7 engagement<\/strong> into a \u201cheat map.\u201d<\/li>\n\n\n\n
            • Model sentiment drift over time<\/strong><\/li>\n\n\n\n
            • Or detect hidden risk<\/strong> (high engagement + negative sentiment)<\/li>\n<\/ul>\n\n\n\n

              That\u2019s where this turns into a real signal, not just labels.<\/p>\n\n\n\n

              \n\n\n\n

              1. What engagement_score<\/code> really measures<\/h2>\n\n\n\n
              \n

              Engagement score = how much energy + commitment<\/em> is present in the interaction.<\/strong><\/p>\n<\/blockquote>\n\n\n\n

              Not sentiment.
              Not importance.
              Not outcome.<\/p>\n\n\n\n

              It answers:<\/p>\n\n\n\n

              \n

              \u201cHow invested is the person in this conversation right now?\u201d<\/em><\/p>\n<\/blockquote>\n\n\n\n

              Think attention \u00d7 effort \u00d7 continuity<\/strong>.<\/p>\n\n\n\n

              \n\n\n\n

              2. Signal strength (the inputs)<\/h2>\n\n\n\n

              signal_strength<\/code> is evidence<\/em>, not judgment.<\/p>\n\n\n\n

              Common observable signals<\/h3>\n\n\n\n
              Signal<\/th>Examples<\/th><\/tr><\/thead>
              Time<\/td>Long messages, long calls, repeat back-and-forth<\/td><\/tr>
              Depth<\/td>Personal detail, specificity, vulnerability<\/td><\/tr>
              Effort<\/td>Preparation, documents shared, follow-ups<\/td><\/tr>
              Agency<\/td>Decisions made, commitments stated<\/td><\/tr>
              Continuity<\/td>Refers to past conversations, future plans<\/td><\/tr>
              Risk<\/td>Disagreement, asking for advice, asking for help<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

              Each of these increases signal strength.<\/p>\n\n\n\n

              \n\n\n\n

              3. Example scales (practical, not academic)<\/h2>\n\n\n\n

              Option A \u2014 Simple 1\u20135 scale (recommended)<\/h3>\n\n\n\n

              This works shockingly well.<\/p>\n\n\n\n

              Score<\/th>Label<\/th>Observable signals<\/th><\/tr><\/thead>
              1<\/strong><\/td>Passive<\/td>Polite replies, small talk, low effort<\/td><\/tr>
              2<\/strong><\/td>Light<\/td>Casual interest, short responses<\/td><\/tr>
              3<\/strong><\/td>Engaged<\/td>Asking questions, staying on topic<\/td><\/tr>
              4<\/strong><\/td>Invested<\/td>Sharing context, opinions, time<\/td><\/tr>
              5<\/strong><\/td>Committed<\/td>Decisions, plans, emotional or professional stake<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
              \n\n\n\n

              Option B \u2014 Expanded 0\u2013100 (only if you need math)<\/h3>\n\n\n\n

              Use if you plan analytics.<\/p>\n\n\n\n

              Range<\/th>Meaning<\/th><\/tr><\/thead>
              0\u201320<\/td>Noise<\/td><\/tr>
              21\u201340<\/td>Awareness<\/td><\/tr>
              41\u201360<\/td>Interest<\/td><\/tr>
              61\u201380<\/td>Engagement<\/td><\/tr>
              81\u2013100<\/td>Commitment<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

              Internally, you can still bucket<\/strong> it to 1\u20135.<\/p>\n\n\n\n

              \n\n\n\n

              4. Mapping signal strength \u2192 engagement_score<\/h2>\n\n\n\n

              You can model it explicitly.<\/p>\n\n\n\n

              Example signals \u2192 score<\/h3>\n\n\n\n
              Signals present<\/th>engagement_score<\/th><\/tr><\/thead>
              Topic hopping, generic replies<\/td>1<\/td><\/tr>
              Jokes, light banter<\/td>2<\/td><\/tr>
              Asking follow-ups<\/td>3<\/td><\/tr>
              Sharing constraints, goals<\/td>4<\/td><\/tr>
              Making decisions \/ asking next steps<\/td>5<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
              \n\n\n\n

