A ten-week program. At the end of which, you know exactly how consumers think about your product, why, and what you can do to increase their ratings. Translating customer feedback into actionable insights. By means of advanced text analysis and with a model that precisely predicts which aspects will increase the rating of your product the most. That’s SfCI, a new program from CQM. For anyone who wants to develop a winning product.
Online reviews, messages on social media and feedback on customer forums. These are the things that largely determine the success of your new product. Potential buyers are won over by rave reviews or, if the reviews are bad, purchase your competitor’s product. Reviews and star ratings from consumers are an absolute gamechanger. They appear in their tens of thousands across the internet and can make or break your product. There are all sorts of tools with which you can monitor these opinions, but CQM takes things a lot further: we find out what motivates your customer to give a rating, and predict which improvements to your product itself, or to elements surrounding your product, will have the greatest effect on those ratings.
SfCI stands for: Straight from Customer Insights. CQM has developed a program with workshops where with you we collect, structure and analyze this essential information, and couple it with concrete measures to improve your product’s scores. This can immediately help the rating of your current product, for example through better service or marketing. But it also provides objective insights for product improvement or developing new products.
The program normally runs for ten weeks but can last longer. However, the program’s five steps never change. Following a kick-off, each step ends with a workshop where we discuss the results together and decide on next steps. The program ultimately results in three or more actionable insights: concrete measures that you can take to improve the rating.
Step 1 in the five-step plan is determining the starting points. We determine which products we are going to consider, and this can include products from competitors? We also determine the drivers: the topics that influence the star rating. In step 2, we collect all texts from the Internet, e-mail traffic and complaints databases. This involves analyzing open, unstructured (or ‘verbatim’) texts. This step is automated, including any translations. In step 3, algorithms and models extract the key topics from the texts. These are linked to the drivers. The texts are also given a sentiment. Step 4 links the drivers to the star ratings: which drivers can most increase the overall star rating? In the fifth and final step comes the real insight: predicting what will happen when certain aspects of the product itself or related to the product are improved, also compared with competing products. This finally eventually leads to the three or more actionable insights.
CQM has now completed the new SfCI program with a number of customers. “In each case, it has yielded concrete insights from which the customer could really benefit,” says CQM senior data scientist, Peter Stehouwer, initiator of the program. "Standard tools tell you what is being discussed, but don’t make the link with the star rating. It’s quite possible for a consumer to discuss issues A and B at length, while only A actually has a significant impact on their star rating. With SfCI, we delve deeper into the how and why with our client. And in a quantified, and thus objective, manner. In the current open but highly competitive economy, with opinionated consumers and multiple communication channels, this is the way to arrive at a winning product.”
CQM’s 5 steps SfCI program
1. Understand & Scope
2. Scrape & Prepare
3. Drivers & Sentiments
4. Model & Validate
5. Insights & Predictions
Social media messages from around the world about a specific baby stroller segment were analyzed. Over 18,000 posts were automatically summarized into a manageable number of clusters. This helped Dorel give direction to their marketing strategy and the development of new products.
Reviews of a specific type of product were analyzed to answer such questions as: why does product A score 0.4 stars less than product B? Based on internet reviews and ratings of products from both Philips and their competitor, our analyzes predicted the impact of development decisions, such as product features, on consumer satisfaction.
Analysis of text data as input for marketing and quality processes.
Want to know more about the new SfCI program from CQM?
Get in touch with Peter Stehouwer.