Harnessing the power of IBM Watson
Saturday decided to work with IBM Southeast China to implement IBM® Watson™ Predictive Customer Analytics—a solution that helps companies personalize the customer experience. Watson Predictive Customer Analytics integrates and models data from multiple sources, profiling and segmenting customers based on their buying behavior, web activity, social media presence, demographics, and many other factors. Based on these profiles and segments, it defines “next best actions” based on the derived customer preferences.
Mr. Wen Zhu comments: “We saw this project as a real partnership with IBM, that would deliver mutual benefit for both organizations. As an early adopter of Watson Predictive Customer Analytics, we would gain access to the latest IBM technology and expert support, helping us explore the possibilities of big data analytics and gain deeper understanding of our customers. At the same time, our experience would help IBM optimize and refine the solution further. Together, we believed we would be greater than the sum of our parts.”
IBM and Saturday divided the project into three phases. The first phase was a micro-consulting engagement to set out a five-year strategic plan for the company—focusing on diversifying into new product categories, increasing the emphasis on online and O2O channels, and leveraging new technologies to enable this. The second phase was to integrate the company’s internal customer and sales data with the Watson Predictive Customer Analytics solution, and the third involved the integration of external data sources.
“We started by moving our largest brand—Saturday—onto the new platform, and building some models for customer profiling,” explains Mr. Wen Zhu. “Where is the customer from? How high are the heels she likes to wear? What colors does she prefer?
“We have also started performing some sales and market analysis, looking at the time and date of purchases, the regions, and which customers are involved. And we are also doing some semantic analysis of external data sources, such as assessing the sentiment of positive and negative comments on social media and websites. This is a new capability for us—we couldn’t do it before on such a large scale, and it is much more efficient with the IBM solution than trying to do it manually.”
To support these analyses, Saturday is using techniques such as product recommendations and frequency and monetary value models, which are built into the Watson Predictive Customer Analytics solution.
Mr. Wen Zhu comments: “We could do most of these analyses before using other tools; the difference with Watson Predictive Customer Analytics is that the data is modeled and the results are produced automatically, so we can monitor changes on a daily or weekly basis. We’re not dependent on one person having time to create a report—the system can provide insight whenever the business needs it. Previously our view of our customers was fragmented across many tools and platforms—now we have one platform that gives us a clear view of everything.”
Immediate insight into sales, marketing and product design
Saturday is already beginning to see the benefits of IBM Watson Predictive Customer Analytics—and as it continues to integrate more brands, business systems and data-sets with the solution, it expects these benefits to increase over time.
Mr. Wen Zhu says: “Our marketing team can use the platform to measure the effectiveness of their campaigns by monitoring customers’ responses online: how often a customer makes a comment about one of our campaigns, whether their comments are becoming more positive or negative over time, and even some of the content of what they are writing.
“For example, we recently ran a campaign with a media company, where their actors were wearing our shoes in a movie. We were able to monitor users’ comments online, and also measure the conversion rate for our online marketing activities.”
Saturday’s product design teams are also using the IBM solution to monitor customer feedback on color preferences, heel height, comfort level, and other key attributes of their products.
“It’s interesting to see that each region of China has its own preferences,” says Mr. Wen Zhu. “This helps us design products that are appropriate for particular groups of customers, which should help us to maximize sales.”
He concludes: “Before this project started, we were sure that big data analytics would be valuable for our company, but we didn’t know exactly where the value would be. Our objective was to work with IBM to explore our data and learn where the opportunities are. Now we are in a much stronger position to take the next steps on our journey with Watson Predictive Customer Analytics, and we hope to achieve at least a 0.5 percent lift in sales volumes as a result.”