The Text Analytics Discussion Today
I don’t need to tell you that there’s a wealth of useful information available on the internet, you’re reading some of it right now! But it’s not just business people and/or bloggers who are posting their thoughts; it’s your customers, too. The more customers created text in cyberspace, the more marketing professionals realized they couldn’t read it all, and – like Jeff Bezos or Tony Stark – they saw an opportunity for a tech solution and text analytics data mining was born. The advancement of text analytics tools trailed only a few years behind the social media boom, and as with most new to market technologies, initial reactions were polarized:
Proponents on one side: New technology allows businesses to read Facebook and know everything about our customers I WELCOME THE RISE OF OUR ROBOT OVERLORDS
Opponents on the other: What is this flim-flammery and software snake-oil peddling? Computers can never understand the deeper meaning behind a teenager’s tweet about the new Kanye album!
As usual, the truth lay somewhere in between. Certainly the software has its limitations, but the transaction-focused firms in the B2C space quickly flocked to the software as they found that text analytics software is good at reviewing text in large quantities (text mining) and assigning basic meaning or a positive/negative emotional tone (sentiment analysis).
I recently had a terrible Thanksgiving travel experience with US Airways, and there’s value to their business in knowing I was simultaneously booked on two flights out of Philadelphia on the same day and blamed by gate crew members. (This value is compounded by the fact that I’m sharing this experience with you.)
Unfortunately for companies in the B2B space, it’s usually not possible and very rarely good enough to simply spot a grouchy customer in the blogosphere, and send him a coupon. These clients generally only provide feedback when you ask for it, so customer survey text analysis and employee survey text analysis will be the best source. Also note that these customers maintain ongoing relationships with businesses, and when faced with a complex problem, their partners (in this case, you!) are expected to understand and address this problem in a targeted way. Happily, text analytics can also help these businesses.
How Text Analytics Works in the B2B Space
1. Tracking Customer Feelings Regarding Key Topics
Text analytics software can understand when a positive or negative word relates to a particular subject within a block of text, and assign a simple score to it based on the gravity of the word chosen. In the example below, we see a strong negative word (hate) assigned to the subject (auditor):
“I hate that my auditor made a mistake, and I had to work long hours to correct it.”
Now, if you are the accounting firm monitoring feedback, you will have a score for your Auditor group. Compared to a recent time period, an analyst would note that the Auditor score is trending negatively and investigate further by reviewing the specific comments, and plan follow-up.
2. Learning What is Top of Mind for Your Customers
Suddenly everyone is talking about a new issue, and by comparing a recent period to a past one, text analytics can spot this and bring it to your attention. This is more sophisticated than any keyword search and count; the software is now smart enough to understand that various words and phrases mean the same thing by drawing on huge databases of text like Wikipedia. Perhaps a client is struggling to navigate the portal provided by a consulting firm shortly after it was rolled out. The clients may have different ways to describe it, but the software can group them and the topic will pop up as a new theme in feedback:
“I was looking around the site and I couldn’t find any of the KPI information.”
In this case, clients could reference a “site, webpage, page, homepage, hub or portal” and the software will recognize these distinct items as references to a website.
3. Monitoring Customer Perceptions Following a Change in Your Business
Most software evolves in versions, and users are forced to wait for the next generation for new features. Here we have software driven by instructions simple enough to be tweaked on a monthly or quarterly basis. In the previous consultant portal example, the firm is being passive in observing feedback about the rollout. Instead, a firm could take advantage of the software’s flexibility and be proactive, using text analytics to keep an eye-out for clients talking about this particular issue and bring this feedback to interested parties and even follow up with the clients themselves.
Did we find the truth in the middle?
While we are a long way from customer experience management software that directs action on complex issues, after considering the three uses above, doubters should see why we are using this technology more and more. Text analytics can identify trends in positive or negative language around predetermined topics, spot hot new topics, flag mentions of critical issues, and generally lead analysts to the most important feedback. The analyst can then efficiently review a manageable amount of open-ended or freeform feedback, and plan action based on the deep understanding possible from a human reader. In this way, text analytics software can be very effective in the B2B space when man teams with machine.
%20(1).png)