Featured article

How to get started with customer insights from customer service conversations

Note: Customer insights may not make you this happy. I mean they could, but no promises.

If you’re reading this, you’ve probably been thinking or hearing about customer insights. This all-encompassing term sums up the idea that the answer to everything you need to know about your customers, your customer experience, and your customer service agent performance is just waiting for you. And while that’s true, it’s not always that easy. In this blog, we’re going to be exploring how to get started with customer insights, including what they are, what they aren’t, and some potential outcomes that make undertaking this effort worthwhile. 

 

First, what are customer insights?

Good question. In its simplest form, a customer insight is something worthwhile to know about your customer. If that seems a bit… abstract, it’s because every business has a slightly different definition of what’s worthwhile to them. For retailers, understanding buying trends, recurring shipping issues, or new products customers are asking about is extremely valuable. But for telecom companies, it could be competitive offerings or service issues. Whatever those data points are for your business, the aggregation of them into usable answers are your customer insights. 

 

Wait, aren’t customer insights the same as analytics?

Yes and no. “Analytics” imply that you have to do some of the heavy lifting when it comes to not only the initial analysis but also interpretation. Many analytics platforms are also designed for more general purpose data analysis, which means they require you to have both the analytics skills to manipulate the data as well as the domain knowledge to interpret it. Luckily, there are also platforms that are purpose-built for customer service insights, so that you only need the latter. (Phew.)

 

So where can I get customer insights for my business?

The simplest answer is: you already have them – you just need a way to access and interpret them. Most businesses with a significant customer service organization have thousands of customer interactions, but only review a handful. As part of an ongoing quality assurance process, this tends to be anywhere from 2-5%. But other than those samples from QA, the majority of those conversations don’t provide any value. That is where your customer insights are hiding.

Using technologies like AI, you can find patterns within all of your customer interactions – from why customers are contacting you, to which issues are causing the biggest headaches, to which agent teams are struggling with which issues. 

 

This sounds a lot like customer feedback – how are customer insights different?

Surveys would be the ideal place to get a lot of this customer feedback. However, they tend to suffer from two main issues. First, they have a very low response rate, usually in the single digits and barely into the double digits for the absolute best customer experience organizations. And second, they also tend to be skewed toward very happy and very unhappy customers, meaning organizations take the average and assume they are somewhere in the middle. 

Using customer insight platforms, you can analyze all available conversations – across phone calls, emails, chats, message-based apps, etc. This gives you a complete picture of your entire customer experience across all available channels. It can also serve as an early warning system for customer issues that are happening right now – not when you review survey responses next week.

 

I still don’t get it. Do you have some examples of customer insights I can use?

Sure! Here’s one example of using customer insights in the wild, which I call, “The Case of the Coupon Code”. 

  • Here’s the scenario: Your marketing team just launched a multi-channel campaign to drum up new business with an enticing 15% off coupon code. This is a major campaign to draw in new customers with a lot of money behind it… but there’s a critical flaw. Instead of “DISCOUNT15” it was entered into the system as “DISCOUT15” (missing the ‘N’) so those new customers are getting error messages at checkout. Most just abandon their carts, but a few tenacious customers call, email, or chat to tell you what’s going on.
 
  • Without customer insights: Your dedicated customer service representatives apologize to the few customers who contact them, and enter a 15% discount code fixing each customer’s issue individually. While a few of them notice a higher than average number of customer complaints about coupons, they chalk it up to plain old user error. Weeks later, when the marketing team looks at the results, they see the campaign had little to no impact and the spend didn’t contribute to the revenue they expected. They can’t figure out why, and decide to completely change their marketing strategy without ever knowing it was a simple typo that led to the campaign’s downfall.
 
  • With customer insights: The Head of Customer Experience notices a significant spike in customer contacts related to coupon code errors. He sends a picture of his dashboard to the Head of Marketing and asks them to double check if their coupon code was entered correctly. The Marketing team finds the error and adds both coupon codes to the system along with a correction email to impacted customers. A few new customers drop off, but the majority appreciate the timely correction and complete their purchases. Weeks later, the Marketing team reports stellar results and gives a shout out to the Customer Service team for saving the campaign. (You can also imagine a slow clap during the company all hands, but totally optional). 
 

