Sentiment is a proprietary KPI that provides you with an overview of the overall satisfaction of your customers.
After an experience, people tend to leave feedback through multiple mediums and channels: they write reviews, share posts and/or digital content on social media... and much more. All this information is the key to understand how much they enjoyed interacting with you and your business. This is where the Sentiment KPI comes into play: a proprietary score (with a valid range from 1 to 100) that measures the overall level of customer satisfaction, accurately estimating the appeal of your business.
Behind this score, there is an evaluation and data processing pipeline that feeds a proprietary semantic algorithm that takes into consideration multiple attributes of the online content (feedbacks, topics, opinions, languages...). Each of these attributes is weighted, with the aim of estimating a final score that is more precise and inclusive of many aspects compared to a pure base rating. Think of one of the experiences you've had: would you find a review enriched by a detailed comment more useful than one with a simple rating? Taking these facets into account is essential to understand how you are performing and how you can further improve.
To help you better understand the Sentiment KPI, we have created a visual scheme with a brief glossary of the main terms used.
SENTIMENT_SCHEMA.PNG
GLOSSARY
- Network: web platform that users can feed with contents of judgement
Example: booking.com - Cluster: super-set of topics which have similar argument of pertinence
Example: Food, Hotel Staff - Topic: subject on which users express opinions
Example: dinner, breakfast, service - Opinion: judgement expressed by a user on a topic
Example: good, limited, slow - Fragment: piece of content extracted according to criteria of meaningfulness, relevance, synthesis and self-consistency
Example: Breakfast was good, but the service a bit slow
You can find the Sentiment KPI in the following products and services:
- Data Appeal Studio: the sentiment value will refer to the destination you are observing, with the possibility of making comparisons that will help you identify an analysis trend
- Travel Appel: the sentiment value will refer to the single POI you are looking at, to help you make decisions
- POI API: sentiment value referring to a single POI, obtainable in near real-time, which can be requested as an aggregate or timeseries and with the possibility of filtering through multiple parameters
- Location API: sentiment value referred to one or more shapes, containing a certain number of POIs, which can be requested as an aggregate or timeseries and with the possibility of filtering through multiple parameters
- POI Data-Pack: sentiment value referring to a single POI, which can be: calculated in the last 12 months if a basic package consisting of Core & Rich Attributes is configured; calculated as a trend over a custom time window if the Sentiment Trend add-on package is requested.