Text Analytics
Deriving meaningful insights from unstructured text feedback
Text Analytics is the process of deriving meaningful insights from unstructured text feedback. This feedback is an essential resource for discovering and prioritizing potential improvements to the customer experience. When customers comment about their frustration with a long line, disappointment with the product selection, or appreciation for an associate who helped them find exactly what they needed, their feedback provides important clues about the interactions shaping their experience.
The expertise and time traditionally required to analyze large quantities of customer comments has often been a prohibitive barrier to uncovering these insights. Medallia Text Analytics solves this problem by:
-
Categorizing text from multiple data sources into themes and topics. Supported data sources include (but are not limited to) survey responses, call center notes, chat logs, email, and social media reviews.
-
Supporting topics in multiple languages, and providing machine translation support through Amazon Translate, Google Translate, or SYSTRAN.
-
Associating an impact score with each topic to show the influence of the topic on overall customer loyalty.
Note: For records with satisfaction scores only. -
Applying sentiment models to detect whether each sentence in a comment is strongly positive, positive, mixed opinion, negative, strongly negative, or has no opinion.
Watch the following video for an overview of Text Analytics. For more videos about using Text Analytics, see Text Analytics videos.
Important terms
Make sure that you understand the following terms as you work with Text Analytics:
-
Structured feedback — Feedback aligned with pre-configured answer possibilities. For example, you might ask customers for satisfaction or likelihood-to-recommend scores, or to select from a list of stores they visited.
-
Unstructured feedback — Free-form feedback, usually provided as text feedback in surveys. For example, you might ask customers to provide the reasons for the scores they gave. If you are using Medallia Speech, customers also provide unstructured feedback in voice calls. Text Analytics processes unstructured feedback to reveal insights about customer interactions with your company.
-
Topic — A category of phrases in unstructured feedback. For example, a customer might leave the following unstructured feedback in a survey text field:
"I usually like your brand, so I left a high score, but the person I talked to today was so rude!"
In this example, the first part the feedback might be associated with the Brand - Satisfaction topic, while the second part might be associated with the Staff Attitude topic. Topics are available in reports, so you can find the feedback related to each topic.
-
Sentiment — The degree of positivity/negativity customers feel about phrases in topics. In the example comment above, Text Analytics would identify positive sentiment in the first phrase, and negative sentiment in the second phrase. Sentiment is available in reports.
-
Theme — A system-generated group of conceptually-related words and phrases in unstructured feedback. Themes are similar to topics, but while topics are built by Medallia Experience Cloud administrators, themes are built automatically by Text Analytics. For example, if you build a topic for Laptop Screen, Text Analytics might build themes for LED, Size, Picture, Crack, Spill, and so on. Themes are available in reports.
Text Analytics in Medallia Admin Suite
In Medallia Admin Suite, you have precise control over the subjects for which Medallia Text Analytics captures insights. You define the topics for which you want insights, and create the rules that determine which phrases are captured for each topic. For more information, see Topics screen.
To use Text Analytics, roles need the Manage Topics permission under the Text Analytics section listed in Administrative permissions.
For more information, see:
To access Text Analytics, select Text Analytics under the Analysis and Prediction section:
