The information contained in the customer reviews on particular products or services is of great interest both for companies and for consumers themselves. Companies and organizations invest always a huge amount of money to acquire the opinions of guests to direct their marketing mix and affect consumer satisfaction. And ‘it demonstrated that each individual is interested in the experiences of others when buying a product or adheres to a service other than the fact that the opinions expressed online now greatly affect business reputations. For this reason, information on customer opinions are likely to become a key component of Customer Relationship Management (CRM) solutions.

Consumer reviews provide valuable information for companies:

  • First, they help to understand how their products and services are perceived by producing evidence about the satisfaction and expectations useful to determine the needs and current and future customer preferences;
  • secondly, can be helpful in understanding what are the dimensions or the relevant attributes of the product for a specific consumer segment, or to discover new segments based on consumer groups that have common characteristics;
  • Third, the views of customers about the products offered by competitors provide essential information to build a real competitive analysis.

Collecting the views of consumers has always been a difficult and demanding task, usually done by specialized consulting firms that conduct surveys on representative customer groups. The results of these investigations are limited to a small sample of customers and must be extrapolated for the entire population. In addition, investigations are only valid for a short period of time since today from customers change quickly, often as a result of independent events by companies.

With the growth of the Web and the disclosure of review sites, forums, blogs and discussion groups, consumers’ opinions are not limited to word of mouth but more or less explicitly are “dense” in the Web, as the existing applications allow easy sharing of experiences and opinions, positive or negative, now about any product or service. In this case, in the context of reviews, the cost of collecting information for a company would be markedly lower.

The huge amount of online reviews that are generated each day makes all kinds of manual processing virtually impossible. The challenge is therefore to develop new technologies to process this information efficiently and produce essential knowledge to make better marketing strategies. The information contained in online reviews and forums is usually presented as structured and free text (user-generated content), the automatic extraction of the views of users requires the use of advanced Natural Language Processing (NLP). In particular, we see the emergence of a discipline within the NLP called Sentiment Analysis (or Opinion Mining) on ​​the computational processing of opinions, feelings and subjectivity in text. Where ends the work of the Natural Language Processing and Sentiment Analysis, begins what the business intelligence in order to “treat” the huge amount of information is to synthesize and generate knowledge to provide the business user tools (such as market analyst) to support strategic decisions.

The figure shows what can be a typical decision-making and inferential of a marketing strategy based on the web content. The process is typically cyclical in that every marketing action causes a reaction of customers. The correct interpretation of the reaction of customers is the key to sharpen the proposition on the product / service market.



Web Content : is the information base to be treated, comprised of sites, blogs and specialized forums on the product / family of products / services of interest.

Gathering Date : collects data through the download of documents, analyzes them to extract morpho-syntactic and semantic information.

Information Extration : allows to identify phrases mentioning one or more reference products, find phrases related to the environment even if not explicitly mentioned.

Sentiment Analysis : determines opinions and realized as a sequence of steps for classification.

Warehouse Enrichment Date : acquires data of structured Sentiment Analysis and integrates them within the complete corporate repository of information about users, orders and the characteristics of products.

Data Mining : using specific algorithms allows you to bring out the hidden information, allows to cluster the data you “look alike.”

Business Intelligence : enables the provision of an environment for advanced analysis on the data through interactive features and graphical presentation of results.

Market Strategy : is the market strategy applied on the basis of dell’Opinion Mining results, the company’s choices and of course the experience of the market analyst.