              5. Real examples from your domain<\/h2>\n\n\n\n

              Example 1: Salsa Dancing<\/h3>\n\n\n\n
                \n
              • \u201cYeah I dance sometimes\u201d \u2192 1<\/strong><\/li>\n\n\n\n
              • \u201cI love salsa, where do you dance?\u201d \u2192 3<\/strong><\/li>\n\n\n\n
              • \u201cI\u2019m training for a festival, practicing twice a week\u201d \u2192 5<\/strong><\/li>\n<\/ul>\n\n\n\n

                Topic stayed the same.
                Engagement changed drastically.<\/p>\n\n\n\n

                \n\n\n\n

                Example 2: Business \u2014 Marketing Campaigns<\/h3>\n\n\n\n
                  \n
                • \u201cSend me the deck\u201d \u2192 2<\/strong><\/li>\n\n\n\n
                • \u201cWhat would success look like?\u201d \u2192 3<\/strong><\/li>\n\n\n\n
                • \u201cLet\u2019s launch next month, budget is X\u201d \u2192 5<\/strong><\/li>\n<\/ul>\n\n\n\n
                  \n\n\n\n

                  Example 3: Personal \u2014 Aspirational<\/h3>\n\n\n\n
                    \n
                  • \u201cI\u2019ve been thinking about my career\u201d \u2192 3<\/strong><\/li>\n\n\n\n
                  • \u201cI\u2019m unhappy and need to change direction\u201d \u2192 4<\/strong><\/li>\n\n\n\n
                  • \u201cI\u2019m committing to a 90-day pivot\u201d \u2192 5<\/strong><\/li>\n<\/ul>\n\n\n\n
                    \n\n\n\n

                    6. Where this fits in your model<\/h2>\n\n\n\n

                    Think in layers:<\/p>\n\n\n\n

                      \n
                    • Topic<\/strong> \u2192 What<\/li>\n\n\n\n
                    • Intent<\/strong> \u2192 Why now<\/li>\n\n\n\n
                    • Learning touchpoint<\/strong> \u2192 Where in the arc<\/li>\n\n\n\n
                    • Engagement score<\/strong> \u2192 How much energy<\/li>\n\n\n\n
                    • Signal strength<\/strong> \u2192 Why you believe that score<\/li>\n<\/ul>\n\n\n\n

                      This keeps the model explainable<\/strong>, not magical.<\/p>\n\n\n\n

                      \n\n\n\n

                      7. Design advice (important)<\/h2>\n\n\n\n
                        \n
                      • Don\u2019t over-optimize this.<\/li>\n\n\n\n
                      • Human-readable beats perfect.<\/li>\n\n\n\n
                      • Engagement is situational<\/strong>, not identity.<\/li>\n<\/ul>\n\n\n\n

                        A person can be:<\/p>\n\n\n\n

                          \n
                        • Engagement = 5<\/code> in Salsa<\/em><\/li>\n\n\n\n
                        • Engagement = 1<\/code> in Investing<\/em><\/li>\n<\/ul>\n\n\n\n

                          That\u2019s not inconsistency \u2014 that\u2019s truth.<\/p>\n\n\n\n

                          \n\n\n\n

                          Summary<\/h2>\n\n\n\n
                            \n
                          • Signal strength<\/strong> = observable evidence<\/li>\n\n\n\n
                          • Engagement score<\/strong> = normalized judgment of investment<\/li>\n\n\n\n
                          • Use 1\u20135<\/strong> unless you really<\/em> need more<\/li>\n\n\n\n
                          • Anchor every score to something you could point to<\/li>\n<\/ul>\n\n\n\n

                            \/\/todo<\/p>\n\n\n\n

                              \n
                            • Define decay over time<\/li>\n\n\n\n
                            • Weight engagement by intent<\/li>\n\n\n\n
                            • design a query like: \u201cWhich topics generate high engagement but low outcomes?\u201d<\/em><\/li>\n<\/ul>\n\n\n\n

                              lethal design in a good way.<\/p>\n\n\n\n

                              Engagement is situational, but Energy and Effectiveness is Person Centric<\/h2>\n\n\n\n

