Kurt Marko
We are in a time of unprecedented flux in consumer behavior, customer expectations and company business models created by technologies that simultaneously disrupts established businesses and spawns new ones. The genesis is the symbiotic interaction of three seminal tech-enabled developments: mobile devices, cloud services and big data analytics. The turmoil is particularly acute in retailing, however few industries and business processes will escape the ramifications from a world of multiple customer platforms and communication media that combine to form the omnichannel. Businesses, regardless of industry, that successfully exploit new communication channels, service delivery options and unprecedented sources and quantities of data will thrive by providing superior buying and support experiences. Just ask Starbucks’ visionary CEO, Howard Schultz, who articulated the goal every organization should strive for in this era of heightened customer choice and expectations:
Delivering the complete Starbucks’ experience for our customers and creating an authentic emotional connection between our customers and our partner’s has been and continues to be at the heart of everything we do.
The term omnichannel, now common, is often abused, particularly in the context of retail sales and fulfillment. I’ve previously written about the implications and opportunities (introductory overview here and application to sales enablement here) from the seamless meshing of consumer experiences across mobile devices, Websites and in-person interactions using multiple communication paths, phone, IM, email, Web chat and social networks. Still, the omnichannel discussion often focuses on retail, but its implications on customer service and support are equally significant. Used wisely, omnichannel can turn frustrating, unfruitful customer interactions into delightful, loyalty-building experiences.
Collecting, correlating and analyzing data from customer interactions across channels is the key to transforming the customer experience from nightmare to nirvana. The nexus of big data and machine learning in all its forms, including predictive analytics and even neural network deep learning, are the underpinnings of well informed, highly efficient and deeply satisfying interactions that benefit both customers and business.
Customer interactions for the mobile, social era
The impetus is clear: ample data shows significant changes in the relationship between businesses and their customers. For example, asked about various customer support options, although one survey found two-thirds preferred speaking with a live person, 23% would rather deal with the company website. The catalyst is the pervasive use of mobile devices and apps. With smartphones now an ever present companion, people expect immediate access to everything they consider relevant to the task at hand. But no one’s monogamous: smartphones aren’t the only communication tool at hand meaning people routinely initiate interactions on one device/channel and take them up later (or at least expect to) on another.
Everyday
conversations provide a useful metaphor for the best omnichannel
customer relationships. Translated to business, it means delighting
customers requires emulating the qualities of fruitful conversations.
The best conversations are continuous, naturally picking up from where they left off, but they are also interactive, efficient, and mutually beneficial,
providing tangible benefits in new and useful information. In an
omnichannel world, customers change devices at a whim and expect the
ability to start a task at one point in time, pick it up later and
maintain continuity throughout the event. Transforming the customer
experience into a natural conversation improves brand strength and
engagement and eliminates support frustrations, resulting in happy,
loyal customers.
Audacious goals with pragmatic plans
Redesigning critical and mature business processes like those used for customer support and account management is daunting, but it’s important to have both audacious goals linked to pragmatic plans. As the saying goes, a goal without a plan is just a wish. So, think big, but also strategically. The audacious part means addressing your most important and critical customer support processes. These entail support processes and tasks that have high volume, span multiple channels and are of high value to both the business and customer.
But being pragmatic requires identifying the areas where changes provide the most bang for the buck. Thus, it’s imperative to identify and prioritize channel pairs, like Web and phone support or Web and interactive chat that customers are most likely to use together: the logical fits in an omnichannel journey.
For example, banks assisting customers with payment questions often find that the number of people calling for phone support within 24 hours of being on the Web is five-times higher than average. This correlation is a clue to problems with the underlying Web self-service process, since callers obviously didn’t get satisfactory answers during their online journey. Dropping back to a secondary support channel kills customer satisfaction and increases support costs.
Customer-centric design
Too often organizations fixate on internal processes and end up making things easy for support agents or in-house bureaucracy without actually improving the customer experience. Instead, examine the customer experience problem from the outside in, as the customer experiences the process. An effective, structured technique for such outside-in analysis entails building customer personas: prototypical users representative of various customer segments and demographic groups, with correspondingly divergent customer needs and omnichannel journeys through your support system. Each persona is an archetype or idealized representation of key customer groups. Following their typical paths over several support channels through your system can reveal process bottlenecks, roadblocks and opportunities for automation.
A Big Data Foundation
Online retailers and advertisers were pioneers in understanding the value of data collection and analysis, but every organization has mountains of transaction data that can be mined to improve the customer experience. But it’s not enough to aggregate a mishmash of structured and unstructured information into a data lake. Machine learning and predictive analysis requires a complete understand of the customer. That means transaction records should also map onto a single data structure that includes auxiliary customer information from other systems like CRM, BPM, third-party consumer marketing services and customer satisfaction surveys and metrics.
Start with transaction data and metrics that can shed light on customer journeys through your support maze. Look for indicators of process inefficiency, customer frustration and cross-channel breakdowns or gaps. Identify the attempted user tasks and measure their rates of success and transfer. Then look for data identifying problems such as Web-to-call agent transfers that are clues to user frustration and problem escalation.
