Zendesk Launches Satisfaction Prediction

Zendesk today at its Relate Live user conference in New York introduced Satisfaction Prediction, a machine learning and predictive analytics feature for customer satisfaction.

Satisfaction Prediction leverages historical satisfaction survey results to predict conversations at risk of bad customer satisfaction in real time, allowing organizations to take a data-driven approach to customer service.

The machine learning that underpins Zendesk's Satisfaction Prediction learns from signals that could precede negative customer satisfaction, such as the amount of effort involved to solve a ticket, latency between user and agent responses, and language used within the conversation, paired with the customer's satisfaction rating of the interaction. Satisfaction Prediction learns from these signals, enabling it to predict whether specific interactions will receive good or bad satisfaction ratings.

Zendesk CEO Mikkel Svane, in his morning keynote, said machine learning is one of those disruptive technologies that will soon be the future of customer service.

Through years of analyzing customer interaction data, Zendesk believes it can now provide customer service teams with an early warning system for poor customer satisfaction by predicting and pinpointing customer conversations that are most likely to lead to high or low satisfaction scores.

"Zendesk's Satisfaction Prediction capabilities bring a sixth sense to increasingly complex customer conversations, empowering organizations to better anticipate a customer's level of frustration before bad interactions can occur," said Adrian McDermott, senior vice president of product development at Zendesk. "By introducing cutting-edge machine learning technology to Zendesk's customer support platform, we're using data-driven insights to help organizations build better long-term relationships with their customers."

Satisfaction Prediction uses predictive analysis of customer signals to generate scores (ranging from 0 to 100) each time a customer service ticket is created or updated. This allows agents or managers to prioritize workflows, drive business rules, or trigger downstream integrations. A machine learning model is automatically generated through an analysis of live account data to create a unique, personalized customer service prediction model for each Zendesk customer.

Keeping customer service interactions from going sour, Svane said, is more important today than ever before, due in no small part to social media's influence. "As consumers, we have an incredibly loud voice, thanks to social media. Anyone can turn into a promoter or detractor for any brand very quickly," he said. "The voice of the customer has never been louder."

Providing a level of personalization available from products like Satisfaction Prediction, Svane added, can really make or break a customer service interaction, particularly as consumers embrace the subscription economy.

The transition from traditional, transactional customer relationships to more subscription-based, convenience business models will require a customer service transformation as well, he said.

"There has to be a change in how we think about customer relationships," Svane said. "It's about the ability to retain customers and keep customers loyal over their lifetimes."

Svane outlined three principles for providing great customer service in today's super-connected world:

  1. Design service into the experience. Customer centricity, he said, needs to be embedded into every experience, including interactions with IVRs and mobile apps, that a consumer could have with the company.
  2. Empower your people to personalize every customer service interaction.
  3. Be effortless, not delightful. "Great gimmicks don't build long-time loyalty. Going above and beyond to fix one bad experience doesn't build loyalty," Svane said.

Pete Stein, the former CEO of Razorfish, also emphasized the power of predictive analytics and personalization as customer service tools. "The brands that are doing service well are predictive and doing some creative things with their data," he said.

"You have to create something that was meant for [each individual customer] and connects them with your brand," Stein said. "You need to create something that is like a gift just for them."

Related Articles

Zendesk has launched Automatic Answers, a feature powered by machine learning to help customers solve their inquiries without having to go through contact center agents. Companies can embed the technology right into their Web sites or mobile apps.

Posted July 13, 2016

Customer experience management solution provider InMoment has teamed up with predictive analytics vendor RapidMiner to deliver a more sophisticated Voice of the Customer solution that not only provides insight on what customers are doing at different phases of the customer journey, but also why they're taking certain actions.

Posted January 15, 2016

Metrics are the lifeblood of every contact center's performance, yet many of them aren't measuring the right performance indicators, the ICMI determined in a recent study. The findings, announced at the ICMI Contact Center Conference earlier this week, revealed that as contact centers evolve to be increasingly multichannel hubs, companies are struggling to maintain consistency across channels and customer touch points.

Posted October 23, 2015