                              ENERGY ASSESSMENT (Circle 0-10)<\/p>\n\n\n\n

                                \n
                              • Pre-interaction mood: Dread [0 1 2 3 4 5 6 7 8 9 10] Anticipation<\/li>\n\n\n\n
                              • Post-interaction feeling: Drained [0 1 2 3 4 5 6 7 8 9 10] Energized<\/li>\n\n\n\n
                              • Conversation flow: Strained [0 1 2 3 4 5 6 7 8 9 10] Natural<\/li>\n\n\n\n
                              • Emotional safety: Unsafe [0 1 2 3 4 5 6 7 8 9 10] Safe<\/li>\n<\/ul>\n\n\n\n

                                EFFECTIVENESS ASSESSMENT (Circle 0-10)<\/p>\n\n\n\n

                                  \n
                                • Follow-through: Never [0 1 2 3 4 5 6 7 8 9 10] Always<\/li>\n\n\n\n
                                • Value created: None [0 1 2 3 4 5 6 7 8 9 10] Transformative<\/li>\n\n\n\n
                                • Proactivity: Passive [0 1 2 3 4 5 6 7 8 9 10] Initiator<\/li>\n\n\n\n
                                • Efficiency: Time-waster [0 1 2 3 4 5 6 7 8 9 10] Multiplier<\/li>\n<\/ul>\n\n\n\n

                                  Usage: Scoring and creating a recommendation engine to determine a contact frequency<\/h3>\n\n\n\n

                                  The scores mean:<\/p>\n\n\n\n

                                    \n
                                  • E: 73<\/strong> = Engagement Score<\/strong> (0-100) – Measures encounter frequency and patterns. Calculated from total encounters, recent activity (last 90 days), channel diversity, and consistency.<\/li>\n\n\n\n
                                  • R: 32<\/strong> = Relationship Score<\/strong> (0-100) – Measures relationship quality based on NAP intimacy level, NAP impact level, and relationship category (Power\/Master Mind gets bonus points).<\/li>\n\n\n\n
                                  • Rc: 95<\/strong> = Recency Score<\/strong> (0-100) – Measures recent interaction patterns. Based on activity in the last 30 days and days since last contact.<\/li>\n<\/ul>\n\n\n\n

                                    These three scores are combined with weighted averages to determine the recommended contact frequency for that person. Higher scores generally indicate that more frequent contact is warranted.<\/p>\n","protected":false},"excerpt":{"rendered":"

                                    1. What sentiment_score really measures Sentiment score = emotional valencePositive \u2194 Neutral \u2194 Negative It answers: \u201cHow does the person feel about this topic right now?\u201d Not: You can have: That\u2019s why this field matters. 2. Recommended scale (keep it simple) Option A \u2014 Discrete scale (best for explainability) Score Label Meaning -2 Strongly Negative […]<\/p>\n","protected":false},"author":2,"featured_media":19260,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[32,31],"tags":[],"class_list":["post-25381","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-bi-dashboards-analytics","category-bi-data-warehouse"],"jetpack_featured_media_url":"https:\/\/www.designandexecute.com\/designs\/wp-content\/uploads\/2023\/11\/The_Zachman_Framework_of_Enterprise_Architecture.jpg","_links":{"self":[{"href":"https:\/\/www.designandexecute.com\/designs\/wp-json\/wp\/v2\/posts\/25381","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.designandexecute.com\/designs\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.designandexecute.com\/designs\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.designandexecute.com\/designs\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.designandexecute.com\/designs\/wp-json\/wp\/v2\/comments?post=25381"}],"version-history":[{"count":5,"href":"https:\/\/www.designandexecute.com\/designs\/wp-json\/wp\/v2\/posts\/25381\/revisions"}],"predecessor-version":[{"id":25392,"href":"https:\/\/www.designandexecute.com\/designs\/wp-json\/wp\/v2\/posts\/25381\/revisions\/25392"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.designandexecute.com\/designs\/wp-json\/wp\/v2\/media\/19260"}],"wp:attachment":[{"href":"https:\/\/www.designandexecute.com\/designs\/wp-json\/wp\/v2\/media?parent=25381"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.designandexecute.com\/designs\/wp-json\/wp\/v2\/categories?post=25381"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.designandexecute.com\/designs\/wp-json\/wp\/v2\/tags?post=25381"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}