Use machine learning to mine data
Remember, the goal isn’t to merely provide an integrated omnichannel experience, but to be proactive not reactive: anticipate customer needs and prevent problems, don’t just solve them. Take a cue from a hockey legend, Wayne Gretzky: Skate to where the puck is going, not where it already is. For example, one of the biggest customer turnoffs is having to ask the same question or explain the same situation several times to different support personnel or online forms. Instead, anticipate problems and don’t make the customer ask a question in the first place. Predictive analysis can eliminate frustrations by not only providing context for customer queries, but by using statistical modeling and forecasting, can anticipate problems and queries.
Perhaps the most common customer interaction for any service provider involves billing questions.
When the number for a particular customer is abnormally high, it’s a sure sign of potential problems. Combining detailed call and Web logs with predictive analysis can create meaning out of seeming noise. What may be several customer interactions spread out over time now become an omnichannel conversation.
Applied to financial institutions or telecom operators, machine learning
of customer norms can identify whether the person was just trying to
change their address or traveling internationally and checking for
unusual charges. Likewise, machine learning can identify bank customers
with ‘low balance anxiety’ that repeatedly query an online or mobile app
to see whether the next bill payment will clear. Redesigning the
customer experience using data and machine learning enables proactively
sending a text message or IVR call to these customers when their balance
drops below levels required to cover upcoming bill payments by
understanding their normal monthly spending habits.
Audacious goals with pragmatic plans
Redesigning critical and mature business processes like those used for customer support and account management is daunting, but it’s important to have both audacious goals linked to pragmatic plans. As the saying goes, a goal without a plan is just a wish. So, think big, but also strategically. The audacious part means addressing your most important and critical customer support processes. These entail support processes and tasks that have high volume, span multiple channels and are of high value to both the business and customer.
But being pragmatic requires identifying the areas where changes provide the most bang for the buck. Thus, it’s imperative to identify and prioritize channel pairs, like Web and phone support or Web and interactive chat that customers are most likely to use together: the logical fits in an omnichannel journey.
For example, banks assisting customers with payment questions often find that the number of people calling for phone support within 24 hours of being on the Web is five-times higher than average. This correlation is a clue to problems with the underlying Web self-service process, since callers obviously didn’t get satisfactory answers during their online journey. Dropping back to a secondary support channel kills customer satisfaction and increases support costs.
Customer-centric design
Too often organizations fixate on internal processes and end up making things easy for support agents or in-house bureaucracy without actually improving the customer experience. Instead, examine the customer experience problem from the outside in, as the customer experiences the process. An effective, structured technique for such outside-in analysis entails building customer personas: prototypical users representative of various customer segments and demographic groups, with correspondingly divergent customer needs and omnichannel journeys through your support system. Each persona is an archetype or idealized representation of key customer groups. Following their typical paths over several support channels through your system can reveal process bottlenecks, roadblocks and opportunities for automation.
A Big Data Foundation
Online retailers and advertisers were pioneers in understanding the value of data collection and analysis, but every organization has mountains of transaction data that can be mined to improve the customer experience. But it’s not enough to aggregate a mishmash of structured and unstructured information into a data lake. Machine learning and predictive analysis requires a complete understand of the customer. That means transaction records should also map onto a single data structure that includes auxiliary customer information from other systems like CRM, BPM, third-party consumer marketing services and customer satisfaction surveys and metrics.
Start with transaction data and metrics that can shed light on customer journeys through your support maze. Look for indicators of process inefficiency, customer frustration and cross-channel breakdowns or gaps. Identify the attempted user tasks and measure their rates of success and transfer. Then look for data identifying problems such as Web-to-call agent transfers that are clues to user frustration and problem escalation.
Use machine learning to mine data
Remember, the goal isn’t to merely provide an integrated omnichannel experience, but to be proactive not reactive: anticipate customer needs and prevent problems, don’t just solve them. Take a cue from a hockey legend, Wayne Gretzky: Skate to where the puck is going, not where it already is. For example, one of the biggest customer turnoffs is having to ask the same question or explain the same situation several times to different support personnel or online forms. Instead, anticipate problems and don’t make the customer ask a question in the first place. Predictive analysis can eliminate frustrations by not only providing context for customer queries, but by using statistical modeling and forecasting, can anticipate problems and queries.
Perhaps the most common customer interaction for any service provider involves billing questions.
When the number for a particular customer is abnormally high, it’s a sure sign of potential problems. Combining detailed call and Web logs with predictive analysis can create meaning out of seeming noise. What may be several customer interactions spread out over time now become an omnichannel conversation.
Machine learning and optimization are key
There are two elements to a predictive omnichannel system: optimization and self-learning, The best predictive systems improve over time, learn from previous events, adapt to changing conditions and optimize to improve key performance metrics. These attributes are especially important in customer support systems, where the customer mix, channel usage, quality and quantity of data, and business priorities are quite dynamic.
Ultimately, first-class customer support in the omnichannel era means analyzing volumes of data to understand their needs and connect the dots along a customer’s journey. The business mantra must be: learn, anticipate and simplify.
Credit: Forbes
No comments:
Post a Comment