Tuning in to customers: Optimizing the online experience
Creating a positive customer experience online requires a sophisticated combination of customer support, analysis and Web design. Customers are increasingly impatient with cumbersome purchase processes or unresponsive service, and the fiercely competitive market for a dwindling supply of dollars is motivation enough for companies to strive for the best quality service they can muster.
Flash retailing is a phenomenon that has attracted millions of followers who crave both bargains and designer items. In the world of brick and mortar stores, flash retailing is also called "pop-up retailing." A store is set up for a few days to sell items (often women's apparel) to private groups at deep discounts. Private online sales use a similar model in which the duration of the sale is limited and the purchasers sign up to be members of the site. The flash retailers purchase or consign overstocked inventory from other companies and sell it to consumers.
One of the leading online flash retailers in the United States is HauteLook, which was founded in 2007 and purchased in 2011 by Nordstrom. As a startup, HauteLook initially handled member calls with a small group of associates, but its rapid growth pointed to the need for a coordinated multichannel approach to handling member inquiries. The company selected RightNow CX, a customer service software product from RightNow, which is being acquired by Oracle.
RightNow CX is designed to provide positive customer experiences by integrating customer data, communication channels and relevant knowledge. "It is our job to let members define how they interact, whether that is via phone, e-mail, chat or posting questions," says Chris Purpura, director of member care at HauteLook. "Our goal is to make it as easy as possible."
HauteLook began its use of RightNow CX by formalizing its call logging process to determine the reasons for calls. "We set up categories and subcategories that enable us to find out what issues our members were concerned about," Purpura says. "One of the conclusions this data led to was that people needed more information about shipping."
Because HauteLook obtains its merchandise directly from its brand partners, shipping is a two-step process-first the product goes to HauteLook's warehouse and then it is sent out to the member. "We broke down the shipping status more clearly and began showing the arrival time as a date range rather than an expected number of days," Purpura says. "This change provided the information that members needed to track the status of their order."
Preserving the fun of shopping
HauteLook added other features incrementally; for example, chat was first placed at the point of checkout to help ensure that members received support they needed to complete their purchases. Chat was also enabled on the "Getaways" site, which offers hotel rooms at discount prices. "Live chat is the preferred method of communication for Getaways, and the conversion rate is higher when that option is used," Purpura adds.
For the future, HauteLook is planning to incorporate RightNow's Smart Assistant, which supports a knowledgebase that contains answers to frequently asked questions. "We would like to present that option in a very non-intrusive way, as a first step for members who want to use it, while preserving a positive experience," Purpura explains. "Shopping should be fun, and we want to make it easy."
Mobile technology is becoming increasingly important because it allows users so much flexibility; they can accomplish tasks or move a transaction along while away from their desks. "When our clients leveraging RightNow's mobile solutions receive inquiries from mobile devices, they are able to provide support appropriate for that channel," says Susie Boyer, senior product manager at RightNow, "and all of the information is consolidated within the client's central RightNow knowledge foundation and database."
One frustration that can severely mar the customer experience is not being able to find information on a website. RightNow has introduced Intent Guide, which supports search by determining the intent of the customer. Using a natural language interface, Intent Guide improves the search experience by helping customers locate the information or transaction they need.
RightNow also monitors social media channels such as Facebook or YouTube pages, RSS feeds and Twitter. "We can detect sentiment and score it, to make it actionable," Boyer says. However, the monitoring should have a purpose. "It does not make sense to just monitor these channels, you should be prepared to take action," she adds.
Improving customer satisfaction
Humana offers a range of healthcare services including group and individual insurance, Medicare plans, dental and vision insurance, and pharmacy services. In addition, it provides programs to promote wellness, and advisory services for several hundred medical centers. To better understand the requirements of the healthcare providers that use Humana's services, the company wanted to analyze information that was being captured in unstructured form through surveys and from call centers.
Since Humana was using software products from SAS, it opted to extend its use of that platform to accommodate the new initiatives. "Humana already had a large installed base of SAS applications, and SAS stood out compared to other potential alternatives," says Greg Hayworth, a scientist who develops predictive models and other advanced analytics projects for Humana's Provider Network Operations.
One issue that Humana wanted to explore was why the online information resources on topics such as eligibility were not being used as much as had been expected; instead, providers were contacting the call center. "The call center associates were tagging the primary reasons for the call," Hayworth continues, "but 80 percent of the answers were the first two items on our checklist, which did not provide us enough insight to be actionable."
Humana began to use SAS Text Miner to analyze the detailed notes kept by the call center associates, and achieved a much more granular understanding of why providers were calling. "We were able to take some specific steps in terms of prioritizing the information on the website," Hayworth says. "We could then address the more nuanced questions that our providers had not been able to answer previously with the self-service tools."
After the adjustments in the website, provider calls declined by 10 percent. "Without the data to explain what was going on," Hayworth explains, "we would not have been able to build the business case to support the additional features we included in the website." Adding new functionality improved the self-service for the website, allowing the customer service representatives to focus on the more complex questions.
Humana carried out a similar text analysis on free responses in a regular survey it conducts of its healthcare providers. "Previously, we did not have a good tool for interpreting this information," Hayworth says. "Reading each one was an inefficient approach and we were not benefiting from the information we collected." Humana has queued up several other projects, such as one comparing authorizations from nurses who review medical claims to the medical services that were provided, and an analysis of contracts to compare contract provisions with claims.
SAS offers several different text analytic products including one for text mining and one for sentiment analysis. "Our analytical method includes statistical machine learning plus a rules approach," says Fiona McNeill, global product marketing manager for SAS Text Analytics. "The outcome is a model that can be applied to incoming material and presented as analytical results in reports or other delivery formats to improve organizational performance." Subject matter experts can also refine the model through a point-and-click interface, adding or modifying categories.