Guide · Sentiment analysis · 8 min read

What is sentiment analysis and how to do it by yourself

Sentiment analysis tells you how your target audience feels about your brand, product or service — at the scale of the whole internet, in minutes.

Analyst reviewing sentiment dashboards on dual monitors

Sentiment analysis, defined

Sentiment analysis is the automatic classification of text by emotional direction: positive, negative or neutral. Modern systems add emotion (joy, anger, surprise), intensity and aspect — which product, feature or location the feeling attaches to.

How the AI actually works

Today's models are trained on millions of labelled social posts, reviews and articles. They weigh word choice, negation (“not bad”), emoji, punctuation and context windows — which is why a good model catches sarcasm like “oh great, another update 🙄” instead of filing it as praise.

Document level: the overall feeling of a post or article.
Aspect level: feelings about specific features (“battery great, support awful”).
Comparative: your sentiment plotted against named competitors.

Doing it yourself: a 5-step playbook

1. Collect widely. Pull mentions from social, reviews, forums and news — sentiment on one channel lies. 2. Clean the data. Remove duplicates, bots and irrelevant homonyms. 3. Classify. Run an AI classifier rather than keyword lists; “sick” means opposite things in gaming and healthcare. 4. Slice by aspect. Break results by product, campaign and region. 5. Track the trend. A single score is trivia; the slope is strategy.

Metrics that matter

Watch net sentiment (positives minus negatives as a share of total), sentiment per thousand mentions, and reputation score over time. Spikes deserve investigation; slow drifts deserve strategy.

Where Goodstanding fits

Goodstanding runs this entire playbook automatically: collection across 25M+ sources, sarcasm-aware classification in 108 languages, aspect breakdowns and daily reputation scoring — with every number linked to the mentions behind it.

Try sentiment analysis on your brand.

Backfill a year of history and see your sentiment trend in minutes.