Ok, I’m in. How do I get started with customer insights?

Like most projects, it’s important to get the early steps right, but also absolutely critical to understand the end goal. As in the story above, you need buy-in from across your organization to put these insights into action, otherwise it’s a futile exercise.

1. Tap into your customer conversations: Whether you’re trying to cobble together a customer insight tool yourself or go with an already established platform, you need to be able to access your customer conversations in whichever customer service platform (CSP) you use. Many customer insights platform vendors already have pre-built integrations with the major CSPs and other voice, chat, and email providers, so don’t underestimate the time and maintenance effort required for integrations. 

 

2. Centralize and standardize: Once you can access your customer interactions, you need to pull them together. This is another subtle, but key point – you’ll get way more value looking all your channels together than in isolation. Like in the coupon example above, some patterns may be small within individual channels, but larger when combined. Also, most customer insight platforms enable you to filter by individual channels if you wish.

It’s also important to transcribe voice conversations into usable text using some form of transcription service. Not only does this make it easier for your teams to review interactions, it’s also easier for most customer insights platforms as well – and some even do it for you. Lastly, depending on your industry, your data may have to go through some kind of compliance process like personal identifiable information (PII) redaction. Again, most leading customer insight platforms can perform this step.

 

3. Categorize and analyze: Now that your interactions are ready, the platform can analyze them for a number of different customer insights. These include customer sentiment, the scale from happy to unhappy during a conversation; contact drivers, why your customer is reaching out to your organization; and reasons to review, triggers or flags that indicate highly positive or negative context. 

This last group includes asking for your agent’s supervisor, the customer saying they will report you to the Better Business Bureau, or giving your agent a compliment. All things that you probably want to look at first. Having the context behind every interaction can make it much easier to identify patterns, trends, and anomalies in customer behavior and feedback. It also provides an easy way to identify which conversations your agent team leads should be coaching on.

 

4. Take action throughout the organization: The ultimate goal of gathering customer insights is to take action. Once you can identify these patterns, you then need to prioritize which areas of your business need immediate attention – even if that is outside the Customer Experience organization. This is where getting the rest of the business onboard is key. 

But instead of playing Chicken Little and only alerting your peers in Marketing, Product, or Operations when the sky is falling, try sharing insights on a regular basis. This establishes you as the owner of these customer insights and makes the leaders of those departments more familiar with what insights look like and how they are generated. 

This way, they understand the context when you suggest making changes to your product based on customer feedback, revisiting your user documentation to address common queries, or even altering your marketing strategy to better meet customer expectations. The key here is not just to react to negative feedback, but to proactively use customer insights to drive business decisions and strategies as one collaborative team.

 

5. Measure, optimize, and iterate: Finally, it’s crucial to measure the impact of the changes you’ve made based on customer feedback. This involves going back to your customer service interactions and looking for changes in sentiment or the frequency of certain types of feedback. You can also use a standardized metric derived from these conversations so you have an objective benchmark upon which you measure every agent, every team, and your entire customer service organization. As stated above, this provides

 

Last word on customer insights

Using customer service interactions as a source of customer insights doesn’t have to be overwhelming. In fact, it should be significantly easier than managing survey responses from different channels and platforms. But the most important thing to understand is what you get out of it. Your customers are telling you every day exactly what they need, how they feel about your experience, and what you can do better. Having a customer insight platform removes the need to bombard them with ill-timed surveys after a potentially underwhelming experience and instead gives you the tools to make sure it doesn’t happen again. With this mindset, and the rest of your organization bought-in to acting on the insights you gather, you can ensure that your business is aligned with your customers’ needs and expectations and even anticipate them.

 

 

Image courtesy of DALL-E

Prompt: create an image of a working professional turning their customer conversations from phone, chat and email into usable insights in the form of line graphs and bar charts

AI Explainers: Using AI for Customer Sentiment

Wondering how AI can make it easier to understand your customers? Hear from our Lead Machine Learning Scientist on our approach to Customer Sentiment and why tools like ChatGPT fall short.