Every interaction can be broken down to its component parts to assess, infer, anticipate and optimize such interfaces to meet the needs of the individual.
Big data technologies and machine learning methods have transformed our ability to improve interactions for customers (who look for information, solutions, or resolution through personalized experiences) and organizations (who look to provide those personalized experiences across multiple channels in the most cost-efficient manner).
Across all possible applications of big data and machine learning in CX, the following seven rise to the top. These applications, or use cases, achieve the greatest impact by delivering precise customer insights, personalized customer experiences and overall brand value.
Partner in Charge, Customer, Brand and
Marketing Advisory, KPMG in Australia
Proactive customer service offers some of the greatest potential for customer service transformation. The future of customer service will be proactive and preemptive, i.e., customers will receive information and resolutions before they contact customer service, thereby reducing stress, eliminating negative experiences and ultimately saving time and money. A combination of real-time data pipelines, predictive analytics and human-centered design can be used to analyze transactional, behavioral and relational signals that predict customer life cycle events and interaction needs. Armed with these predictions, organizations can preempt the incoming interaction and use the most efficient channels and treatments to resolve the need.
Insurance companies frequently adopt this approach to predict when their customers’ children will need drivers’ licenses or when customers will move to a bigger house. When done right, this advanced capability means an organization can proactively offer new products and services at the time the customers need them the most.
Using historical customer interaction trends, advanced scenario modeling and predictive analytics, brands can better anticipate the variances in contact volumes. Real-time data across channels (web, app, chat and social), broken down by product, customer and geographic dimensions can drive service workforce planning and optimization. Marketing activities (such as limited time offers), local/micro shopping trends and competitor actions can drive traffic up or down, and influence customer interaction needs.
Appropriate staffing saves companies money by allocating resources to the most immediate needs of the business – including frontline customer service, new product training, learning and development, and innovation. Additionally, the right level of customer support will also have a direct impact on customer and employee experience and satisfaction.
Personalized marketing is proven to be effective, but it has to be data-driven and adaptive. Predictive analytics can analyze data nearly instantly, to help companies make dynamic marketing offers. Hotels such as IHG use analytics to predict the in-room features and mix of options it will take for an individual customer to stay at their hotel. The data can suggest if a customer should be offered an upgraded room, food or additional service based on history and preferences, to maintain loyalty or drive additional visits.
In a traditional approach, many pricing models were customized for customer characteristics such as age and socioeconomic status. Today, organizations are applying advanced analytics to more granular usage and engagement data, to optimize pricing schemas that reflect their customer base more accurately.
This is especially prevalent in insurance companies. Both Aviva in the UK and Progressive in the US use telematics programs and in-car sensors to gauge how often and how well customers drive. This data personalizes the rates for each customer based on driving styles and risk profiles.
Data can pinpoint which customers are most at risk of leaving and the economic value exposure they represent (e.g. lost revenue stream). Companies that use predictive analytics to identify ‘flight risk’ factors can greatly improve their customer retention via targeted initiatives. FedEx uses data to predict with 60-90 percent accuracy which of its customers will defect to a competitor.2
Similarly, Sprint can identify segments that are 10 times more likely than other customers to cancel. By using data to identify the factors that lead to churn, and the customers most likely to churn, companies can remediate some of the customer pain points through targeted programs, actions and messages, to improve retention.
Using predictive analytics, especially reinforcement learning, organizations can tailor a customer’s experience in real-time. This advanced capability is at the core of ‘algorithms of services’ like Netflix and Spotify. A customer’s actions, such as watching a certain show or skipping certain songs, and instantaneous feedback (ratings, thumbs up/down), drives their next recommendations.
These dynamic feedback loops can inform real-time content delivery, and significantly drive customer engagement with products and services at moments that matter most.
The most fundamental use of analytics is for predicting customer needs. This is largely what makes apps by the likes of beauty products retailer Sephora so successful. By using historical product purchase and interaction trends, brands can predict repeat purchase propensities and new product and service recommendations.
These precise insights enable targeted offers that help drive marketing efficiency and sales effectiveness. Similarly, L’Occitane uses AI and predictive analytics to ensure that every section of its site meets individual customer needs.
Predictive analytics helps brands look towards the future and improve their customer experiences. These seven types of predictive analytics show just how much new data can do when used correctly and strategically. (Based on Forbes.com, 7 Kinds of Predictive Strategies for Customer Experience: Blake Morgan, Jan 2019).
*2 – Predictive Analytics 2016, Eric Siegel