In the customer centric world, it is important and imperative for organizations to be proactive about the preferences of their customers and take effort to listen to their feedbacks. Such effort would help the brand understand the customer’s future needs, and cater to it in a cost effective way creating, a win-win scenario for the customers and the organizations. There is a commonly known fact in the industry that, retaining an existing customer involves lower investment in terms of marketing dollars and time than acquiring a new customer. Long term customers are very loyal and inclined to recommend the products or services through word of mouth, social recommendations, online reviews, and thereby, act as a brand advocates in front of their friends and family members. Corollary to the concept, dissatisfied customers can cause significant damage to the reputation of the brand. Considering the importance of customer satisfaction and loyalty, identifying the underlying factors that impact the customer satisfaction is crucial for business.
Surveys and reviews are the most cost effective ways to receive direct and unbiased feedback from the customers. Monitoring the reviews or survey feedbacks from the customers and prospects not only helps the brand to understand the perception of their target audience, but also create actionable opportunities for the Online Marketers, Online Merchandisers, Call-Center and Warehouse Managers to improvise their Order Cancellation Rate, Merchandise Return Rate, Call-Center Operational Cost, Marketing Cost-per-Acquisition and Margins at the Category and Product Level.
The quantitative data coming from the Web Analytics or in-house reporting solutions can provide, a Health Scorecard for the business and generate opportunities for optimization on the website but, it doesn’t provide any insights in-to what challenges individual customer segments are facing, and how can we proactively avoid those challenges at the operational level. In order to gain deep-dive insights in to operational challenges faced by the prospective customers during their interaction with the online storefront or, during their offline interaction with the call-centre reps or, their order fulfillment experience; it requires mining of their qualitative data generated via Survey Platforms and Reviews and Ratings Solutions.
Reviews and Survey feedbacks can be quantitative (ratings, scores etc.) or qualitative. While, the quantitative data can be analyzed easily, the qualitative data involves the much complex task of summarizing the verbatim. The text needs to be churned using appropriate text mining techniques to understand the context of the reviews and the expressed sentiment. For one of our prestigious client, the quantitative survey data related to a client’s website showed a negative trend, but when we combined the outcome of our text mining exercise of their qualitative survey data, we learned that, the visitors were disappointed as they could not locate a particular store using the store locator feature on the website. The issue was immediately fixed based on this insight, and generated an additional 15% footfalls in the local branded store. So, in nutshell, Monitoring reviews or feedbacks on regular basis can help the brand to bridge the gap between user expectations and brand’s offerings.
Nabler’s Predictive Analytics team was approached by a large multi-channel ecommerce retailer based out of United States to categorize; the reviews submitted by their customers on the website into various pre-defined parameters. Based on this categorization, Client wanted to share a list of priority actionable
items with their Merchandising, Call-Centre and Warehouse teams, which helps them to improve the quality, durability and fit of their merchandise, reduce the abandonment rate during the checkout process, reduce the % of calls in the call-centre, and deliver a memorable shopping experience to their customers during Pre and Post Fulfillment.
Nabler team implemented the technique of training a text mining algorithm using a sample review dataset shared by the client. The algorithm later categorized the textual data based on parameters and sentiments based on the previous machine learning knowledge. The reviews were categorized into parameters relevant to the business such as delivery, customer service, quality and cost of product. The reviews were further categorized, to identify the sentiment related to each of the pre-defined parameter. For ex: A product review can be categorized as positive for price while negative in terms of delivery. This helps the brand retain the cost of the product while addressing the delivery related issue.
Based on our Text Mining exercise of their Qualitative Reviews data: