Customer Reviews Analysis Through Pulse

Chatmeter’s Pulse sentiment analysis and text analytics tool gives you the ability to delve deeply into the content of customer reviews. This analysis offers ability to:

• Understand your customers and how to provide value to them
• Identify opportunities for operational improvement
• Compare customer perceptions of locations and groups
• Track customer sentiment over time
• Monitor customer feedback on new products or initiatives

Pulse works by using machine learning and artificial intelligence (Sentiment Analysis and Natural Language Processing algorithms) to analyze the text of customer reviews. By using Pulse to analyze this text, you gain advantage from monitoring how people are feeling about your business, what subjects they’re talking about, and how they’re describing those subjects.

In this article, learn how to use the Pulse Summary and Explorer tabs. Chatmeter also reports Pulse data for location and group comparison. You’ll find that information in Insights (Dashboard) > Location Comparison on the Pulse tab. For information about Pulse location comparison, see the help article Pulse Comparison Reports.

Find the Pulse page at Insights (Dashboard) > Pulse.

The Summary tab shows how sentiment (defined as positive, neutral, or negative) is changing over time. You can use it to find points where there was significant change in sentiment or review volume. Detailed usage instructions follow in Using the Summary tab.

Use the Explorer tab to delve into the categories and topics that your customers are talking about. Detailed usage instructions follow in Using the Explorer tab.

Pulse terminology

Pulse uses specific terminology to describe factors involved in the data and analysis, including:

• Sentiment—The feeling of a statement determined by AI and Machine Learning models. Sentiment values are Positive, Neutral, or Negative.
• Category—The general area of the business that a topic falls into
• Topic—The main subject of a statement
• Attribute—The words used to describe the topic of a statement
• Mentions—An instance of a topic used in a statement
• Negative point of interest—A point in time when the average sentiment falls below the standard deviation, pointing to a rise in negative customer sentiment.
• Positive point of interest—A point in time when the average sentiment rises higher than the standard deviation, pointing to a rise in positive customer sentiment
• Volume point of interest—A point in time when the volume of reviews is higher than the standard deviation, pointing to an influx of customer feedback.

Using Pulse page filters

To structure a Pulse report, use the Global Filter to select the locations, accounts, or groups you’d like to include.

Use the Page Filters to select the following factors to structure your report:

• Saved Filters—You can select from filters you’ve created for re-use. For instructions for creating a saved filter, see Using Saved Filter Sets.
• Providers—Listing providers for which you’d like to view your locations’ or groups’ pulse metrics
• Categories—The general area(s) the business of the location or group
• Topics—The main subjects of the statements you want analyzed
• Sentiment—The feeling of a statement, determined by artificial intelligence and machine learning models. Sentiment can be positive, neutral, or negative.
• Duration—By default Last 30 days, but can another selected range or a customized range

Using the Summary tab

The Summary tab has two graphs that show how sentiment (defined as positive, neutral, or negative) is changing over time. Use the filters, as described in the previous section, to specify the data factors you want in your report.

You can also select the level of granularity for your data selections for your Summary tab report by selecting the Granularity selector on the Average Sentiment Over Time graph or the Sentiment Distribution Over Time graph. Choices are 1 to 5 days, one week, 30 days, 90 days, or one year.

The Average Sentiment Over Time graph has a line showing average sentiment score over time (the blue line). The dots on the line indicate sampling dates. Hover over the dot to get the specific score and the number of reviews represented in the score. The orange line shows the volume of reviews.

The three types of points of interest are indicated by green (positive point of interest), red (negative point of interest), or gray (volume point of interest). Click on a point of interest to get more information about what’s driving the change. It will report the percentage change from the previous period and the specific topics leading to the change. Select View more about this time period for more detail as displayed on the Explorer tab’s sunburst graphic.

The Sentiment Distribution Over Time graph shows how many statements are negative, neutral, or positive over the selected time.

Using the Explorer tab

The Explorer tab contains graphs and tables to reveal information about the categories and topics your customers are talking about in their reviews. Toggle to the Explorer tab on the Pulse page at Insights (Dashboard) > Pulse.

The top pane contains:

• Sunburst—An interactive chart that shows categories, topics, attributes, and their average sentiment
  
  

  Categories are the innermost ring. Select a category to see details of the topics and attributes associated with the category. We selected the category shopping to display these details:
  
  Hover over the graph to expose details about specific areas of interest.
  

  You can also click on an area of the sunburst to zoom in to a specific category or topic. We clicked on “service” topic to display this detail:
  
  Note that the Data Metrics and Distribution charts now show data for service, too.

• Data Metrics: A table view of category, topic, and attribute data. You can also find the average sentiment and number of mentions of each data point, as well as the percent change in both metrics from the previous period.
  
• Distribution: Toggle from Data Metrics to Distribution by selecting the Distribution tab to see a visual breakout of the numbers of mentions of each category, topic, or attribute that are positive (green), neutral (yellow), and negative (red). Select a bar to display the details on the data.

Using the Mentions table

As you interact with the sunburst, Data Metrics, or Distribution tables, the Mentions table in the pane below those graphics updates based on what topic, category, or attribute you are looking at. In the Mentions table, you’ll find the individual statements, their sentiment, and the reviews they come from. You can change the sentiment of a statement if it is inaccurately assessed by the software. Use the keyword search to see statements containing a particular word.