What it means
The process begins by collecting public posts that mention a chosen topic, such as a company name or a share ticker (the short code used to identify a listed share). Software then classifies each message by its tone, using word lists or machine learning models that have learned from labelled examples.
The results are totalled over a period and displayed as a score or chart. Different users have different goals.
Traders look for changes in mood that might precede price moves, risk teams watch for a wave of negative comment about their own company, and marketing teams track reaction to a launch. The same underlying data can answer all three questions.
Interpreting the numbers takes care. Sarcasm, slang and emojis confuse automated classification, and a sudden increase in volume may come from a small group of active accounts or from automated bots.
A reading is more reliable when it is based on a large number of messages from many distinct users. Sentiment is best seen as a complement to conventional analysis.
Studies and practitioners have found that online mood can sometimes lead or coincide with market moves, but the relationship is unstable and often disappears after it becomes widely known. Used alone, it can produce false signals, especially around news events.
Governance matters too. Teams should record where the data comes from, how it is cleaned and what the indicator has historically predicted, so that decisions based on it can be explained.
A sentiment score is a tool for asking better questions, not a substitute for judgement. Time frame matters as well.
A one-day spike in negative comment may fade before it affects anything, whereas a steady decline over several weeks is more likely to reflect a real change in how customers or investors feel.
In practice
Real-world examples.
Example
A fund analyst tracks a social sentiment score for a consumer electronics company before its results. The score drops from 35% to 8% after reports of a faulty product. She trims her position slightly and waits for the official results. When the results arrive, sales are below the forecast, and she is glad she did not rely on the score alone to justify a large trade.
Example
A bank's communications team notices that net sentiment about its mobile app has fallen sharply after an update. It identifies the cause as a login problem and fixes it within a day. Sentiment recovers within the week.
Example
A restaurant group compares sentiment scores for each of its branches. One location consistently scores 20 points below the others, with comments about slow service. Management investigates and replaces the shift manager, after which the score improves.
Formula
Calculation
Net sentiment score = (Positive mentions - Negative mentions) / Total mentions x 100
Suppose a company is mentioned 10,000 times in a week, with 6,000 positive messages, 2,000 negative messages and 2,000 neutral ones. Net sentiment = (6,000 - 2,000) / 10,000 x 100 = 4,000 / 10,000 x 100 = 40%. The following week, mentions rise to 12,000 with 5,400 positive and 4,200 negative, which gives (5,400 - 4,200) / 12,000 x 100 = 10%. The fall from 40% to 10% shows a marked worsening of mood.Case study
Seen in the real world.
Clearwater Securities is an illustrative, fictional brokerage that tested a social sentiment indicator on a basket of 50 shares. Its research team recorded daily scores for a year and compared them with price changes over the next week.
They found a small positive relationship for large, widely discussed companies, but almost none for small companies with few mentions. Several of the strongest signals came from coordinated promotion by a handful of accounts rather than genuine opinion.
The team decided to use the indicator only for large companies and only alongside financial analysis. The illustrative lesson is that a sentiment score can add information, but its quality depends on the volume and authenticity of the underlying posts. The brokerage also began filtering out accounts that posted identical messages, which improved the quality of the data.
Watch out
Common mistakes.
- Trusting a score built on a small number of posts, when a handful of accounts can dominate the result.
- Ignoring bots and coordinated campaigns that distort the mood being measured.
- Treating sentiment as a reliable forecast of prices, when the relationship is unstable and often weak.
Questions
People also ask.
How is sentiment scored?
Software classifies each message as positive, negative or neutral, then the counts are combined into a score such as net sentiment.
Can sarcasm fool the indicator?
Yes, sarcasm and slang are hard for automated tools to read, which is why scores can be misleading.
Should I trade on it?
It is better used as one input among several, alongside financial analysis and risk controls.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%Related