Text and Sentiment Analysis

It is critical that today’s enterprise be able to make full use of all the customer information it has in order to drive the best strategic and operational decisions. And that includes text data such as responses to open-ended survey questions, emails, contracts, and call center notes. Some experts believe that this type of content actually represents more than 80 percent of an organization’s data and it continues to grow at exponential rates. And yet, much of that information is not used at all. Or some of it is accessed by humans reading and painstakingly coding it, which certainly is not practical when the number of text records can number in the thousands, if not millions.

True ‘Voice’ of the Customer

Without the ability to capture, analyze, and store this content, the voice of the customer is not truly being heard. For many companies that voice remains “hidden” in text. Text content gives us the details and tells us the ‘why’ behind quantitative data. Customer scores on surveys, codes for problems that are called in to Customer Support, and the number of visits to a web site are certainly useful. But these data points do not convey why the rating was given, what the nature of the problem was, and why the web site visitor did not complete the purchase online. Linking the unstructured data (text) to the structured (quantitative) data puts all of the information at your fingertips, so that you can quickly make the most informed business decisions.

Through the use of CustomerImpact’s award-winning text analysis software, companies can access, integrate, and analyze the full depth and breadth of their customer feedback so as to produce new insights and achieve maximum value. It enables companies to:

Beyond Key Word Search

CustomerImpact’s solution goes beyond the traditional Search method, which is set up to look for key words and phrases using Boolean operators (and/or/not). The drawbacks to the Search approach are that:

• A search query does not always return information
• Different customers may use different terminology for the same issues and therefore the search may not cover all relevant items
• The search may return irrelevant documents because keywords may have different meanings in different contexts
• Key themes are categorized, but the sentiment (positive or negative feeling) is not provided
• It is time-consuming to actually extract meaning—documents are returned, but the analyst must still take time to read them.

CustomerImpact’s Solution

The CustomerImpact solution overcomes these limitations. Its accuracy is unrivaled. It removes false positives, while highlighting all relevant information. Customer sentiment is revealed. Finally, once the system is set up and prepared, the user can quickly mine all the data, easily extracting key takeaways. Features include:

• Highly scalable
• Integration with structured and unstructured data sources
• Identification of trends and statistical exceptions
• Extends the analytical capabilities of unstructured data (basic analytics can be applied to this data)
• Streamlines the creation of categories, resulting in a more efficient process
• Allows for saving and re-using categories, thereby saving time in analyzing future related surveys
• Offers 30-language capability for global businesses
• Extracts and analyzes the most relevant sentences to underscore the most important messages from the text

CustomerImpact’s highly trained and experienced consultants will work with clients to quickly deliver a full implementation. It can exist as a standalone solution, helping clients derive the most value from their customer feedback. Or it can be incorporated as one of the building blocks of our Voice of the Customer program. Our goal is to transform your text-based customer feedback into customer intelligence so that you can better serve customers, driving customer retention, revenue growth, and profitability.