Interpreting the why
In the first few posts of our Listening Enterprise series, in which we are outlining the difference between “listening” and “telling” organizations, we detailed a few key ways in which listening organizations are uniquely suited to spot where even long-held assumptions may be wrong. The first clear marker of what we’ve come to call a Listening Enterprise was about how their frontline team is treated.
Most recently, we discussed the need to listen in a specific manner as the level of conversation complexity rises to new heights. Specifically, the best Listening Enterprises listen in stereo, not mono, mining two-way exchanges for clues about where they fit into the customer’s story.
The next marker reflects the often unpredictable nature of the customer journey and the evolving expectations for CX excellence across many channels.
Customers are entering and exiting new and changing channels, some optimized for auto- or self-service while others remain largely human interactions. Buyers are highly informed coming into interactions, making it difficult for reps to know where to start the education process (and nobody wants to have to repeat things they’ve already told you!). Workforces are still enduring generational shifts that bring new vocabulary.
What’s worse, all of this newness – manifested in the form of new terminology, new dialects (how does one speak to a bot?!?), and new expressions – presents itself in an untold number of combinations. All this adds up to an all-time high in language variability within customer conversations.
To be sure, there are literally thousands of ways to articulate most every important facet of a customer experience. In this world, where scripts of past are rarely prologue for actual dialogue, it feels near impossible to ever anticipate all the ways things must be said. What Tethr customers are finding is in this world, what is said is less important than why it is said.
Listening organizations understand that high language variability means interpretation of dialogue is at a premium. Telling organizations, on the other hand, assume sticking call transcripts into a CRM tool as notes counts as listening. That’s like conducting a United Nations meeting with 100 note takers and zero interpreters. Having a record of what was said is not the same thing as understanding the conversation.
The best listening organizations treat interpretation as an active exercise, rather than an historical review. Customers are in continual dialogue with brands. That meta-conversation can and should feel like an opportunity to mine on an ongoing basis. Just as great research findings almost always beget more questions, interpretation of customer conversations often produce more “why’s” to explore: ‘wait, if THAT’s what they mean there, I wonder if…’. The good news is, quite literally, technology now makes it possible to ask customers big questions at scale one morning and have answers the next – by doing nothing more than conducting the conversations you’re already having!
Business problems like customers bouncing out of the website and into your more expensive call center can be addressed through simple agent prompts such as asking, “what was the last thing you tried before calling?”, then using a tool like Tethr to mine the customer answers. Or lack of clarity on marketing campaign attribution can be addressed by seeding inquiries with sales reps asking “what drew you to us?” and researching what comes next.
But it’s one thing to understand what’s going on inside a customer decision and what your teams said in relation – this is where even the most advanced “telling” organizations would stop. “Listening” organizations examine far deeper to ascertain the connection to experience-related (e.g. NPS, Effort, sales process advancement, brand/product perception, etc.) and economic outcomes (e.g. actual churn, closed sale, etc.).
Companies like this are looking for correlations, adding precision to their understanding and reducing false positives. At a deep level, these teams are trying to measure why certain customers respond in ways that correspond with the supplier’s unique value and gauge the degree to which their teams are striking the right chords.
There are many methods for landing more Why-level insight. Some of the more advanced approaches involve building data models that incorporate key outcome data to run against thousands of phrases within specific call sets. One Tethr customer in the insurance space has extensively measured how specific initiatives to lower customer Effort drove improvements in NPS.
But starting exploration needn’t require a PhD in data science. There’s a simple truth about conversations that, when harnessed, can power quick and nuanced analysis. It’s such a simple truth, in fact, that it’s easy to dismiss or forget: in dialogue, sequence matters.
Let’s say you just hopped off a phone call telling a company you’re done using their product. In that same conversation, the agent offers what amounts to an empty apology. As that company examines what went wrong, the answer is materially different if you said you were leaving before or after the fauxpology. If before, it likely means something else was broken in the customer experience. If after, it may mean the agent themselves contributed to your decision (as Tethr analysis clearly indicates the harmful effects of what will be perceived as “Powerless to Help” language).
Or maybe you received a discount after calling a call center. The transcript would pick up on “discount” and bucket that with all other discount situations. But interpreting the sequence might reveal whether that discount was even necessary. Was this a customer bluffing to leave and a rep too quickly offering a discount as a crutch to save the situation? Here, a discount is bad. Or was the discount offered after the customer mentioned a desire to bundle a few additional purchases together? Here, a discount is a sign of something good. Two completely opposite interpretations, all born out of when things happened in the course of a conversation.
Tethr customers now have the ability to bake such complex “If/Then” combinations directly into how they build and tune language categories. We call this feature “Conditional Categories” and, once employed, it allows users to start measuring key sequences that profoundly change our understanding of customer perception.
Of course, getting to a true understanding of how perceptions are correlated with specific situations requires more formal pairing with actual outcome data. Did they customer buy or not buy? Did they actually leave or were they just bluffing? Broken aspects of the customer experience are best fixed using precise diagnosis tools. But exploring sequences – ‘why the heck would they say that?’ – can light a path and hone parameters for more formal diagnosis exercises.
In our next post in the Listening Enterprise series, we’ll discuss the last marker which involves extending your lenses far and wide. Want to learn more about how Tethr helps companies land Why-level insight? Click here to request a demo.