- DarkLight
Smart Sentiment Analysis
- DarkLight
Save valuable time when moderating reviews with Yotpo's Smart Sentiment Analysis. Yotpo's smart sentiment algorithm analyzes your reviews and automatically identifies positive and negative reviewer sentiment. We then help you leverage this data throughout the Yotpo marketing platform to boost your click-through rate, broaden your social media footprint, and create dynamic Facebook Ads to showcase your best reviews.
- Yotpo accounts that are set to English
- Reviews that are written in the English language
How it works
Yotpo analyzed one million reviews in order to identify the most common words used in one, two, and three-star reviews and then analyzed the most common words found in four and five-star reviews. The words that appeared in reviews most frequently were then classified by polarity where the most common words in one, two, and three-star reviews carry a negative sentiment and the most common words appearing in four and five-star reviews carry a positive sentiment.
Each review is then analyzed by our smart algorithm according to this baseline and words are classified by positive or negative sentiment according to their known polarity.
The result is an output, from 1-100%, which is indicative of the likelihood that the review contains positive or negative customer sentiment.
Smart Sentiment Analysis in Yotpo
Yotpo's Smart Sentiment Analysis is automatically incorporated into core Yotpo Moderation and Social features. The sections below provide an overview of how Yotpo leverages Sentiment Analysis in each respective feature.
Auto publish
Enable Active Auto Publish to allow reviews to automatically appear on-site. Auto Publish is based on the review's star rating as defined by choosing the minimum star rating in the adjacent drop-down menu. If the minimum score is set to 3 stars, all reviews with 3, 4, or 5 stars will automatically be published, while those reviews with 1 or 2 stars will be left unpublished.
Auto publish by sentiment
Checking the Activate Sentiment Filter box will exclude reviews with negative sentiment from auto-publishing to your site regardless of their star rating. Click to read more about configuring Auto Publish in your Moderation settings.
Currently, Yotpo's Smart Sentiment Analysis is only compatible with English reviews and is based on statistic models, which are not 100% accurate.
Reviews will be blocked and flagged by the sentiment filter in the following cases:
- The review does not contain a single English word
- e.g. A review submitted entirely in a non-English language
- The review contains entirely random/gibberish characters
- e.g. A review that contains only random or gibberish characters such as a code, script, etc.
- Reviews can contain random characters such as a serial number, gift code, etc. but must contain other identifiable English words
- Emojis such as smileys ":)" are supported and will not be blocked by the sentiment filter if they are accompanied by English words
- The review does not contain any content whatsoever
- e.g. A completely blank/empty review
Manual publish
If your orders do not pass the positive review sentiment, you can always publish them manually.
You may publish reviews individually or in bulk.
To publish an individual review:
- From your Yotpo Reviews main menu, go to Reviews > Moderation.
- Next to the review you want to publish, click Publish.
To publish multiple reviews at once:
- Mark the check box on each review you'd like to publish
- Select Publish Reviews or Publish Reviews & Pictures (if relevant) from the Action drop-down menu
Social Push
Yotpo's Social Push feature helps you grow your social media footprint by automatically pushing reviews with the best sentiment to your social media pages. Our smart algorithms will choose the best time of day and the best intervals to post top reviews to your Facebook account without running the risk of spamming your followers. Reviews that are pushed to your social media pages are chosen based on star rating, sentiment, the popularity of the product, and the quality of the review.