cotera

Twitter Brand Sentiment Tracker

Track what people are saying about your brand on Twitter. Log every mention to a spreadsheet and get Slack alerts when sentiment turns negative.

Brand monitoringSentiment analysisCrisis detectionSocial analytics

The Challenge

Your brand is being discussed on Twitter right now. Some of those conversations are glowing endorsements, others are frustrated complaints that could snowball. Manually scrolling through search results and trying to gauge the overall vibe is slow and unreliable. You need a structured way to track sentiment over time and get alerted before a small problem becomes a PR crisis.

What This Prompt Does

Search & Classify

Find brand mentions and tag each as positive, negative, or neutral

Weight by Influence

Factor in follower counts so high-reach complaints get flagged first

Log to Sheets

Append every mention as a structured row for trend analysis

Alert on Slack

Get instant notifications when negative sentiment spikes

The Prompt

The Prompt

Task

Use @Twitter/Search TweetsName it "Twitter/Search Tweets" and call it with @Twitter/Search Tweets to find recent mentions of a brand or product. Classify each mention as positive, negative, or neutral. Log structured results to @Google Sheets/Append RowName it "Google Sheets/Append Row" and call it with @Google Sheets/Append Row. If the negative ratio exceeds a threshold, fire an alert via @Slack/Send MessageName it "Slack/Send Message" and call it with @Slack/Send Message.

Example: Track sentiment for "Stripe" on Twitter, log every mention to a spreadsheet, and alert #brand-alerts on Slack if negative mentions spike above 20%.

Input

The user will provide:

  1. Brand or product name to monitor
  2. Google Sheets URL for logging
  3. Slack channel for negative-spike alerts
  4. Optional: negative threshold percentage (default 20%)

Example: "Track Twitter sentiment for Acme. Log to https://docs.google.com/spreadsheets/d/abc123. Alert #brand-alerts if negatives exceed 25%."

Context

Search Strategy

  1. Search Twitter for the brand name, common misspellings, and product-specific hashtags
  2. Pull the latest 40 tweets using the "Latest" type filter
  3. For each tweet, classify sentiment based on language cues:
    • Positive: praise, recommendation, excitement, gratitude
    • Negative: complaint, frustration, bug report, churn signal
    • Neutral: informational, question, retweet without comment
  4. Use @Twitter/Get ProfileName it "Twitter/Get Profile" and call it with @Twitter/Get Profile on high-engagement authors to note follower counts — a negative tweet from a 50k-follower account matters more than one from 50 followers
  5. Append each mention as a row in Google Sheets with: date, author, follower count, tweet text, sentiment, engagement score, link
  6. Calculate the negative ratio for this batch — if it exceeds the threshold, send a Slack alert

Sentiment Classification Guidelines

  • Look at the full tweet, not just keywords. "I can't believe how good Acme is" is positive despite "can't believe."
  • Sarcasm is tricky — if ambiguous, label as "Needs Review"
  • Retweets of negative content count as signal amplification, not separate negative mentions
  • Replies in a thread should be evaluated individually

What Counts as Alert-Worthy

Immediate Slack alert:

  • Negative ratio exceeds the set threshold
  • Any single negative tweet with >500 likes
  • Negative mention from a verified account or >20k followers

Weekly summary (append to sheet):

  • Overall sentiment breakdown
  • Top 5 most-engaged tweets
  • Trend direction vs. previous week

Output

Google Sheets Row Format: | Date | Author | Followers | Tweet | Sentiment | Likes | Retweets | Link | |------|--------|-----------|-------|-----------|-------|----------|------|

Slack Alert Format:

Sentiment Spike Alert: [Brand] Negative mention ratio: [X]% (threshold: [Y]%)

Top negative tweets:

  1. @[handle] ([followers] followers): "[excerpt]" -- [likes] likes [link]
  2. @[handle] ([followers] followers): "[excerpt]" -- [likes] likes [link]

Batch summary: [positive]% positive, [neutral]% neutral, [negative]% negative ([total] tweets scanned)

Recommended action: [Brief recommendation based on the nature of complaints]

Example Usage

Try asking:

  • →"Track Twitter sentiment for our brand and log it to this spreadsheet"
  • →"Monitor what people are saying about Acme on Twitter, alert me if negatives hit 15%"
  • →"Set up a daily Twitter sentiment report for our product launch"