What Can Businesses Learn from Text Mining

Case StudyWhat Can Businesses Learn From Text Mining? Text mining is the discovery of patterns and relationships from large sets of unstructured data – the kind of data we generate in e-mails, phone conversations, blog postings, online customer surveys, and tweets. The mobile digital platform has amplified the explosion in digital information, with hundreds of millions of people calling, texting, searching, “apping” (using applications), buying goods and writing billions of e-mails on the go.
Consumers today are more than just consumers: they have more ways to collaborate, share information, and influence the opinions of their friends and peers; and the data that they create in doing so have significant value to businesses. Unlike structured data, which are generated from events such as completing a purchase transaction, unstructured data have no distinct form. Nevertheless, managers believe such data may offer unique insights into customer behaviour and attitudes that were much more difficult to determine years ago.
For example, in 2007 JetBlue (the American Airline) experienced unprecedented levels of customer discontent in the wake of a February ice storm that resulted in widespread flight cancellations and planes stranded on Kennedy Airport runways. The airline received 15,000 emails per day from customers during the storm and immediately afterwards, up from its usual daily volume of 400. The volume was so much larger than usual that JetBlue had no simple way to read everything that its customers were saying.

Fortunately, the company had recently contracted with Attensity, a leading vendor of text analytics software, and was able to use the software to analyze all of the e-mail it had received within two days. According to JetBlue research analyst Bryan Jeppsen, Attensity Analyze for Voice of the Customer (VoC) enabled JetBlue to rapidly extract customer sentiments, preferences, and requests it couldn’t find any other way.
This tool uses a proprietary technology to automatically identify facts, opinions, requests, trends, and trouble spots from the unstructured text of survey responses, survey notes, e-mail messages, Web forums, blog entries, news articles, and other customer communications. The technology is able to accurately and automatically identify and many different “voices” customers use to express their feedback (such as a negative voice, positive voice, or conditional voice) which helps organisations pinpoint key events and relationships, such as intent to buy, intent to leave, or customer “wish” vents. It can reveal specific product and service issues, reactions to marketing and public relations efforts, and even buying signals. Attensity’s software integrated with JetBlue’s other customer analysis tools, such as Satmetrix’s Net Promoter metrics, which classifies customers into groups that are generating positive, negative, or no feedback about the company. Using Attensity’s text analytics in tandem with these tools, JetBlue developed a customer bill of rights that addressed the major issues customers had with the company.
Hotel chains like Gaylord Hotels and Choice Hotels are using text mining software to glean insights from thousands of customer satisfaction surveys provided by their guests. Gaylord Hotels is using Clarabridge’s text analytics solution delivered via the Internet as a hosted software service to gather and analyze customer feedback from surveys, e-mail, chat messaging, staffed call centres, and online forums associated with guests’ and meeting planners’ experiences at the company’s convention resorts.
The Clarabridge software sorts through the hotel chain’s customer surveys and gathers positive and negative comments, organizing them into a variety of categories to reveal less obvious insights. For example, guests complained about many things more frequently than noisy rooms, but complaints about noisy rooms were most frequently correlated with surveys indicating an unwillingness to return to the hotel for another stay. Analyzing customer surveys used to take weeks, but now takes only days, thanks to the Clarabridge software.
Location managers and corporate executives have also used findings from text mining to influence decisions on building improvements. Wendy’s International adopted Clarabridge software to analyze nearly 500,000 messages it collects each year from its Web-based feedback forum, call centre notes, e-mail messages, receipt-based surveys, and social media. The chain’s customer satisfaction team had previously used spreadsheets and keyword searches to review customer comments; a very slow manual approach.
Wendy’s management was looking for a better tool to speed analysis, detect emerging issues, and pinpoint troubled areas of the business at the store, regional or corporate level. The Clarabridge technology enables Wendy’s to track customer experiences down to the store level within minutes. This timely information helps store, regional and corporate managers spot and address problems related to meal quality, cleanliness, and speed of service. Text analytics software caught on first ith government agencies and larger companies with information systems departments that had the means to properly use the complicated software, but Clarabridge is now offering a version of its product geared toward small businesses. The technology has already caught on with law enforcement, search tool interfaces, and “listening platforms” like Nielsen Online. Listening platforms are text mining tools that focus on brand management, allowing companies to determine how consumers feel about their brand and take steps to respond to negative sentiment.
Structured data analysis won’t be rendered obsolete by text analytics, but companies that are able to use both methods to develop a clearer picture of their customer’s attitudes will have an easier time establishing and building their brand and gleaning insights that will enhance profitability. ENDCase Study Questions: 1. What challenges does the increase in unstructured data present for businesses? 2. How does text mining improve decision-making? 3. What kinds of companies are most likely to benefit from text mining software?
Explain your answer. 4. In what ways could text mining potentially lead to the erosion of personal information privacy? Explain. 5. Visit a website such as TripAdvisor. com (or high street retailer ) detailing products or services that have customer reviews. Pick a product, hotel, or other service with at least several customer reviews and read those reviews, both positive and negative. How could Web content mining help the offering company improve or better market this product or service?
