In Scup operations, it is possible to tag and make sentiment analysis references, videos and posts obtained in your monitoring, helping data gathering be analyzed not only quantitatively but also qualitatively.
It is important that those who perform monitorings understand the content of the references obtained, that is, what the tagging operation allows you to do, once after doing that, it is possible to analyze what people have been most commenting about the brand or product monitored.
What is tagging?
It is a kind of “label” assigned to a reference.
What is Sentiment Analysis?
Sentiment analysis is to specify if a reference is positive, negative or neutral.
Why Should We make sentiment analysis and Tag?
When you make sentiment analysis and tag the references of your monitorings, you start working with the information obtained qualitatively, which is highly important to interpret what people have been saying about brands and products on the networks. It allows you to understand public opinion about products, services and your company’s brand. The quality of information may contain a great knowledge for everyone involved, including for crisis prevention.
Registering Tags
Registering tags is very simple on tabs Setting / Tags and Add Tags. It is possible to create several tags at once by clicking (+) on Add Tags page.
Tips for your Tagging and Sentiment Analysis operations:
• Before starting your classification operation, have in mind your monitoring strategic purpose,
• Create a document with monitoring information for your team reference,
• Establish the tags and the description of which kind of topics can be tagged with that specific tag. e.g.: Tag Trash (everything not related to your brand),
• Use keywords to establish the topics that can be classified by each tag. e.g. Tag Delivery: Deadline, before deadline, late, delay,
• For Sentiment Analysis, many issues can be raised about references, such as irony, so discuss with your team about the interpretation possibilities for that reference and always keep your plan updated,
• Keep your team oriented about brands and companies actions monitored, therefore, the establishment and orientation about which TAG is to be used will be easier,
• For punctual campaigns, use a specific TAG for each event in order to have a more precise analysis. e.g.: Tag fair 1, Tag fair 2.
• Establish rules to make teamwork easier, which can be useful for big campaigns with a great volume of automatic rules, both for sentiment analysis and tagging.
• When a user, for example, mentions two products in a single reference and there are in your plane specific tags for each one of them, establish a priority between the products: which of this information needs to be more observed? Generally, less commented products should be top priority, once they require more attention because they are less mentioned.
• Dividing tags into groups also helps data analysis, for example: Moment, who, influencers, feedbacks, etc. e.g.: Purchase Moment, Use Moment, Cancellation Moment.
• If your team performs services via Scup, establish a work flow, once not all references are subject to services, so establish subject and non-subject tags.
• When references are divided into the team, it is possible to establish tags for each employee and divide among them, thus establishing rules.
• For classification, establish what shall be considered for each sentiment analysis, for example: Shall actions feedbacks with the expression “thank you” be considered positive?
Why did Scup choose a non-automated work?
Our team works with the purpose of making daily actions simpler and less complex for social network professionals, so why didn’t we automate sentiment analysis and tagging?
Because we understand that the references interpretation and especially the feelings are better analyzed by people, an automated job is hardly capable of identifying really relevant tags and classification criteria for a specific kind of business. When a person interprets it, it is possible to view the users’ history, and that helps you better understand the message the user wants to convey. With this kind of analysis, we can say that the information can be interpreted with a more strict criteria and relevance.
As we believe in social media as a way of relationships, it is sensible that people read, interact and classify in a context, so that true relevant information can be obtained by the company. A machine cannot interpret, for Globo, for example, that the establishment and classification of a reference with a tag “Soap Opera” is relevant information.
Examples of Tweets:
In this situation, the user informs that the technician showed up and solved the problem, but we cannot be sure if this Tweet, which causes confusion when being interpreted, is positive because the technician showed up, or negative by the lack of confidence in the job performed?
In this example, the user is ironical when writing “Surprisingly, the connection is hardly ever down!”, so did she expect to have problems and not that she is satisfied because the Internet is not down?
In situations like these, when people’s interpretation makes difference, it is possible to understand and discuss with the team what is the meaning of all that, and be prepared to further situations, something that an automated interpretation would not be able to do with so many details.
scup.com.br