The Sentiment Analysis Dashboard | Yext Hitchhikers Platform

What You’ll Learn

In this section, you will learn:

• The main components of the Sentiment Analysis dashboard
• The insights associated with each keyword

The Components of the Sentiment Analysis Dashboard

Sentiment Analysis is found under Reviews > Sentiment Analysis in the navigation bar.

The Sentiment Analysis dashboard is built to provide the tools and insights to drill down into specific keywords or categories and see the associated customer sentiment and reviews. The two main components are Sentiment by Collection and Sentiment by Keyword.

Sentiment by Keyword - Sentiment by Keyword, found below Sentiment by Collection, is the best way to view and analyze individual keywords. Keywords (primarily nouns) are the words that are identified as subjects being described in the review content.

Keywords can be sorted alphabetically or based upon mention count, sentiment score, and average rating.

Note that keywords are automatically added based upon review analysis. You cannot add or customize the keywords, you can only view keywords that have already been identified in your existing reviews.

Sentiment by Collection - Collections are at the top of the screen and are essentially folders to group related keywords. Collections are created by the user, and we will describe how to do that in more detail in the next unit.

When you first visit this page, there won’t be any collections yet in new accounts. We will show you how you can create one later.

Filtering

You can filter the Sentiment Analysis tab using the Filter toolbar at the top, which includes the functionality to filter by location, time, mention counts, keywords, and sites. This can be useful to get valuable insights such as understanding if one region has better “service” than another. Filtering by timeframes adds deltas to each metric so you can see how they have changed over time.

Keyword & Collection Insights

Whether you are viewing your reviews in a Collection, or individually by keyword, there are a number of different insights that you are able to explore.

Sentiment Score - Sentiment score is the rating of the positivity or negativity associated with a particular keyword. The scale is scored on a scale between -100 and +100.

The majority of keywords have a sentiment score between -10 and +10. That means that any keywords below -10 or above +10 are extremely negative or positive, respectively.

Note

Since Yext relies on natural language processing technology for sentiment analysis, there will be scenarios where the sentiment score may not be accurate: for example,if the review is sarcastic, or if it contains ambiguous keywords (such as “cheap,” which could be either positive or negative in context). The algorithm is constantly learning and improving over time based upon the reviews that come in.

Mentions - The number of times a keyword appears in your reviews.

The most frequently appearing keywords are often the most interesting keywords to consider. These are the items that your customers are most frequently commenting on and therefore may inform larger business decisions.

Average Rating - The average rating of all reviews that contain a given keyword.

While there is a correlation between sentiment and average rating, sentiment is often the better measurement of a consumer’s reaction to a keyword. Consider a 3 star review: “the food was good, but the service was bad.” The star rating on its own will not help a business understand the quality of their food or service but sentiment analysis will parse out the sentiment of these attributes.

Top Modifiers - Words most frequently used to describe keywords, giving them context and their sentiment score. Modifiers are typically adjectives and adverbs. Examples include words like “friendly,” “helpful,” “nice,” “rude,” and “amazing.”

Clicking on a modifier will preview recent reviews that include the keyword and modifier pairing. The highlighted keyword/modifier link at the top left of the preview will take you to a pre-filtered list of all mentions of that keyword and modifier.