All customer support teams have experienced the “stomach drop” feeling that typically accompanies hearing that a big customer is canceling. And although it sometimes feels out of left field, customer distress doesn’t happen overnight. Rewind to a few months ago. Were there any indicators that your customer was unhappy with the relationship? More often than not, the answer to that question is yes.
Perhaps the customer submitted multiple requests for help on a single issue that went unanswered. Or maybe you recently hired a new crop of customer support agents that aren’t able to answer inquiries as fast as your senior employees. Did you see an increase in how long it takes to resolve each ticket? What about the quality of agent responses, have ratings from the customer dropped in recent months?
As your business scales and your support team grows, customer health becomes difficult to measure. In a B2B environment, where there are multiple users on the other end of the line at each customer, predicting and preventing customer churn is even more challenging.
Enter: IBM Watson. Used by industry-leading companies worldwide, IBM Watson is a sophisticated question-answering computer system that combines AI and data-driven software to perform complex analyses of unstructured data. In customer support, this technology can improve an agent’s understanding of customer distress and assist in painting a complete picture of account health.
Using data-driven indicators to identify customer distress
Tracking the impact of one-off “negative” interactions with a single customer over time is challenging without the help of innovative tech. Aggregating customer interactions into a single B2B customer support platform allows agents to gain a single, 360-degree view of each customer.
Without powerful support software, teams have a limited view of the customer relationship. Just because a single contact is happy with their experience, it doesn’t mean all contacts at the customer are pleased.
In the best customer support platforms, users get a real-time view of these early indicators via a customer distress index (CDI). This is a score assigned to each company based on information including tickets created in the last 30 days, tickets currently open, average time open, average time to close, total tickets created during the lifetime of the customer’s account, agent rating, ticket severity and more. It’s a great way to understand immediately how a customer feels about your business without needing to dive deep into the data.
The best part is that your company can weigh each metric based on your own personal importance for an individual customer. For example, a customer may have requested in the past that all tickets be closed within 30 days. For this customer in particular, you can place a strong emphasis on ticket close time when the CDI is being computed.
The importance of a proactive approach in managing customer relationships
With a CDI solution in place, your business will know when customers aren’t happy. The next step is making them satisfied customers once again. The days of customer support agents waiting around for questions or comments from customers are long gone. Especially in highly-competitive fields like B2B SaaS, competitors are actively targeting your customers, pointing out potential weak spots with your business and promising better results. To provide the best possible experience for your highly-valued customers, it’s important that support teams are proactive, not just reactive.
Limiting your support team by allowing them to only communicate with customers around just the ticket topic is doing the entire organization a disservice. A better approach to proactive customer support is to encourage agents to open up the conversation when they are working on a ticket with a distressed customer. When the issue is resolved, an agent can if everything else is going OK with them and your business. If the customer speaks up, it’s a great way for support to make a warm hand-off to the customer success team. They are the customer retention specialists, not the support team.
Leveraging IBM Watson sentiment analysis to make better decisions
Regardless of whether the success or support teams are talking with customers, both groups can do their research much easier thanks to IBM Watson. When responding to a customer, written communication like email and chat are too often left open to interpretation. Small changes in punctuation or word choice can change the entire meaning of a comment. For example, “This will work!” and “This won’t work.” are similar in syntax but mean completely different things when it comes to customer satisfaction.
In these scenarios, determining the sentiment of the message, and subsequently its level of importance, can be challenging for a support agent. To expedite this process, support teams can deploy IBM Watson’s natural language processing (NLP) to analyze customer communications through tone, word choice, punctuation and level of formality. With this kind of sentiment analysis, tickets can be automatically categorized as “satisfied,” “excited,” “polite,” “sad,” “frustrated,” “impolite” or “sympathetic,” which helps agents understand how they should respond and prioritize their work.
Sentiment analysis can also be useful in understanding how your customer support team’s written communications will be perceived by your customers. With this information, managers can highlight training opportunities for agents in order to improve the tone of their responses.
Scoring sentiment to paint a picture of customer health
Because most communications from a customer contain several different sentiments, intelligent support software will highlight each sentiment and the corresponding degrees of confidence. For example, text that is analyzed with the results of “Frustrated 66% Sad 74%” would indicate with a 66% confidence that the author of the text was expressing frustration and 74% confidence that the text expressed sadness.
While valuable at the action and ticket level, sentiment scoring really gets interesting at the customer level. Software with IBM Watson looks at the past several years of sentiment data across the account and displays customer sentiment on a scale from 0 – 1,000, where “negative” sentiments lower the score, and “positive” ones increase it.
Avoiding customer frustrations before they arise will ultimately aid in reducing churn for your business, and luckily, you don’t need a magic 8-ball to predict when a customer is at risk. If support teams are armed with the tools and data they need to understand how their interactions impact customers, the overall health of your customer base will increase. Innovative technology, sentiment analysis, and sentiment scoring are important pieces of that puzzle, especially in B2B where customer relationships are inherently more complex and not only what you say but how you say it truly matters.
social experiment by Livio Acerbo #greengroundit #thisisnotapost #thisisart