How text analytics helps you tune into what your customers are already telling you.
Mining the structured data in your research will tell you what your customers did, but text analytics and a little linguistics experience will tell you what your customers actually thought.
The last 20 years belonged to the Marketing Director who understood people. The next 20 will belong to the Customer Experience Director who understands data.
Surveys using structured data, those tick boxes you’ve been asking your customers to fill in for so long, will always be good for telling you what happened. But they won’t tell you why it happened.
Wrapped up in unstructured data – those ‘free text’ sources, such as customer emails and contact centre transcripts or the last question on the CSat survey where you let customers tell you in their own words what they thought of you – are the customers’ answers to questions they wished you’d asked. Given freedom to speak their minds, customers do … revealing the thoughts behind their actions.
Text analytics isn’t word clouds. (That’s especially handy when you realise how many different ways people can find to spell the word ‘aisles’.) The best software out there understands themes, even when the theme’s key words aren’t mentioned.
Think of it this way: if you and I caught the lift to the top of London’s newest skyscraper, you might say “Wow! I can see for miles.” And I might say, “Yeah, I can see the whole of London and the hills beyond.” Good text analytics software understands that we’re both talking about the view even though neither of us used that word.
It’s now possible to process, statistically and reliably, 100,000 or a million of your customers’ comments almost instantly (and certainly before your next batch of data arrives). These days, there’s no need to spend time building dictionaries, “engine training”, or other bespoke preparation. At Verbal Identity, we’ve been able to produce valuable insights, using pre-existing data sources, in around two weeks. And that was achieved with no disruption to our clients’ departments or systems.
If you’re interested in increasing your response rates to CSat surveys, setting your customers free to talk more means you can reduce the number of questions you ask. Sometimes down to just two: Rate us 1-10, then tell us why.
But the ability of text analytics to determine what’s important won’t get you any further on its own. To determine the true root cause, you’ll need to hand over the verbatims to a linguist. Why?
Well, what would you do if you were told that 38% of your unhappy customers were unhappy because of the staff and in particular, “The girl was rude”?
This comment was found frequently in a project Verbal Identity conducted recently for a Big 4 UK supermarket. Our client was delighted that –finally – they had a statistically robust, reliably repeatable way of measuring attitudes to staff. And they asked us, as linguists, what ‘rude’ meant.
We told them, as linguists, the most telling part of “The girl was rude” wasn’t “rude”, it’s at the other end of the sentence: “the girl”.
A quick bit of linguistic investigation showed us what ‘the girl’ means: it means ‘an untrained, unskilled, useless, and inferior person.’
The girl? The average age of our client’s retail staff was 45, their average length of employment was 15 years. They had on the job training at least twice a year.
The staff member may have been rude. But the real problem is the customer’s perception of the role of the person they spoke to. Shelf-stacker. Girl.
Text analytics, when combined with skilled linguistics, can point you to the very root of the problem.
In fact, language creates concepts in people’s minds and so customers’ comments are a unique source of both explanations and solutions.
In the case above, we reminded our clients that there’s a good working example of how you can create favourable impressions of staff using their job titles. You can find these people in every major city of the world, working for the world’s most valuable brand. They’re standing at the Genius Bar.
Chris is Head of Voices at Verbal Identity. He founded the company in 2011.
He is a multiple award-winning copywriter and brand strategy consultant, and has developed insights and language for British Airways, Selfridges, Sky, Ocado, the Guardian, SABMiller, and international hospitality and automotive brands. He is one of the people who claim responsibility for the line, “You never actually own a Patek Philippe, you merely look after it for the next generation.” He also writes for a national newspaper and a luxury magazine.
Josh is an Analyst and Copywriter at Verbal Identity.
He is a trained linguist (BA&MPhil, Oxon), with specialisms in linguistic theory and early narratology. He uses this theoretical base together with a keen practical understanding of storytelling to analyse brand DNA and write compelling brand narratives.