From 2011 We Trust in Value
Text & Sentiment Analysis
All companies have historically generated and received and keep generating and receiving large amounts of text, mostly in unstructured documents.
Storing these documents by indexing and tagging them as required by the traditional Document Management and Search Engines Platforms is a very cumbersome and prone to errors and inaccuracies work.
Text Analysis solutions can automate this effort through semantic search; for example, they can extract all and only the information relevant to a given search, information that in turn can be used to create reports of knowledge; unforeseen relationships among documents can be identified as much as “moods and inclinations”.
There are two main areas of intervention.
At this stage, natural language analysis tools sift and automatically categorize large volumes of multilingual text, even through the creation of metadata and the development of taxonomies.
In this way an analyst can identify behaviors or correlations not evident at first sight.
An example of content categorization is in the analysis of the trouble tickets opened by a Help Desk operator, often working for a third-party company, or directly by the customer in self-service mode: in such cases, the formal categorization offered by the Trouble Ticketing platform is often not complete or correct.
Another example is the analysis of electronic health records: they contain unstructured information that can be used to enrich the data made available by the traditional health information streams. Such information are not only useful to doctors, but, in an aggregate and anonymous form, to pharmaceutical companies, health departments, insurances.
Recent analyses show that, when deciding on a purchase, potential customers increasingly rely less and less on advertising or manufacturer’s advice and more and more on friends’ suggestions; but friends stay increasingly in touch through social networking and whole discussion forums are dedicated to products and services: a way to exchange experiences in order to avoid improper purchases.
A way that is already having a big success in consumer electronics and travel especially.
Knowing what is written on the Internet therefore becomes absolutely essential for a company to know its own reputation and evaluate how this is affected by the launch of a product or by a new advertising campaign or, on the other way round, to understand which services and solutions their clients would like to have, gaining this way an important competitive advantage.
This is what Sentiment Analysis solutions do: automatically and periodically navigate digital content such as websites, blogs, social networks but also any internal sources, such as press releases, and then perform statistical and language analysis to extract the information of interest, build indexes and create reports and graphs that summarize the thought of consumers on one’s brand and products – or products and the brands of competitors.