The Insights Report, powered by IBM Watson, uses artificial intelligence to automatically discover keywords in review content and bring unknown information to light.

What is Insights Report?

Please allow up to 24-72 hours for the report to populate data for your location(s). Please also note that this feature is currently in Beta mode. If you run into any issues, please contact [email protected].

The report presents data in 3 different views and can also monitor Tags:

The Impact chart is a visual display showing discovered keywords and how they impact your ratings and reputation. Understand the rating, quantity, mentions, and sentiment of each discovered keyword.

Keywords are plotted on the chart by:

Hover over a keyword for the number of mentions and average review rating

When review content is analyzed and processed using NLP it uncovers keywords from your review content. The Insights Report then plots the keywords on a chart, and is broken down to 5 data points:

The size of the circle represents the total number of mentions for a discovered keyword. Hover over each circle to view the total mentions and average review rating for that keyword.

Sentiment helps you to understand your customer’s feelings about discovered keywords by assigning a Sentiment Rating. This can enable you to extract new insights about your product, service, or brand to better understand customer experience.

The Sentiment Rating is generated using Natural Language Processing by IBM Watson:

Greater than 2.5 – 5.0 = Positive Sentiment (Green)

Less than or equal to 2.5 = Negative Sentiment (Red)

The goal of assigning sentiment to a keyword is to provide an understanding of a customer’s attitude about a specific keyword separate from the review rating.

In the example above the customer gives a 1-star rating. However, their experience with “cold press coffee” is positive. The content of this review is processed to assign positive sentiment to “cold press coffee” regardless of the negative review rating.

‘Long line’ is also discovered and assigned negative sentiment.

Understanding the sentiment of discovered keywords from review content outside of the overall rating provides a deeper understanding of your customer’s experience. Use this understanding to prioritize action more precisely.

Trends identifies the percent change in the number of times a discovered keyword is mentioned between two periods. Change the time periods to understand trends, compare date ranges to uncover keywords with the most mentions, track increases or decreases in keyword mentions, and identify the top location for a discovered keyword.

Choose to compare the last 1, 7, 30, or 90 days to the matching previous time period to uncover keywords with the most mentions and track any increase or decrease in the number of mentions of a keyword within the selected period. You can also identify the top location for a discovered keyword.

If you want to understand how specific aspects of your business impact reviews and ratings using the Insights Report, you can do so using Tags (or Auto-Tagging).

To view Impact, Sentiment, or Trends for Tags, selecting Tags in the upper right corner

Each section of the Insights Reports can be used to monitor Tags by changing the selection to display Tags.

Natural language processing, sentiment analysis, and machine learning make it possible to discover the impact, sentiment, and trends of all of your review content and bring previously unknown information to light. With IBM Watson powering your Insights Report, you’ll gain a deeper understanding of what’s delivering a 5-star experience and what’s not, empowering your business to take action.

IBM Watson uses artificial intelligence (AI) disciplines to generate the content of the Insights Report including; Natural Language Processing, Sentiment Analysis, and Machine Learning. As a result, review content can be understood at scale, and with velocity, to allow you to respond quickly to shifts and signals from your customers.

Natural Language Processing (NLP) is the use of technology (AI, computational linguistics and computer science) to allow a machine to understand a human’s natural language.

NLP breaks down language into shorter, elemental pieces with the ultimate objective to read, decipher, understand, and make sense of language in a manner that is valuable.

Sentiment Analysis is a field within NLP that works to identify and extract opinions from text. The unstructured content of your reviews is automatically transformed into structured data, so that the expressed opinions about an identified keyword or phrase can be surfaced as a Sentiment Rating.

Sentiment ratings assigned to keywords discovered in your review content help you understand the nuances of your customer’s experience over time and identify what’s behind shifts that happen.

Machine Learning builds algorithms that allow computers to learn to perform tasks from data instead of being explicitly programmed. It then “tests” and “corrects” the computer to further the accuracy of the computer’s “understanding”.