Machine learning. Artificial intelligence. Big data. Deep learning. These terms have saturated the modern business lexicon and permeated the zeitgeist. What is your experience with these buzzwords? You’ve certainly read about them and likely talked about them, but have you implemented them? Are you leveraging them to improve your customer experience? Would you like to learn where to begin?
Centriam is owned and operated by data scientists who are obsessed with getting organizations to act on customer data. Actions that drive customer experience and company profits. Modern machine learning is an important component to drive actions. To help leverage customer data using machine learning, we’ve written a guide for moving Beyond the Hype to enhance customer experience with machine learning. To kick-off our machine learning series we are starting with a few examples. Centriam identified three distinct areas where these modern algorithms can be applied and selected illustrative examples for each.
Where no analytics exist, develop a new way to measure existing processes. Many current business operations generate data that never gets investigated, often because the data is too large or unstructured. Machine learning can unlock the latent potential of these everyday activities.
Adobe is utilizing machine learning to analyze text in help desk tickets. By studying trends in system failures, the system can identify those events which precipitate outages. Using this information, Adobe can work to prevent or mitigate failure events. In addition, reintroducing the data from the triggering events into their machine learning algorithms will enable Adobe to create “self-healing” capabilities. For instance, using machine learning, they discovered patterns in password reset requests which created large backlogs. Adobe developed automated proactive messaging which drastically limited the requests. Absorbing this work from existing IT staff will free up time for advancing other projects.
Stitch Fix, an online subscription and personal shopping service, is leveraging machine learning to upend our notion of what retail can be. Stitch Fix has no stores, brick-and-mortar or online, but instead ships clothing to its customers. Customers fill out preference surveys, measurements and personal notes, which are supplemented with Pinterest boards, chatbots and historical sales. While this information is available to most brands, they do not utilize it. Stitch Fix’s machine learning algorithms ingest this information and returns clothing suggestions to stylists, who make the ultimate selection. This model eliminates most overhead that retail brands face.
Where analytics is outdated, unlock new potential by updating your methods. For decades, traditional methods have been used for personalization and optimization actions. These techniques were groundbreaking when first developed and have even gotten better over time. But newer machine learning algorithms can vastly improve upon these now-routine methods.
North Face uses machine learning to generate a personalized, interactive online shopping experience. While the outdoor retailer had an existing program for customizing its website, it was limited in scope. The new method incorporates purchase and browser history with a series of preference questions to recommend clothing and gear. By simplifying the searching and sorting process for its customers, North Face decreases friction while enhancing the shopping experience.
US Bank holds an immense amount of customer behavioral data but has not always delivered actionable insights from this wealth of data. Meanwhile, they operated a personalization program without sufficient distinguishing power. Modern machine learning methods enabled them to capitalize on their abundance of behavioral data to build a better personalization engine, unearthing patterns that were previously buried. Progressing from a paradigm of what happened to what will happen has unlocked key growth opportunities. After implementing the program, U.S. Bank saw a more than twofold increase in conversion of top-ranked leads.Where analytics is outdated, unlock new potential by updating your methods. - Click to Tweet!
Procter & Gamble is harnessing the power of machine learning throughout the customer journey. From building skin advisor apps to developing payment processing bots, the consumer goods multinational is actively seeking out cognitive technology initiatives. They are also employing machine learning to improve their existing analytics, benefitting both the company and its customers. They have used these techniques in chatbots, but also to refine marketing spend, optimize their supply chain, and improve trade promotion.
When on the cutting edge, new ideas will be unlocked. Modern technologies and new customer touchpoints precipitate avant-garde analytics. The intense speed of growth and change in automation, computing and robotics will continue to spur new data sources, and new opportunities for personalization, optimization and experience improvements.
KFC is using facial recognition software powered by machine learning algorithms to deduce customers orders. The program gathers data such as facial expressions, gender, and other visual features in concert with order history to generate recommendations while customers are in line. Beyond a mere novelty, the fast food giant hopes to improve the experience for its customers and boost cross-selling.
The Associated Press has employed machine learning to produce content for its wire services. The news service found their staff was insufficient to write quarterly earnings stories for the more than 5,000 publicly held companies in America. Despite the demand, reporters were only able to compose 300 in a timely fashion. So, AP trained algorithms to write short earnings reports for every company, increasing content while freeing their staff to focus more deeply on the larger news trends within the business world.
Where can you apply machine learning to boost customer experience in your organization? While deploying facial recognition software may not be right for your business, there are sizable enhancements that can be made. In a recent review of the landscape, McKinsey reported that newer machine learning algorithms nearly double prediction performance versus existing advanced methods. Improving the accuracy of next best action predictions twofold or being twice as accurate at forecasting churn is extremely valuable. We suggest four areas where machine learning can make a meaningful improvement over traditional methods in our latest white paper Beyond the Hype: Boost Customer Experience with Machine Learning. Outlining where traditional analytics are lacking and giving specific suggestions of modern methods to apply in each situation, we dive into tips for everyone from data scientists to practitioners.