What pieces of information should be highlighted| What can businesses learn from text mining? 1. What challenges does the increase in unstructured data present for businesses? The increase in unstructured data, such as that generated from e-mails, phone conversations, blog postings, online customer surveys and tweets, presents challenges for businesses as it has no distinct form, unlike structured data, which is generated from events such as completing a purchase transaction.
The challenge of having unstructured data means that it can be difficult to interpret a large quantity of data in a short time as there are so many differing pieces of data rather than just a few structured pieces. The need to use tools such as text mining to interpret unstructured data adds extra challenges specifically those related to finance. The cost of implementing such tools can be great; not only does the technology need purchasing; the rate at which technology evolves means there will be costs in the upkeep with regards to updating new software.
Other costs will include staff training; this will have an initial outlay as well as a continuous financial impact as new technologies will require new training. 2. How does text-mining improve decision making? Using text mining improves decision making as it can analyse a vast quantity of data, condense the results into specific categories and reveal information that would have been less obvious otherwise. It can show correlations between many different factors more easily than without the text mining analysis.
Using these less obvious insights gleaned from the information it is possible for a business to make better informed decisions that may never have been thought of if it was not used. Using text mining tools allows companies to build predictive models to gain insight into both their structured and unstructured data. Using these tools it is possible to recognise patterns and common themes amongst unstructured data, particularly those gained from things such as focus groups and blogs. Identifying these themes allows better decisions as it can show correlations between data that otherwise would not have been visible.
An example of this practice is the use of listening platforms such as Nielson Online which can determine the feelings of consumers and allow a company to better make decisions based upon their customers’ wants and needs. 3. What kinds of companies are most likely to benefit from text mining software? Large companies that have information system departments will benefit mostly from text mining software as it will enable them to speed up processes that they are already concentrating on. The text mining software will allow these companies to analyse large amounts of data that would normally take weeks to work through in just days.
Other companies will benefit from smaller packages of the text mining software, particularly those that incorporate ‘listening platforms’. This will allow companies to more easily gauge how they are perceived by their consumers in terms of brand satisfaction and highlight any improvements that need to be rendered. Financial and communications provider companies can benefit from using text mining software by using it to identify their customers’ needs from their customer feedback to interpret better ways in which to retain their most profitable clients.
Marketing companies can benefit from using text mining software to implement predictive modelling to improve marketing and promotions to their target audience and retailers can benefit from text mining software to quickly identify any major issues that occur on store level to better help managers improve their stores. 4. In what ways could text mining potentially lead to the erosion of personal information privacy? Text mining could potentially lead to the erosion of personal information privacy as it gives such an increased insight into the movements and habits of the public.
Although text mining can help make improvements in the services being offered, it also gains a large amount of information about an individual. This insight into one’s personal information further adds to the ever growing ‘big brother society’ or ‘surveillance society’. With the introduction of things such as increased CCTV monitoring the streets and larger quantities of data constantly being stored by companies there is much speculation that personal privacy is quickly being eradicated. Text mining tools may be another way in which this is apparent.
An example of this is text mining tools used on holiday purchases; such a simple task can give an insight into the financial circumstances of an individual from the cost of the holiday to any extras purchased with it, as well as spending habits of that individual and other preferences. One way this information could infringe privacy is if it is then used to market other products specifically to that individual based on their prior purchases. 5. How could Web content mining help the offering company improve or better market this product or service?
What pieces of information should be highlighted? Using Tripadvisor. com to read reviews on a hotel in London it has been possible to see the differing opinions of guests staying there. The hotel needs to utilise these reviews in order to better promote their services and to eradicate any problems. Using web content mining could be the most efficient way to do this. The hotel has 736 reviews of which 630 are positive and 106 are negative. It would be inefficient to manually read hrough this amount of text and cross reference specific points that need addressing. Using web mining tools the hotel could easily find which points they can use to market their services, some which appear to be the accessibility to amenities, particularly the tube station, and which points they need to improve on, particularly apparent is the attitude of the staff. Not only will web mining easily flag up these points it will easily show trends in the feelings of the guests, which could be missed if the reviews were to be analysed manually.
The hotel would also save time and money by allowing the use of web mining as it eradicates most man power and human error. Bibliography Books Kenneth C. Laudon, Jane P. Laudon (2012). Management Information Systems: Managing The Digital Firm. Harlow: Pearson Education Limited. Online Sources Daily Mail Online (2010) Big Brother society is bigger than ever: New technology is ‘undermining privacy by stealth’. Available at: http://www. dailymail. co. uk/news/article-1328445/Big-Brother-society-bigger-New-technology-undermining-privacy-stealth. tml#ixzz1s9qMFfIg (Accessed 10/04/2012) JISC (2012) The Value and Benefit of Text Mining to UK Further and Higher Education. Digital Infrastructure. Available at: http://bit. ly/jisc-textm (Accessed 10/04/2012) Nucleus Research (2007) SPSS Text Mining. Available at: http://www. spss. ch/eupload/File/PDF/Guidebook%20–%20SPSS%20Text%20Mining. pdf (Accessed 10/04/2012) World Academy of Science, Engineering and Technology (2005) Powerful Tool to Expand Business Intelligence: Text Mining. Available at: http://www. waset. org/journals/waset/v8/v8-21. pdf (Accessed 10/04/2012)