TDWI Blog

TDWI Blog: Data 360

Blog archive

Five best practices for text analytics

I’ve recently spent a lot of time talking to companies about how they’re using text analytics.  By far, one of the biggest use cases for text analytics centers on understanding customer feedback and behavior.  Some companies are using internal data such as call center notes or emails or survey verbatim to gather feedback and understand behavior, others are using social media, and still others are using both. 

What are these end users saying about how to be successful with text analytics?  Aside from the important best practices around defining the right problem, getting the right people, and dealing with infrastructure issues, I’ve also heard the following:

Best Practice #1 - Managing expectations among senior leadership.   A number of the end-users I speak with say that their management often thinks that text analytics solutions will work almost out of the box and this can establish unrealistic expectations. Some of these executives seem to envision a big funnel where reams of unstructured text enter and concepts, themes, entities, and insights pop out at the other end.  Managing expectations is a balancing act.  On the one hand, executive management may not want to hear the details about how long it is going to take you to build a taxonomy or integrate data.  On the other hand, it is important to get wins under your belt quickly to establish credibility in the technology because no one wants to wait years to see some results.  That said, it is still important to establish a reasonable set of goals and prioritize them and to communicate them to everyone.  End users find that getting senior management involved and keeping them informed with well-defined plans on a realistic first project can be very helpful in handling expectations. 

Best Practice #2 – Manage expectations among business analysts (and statisticians).  Most people who deal with text analysis believe that  “looking at text data is very different than analyzing structured data.”  It turns out that some analysts get uncomfortable when dealing with unstructured data because they’re used to building models with extremely high accuracy using well-understood structured data.  The reality is that with unstructured text, you might only achieve 70-80 percent accuracy (for example, in sentiment analysis).  Therefore, these analysts need to enter into the analysis with their eyes open to this ambiguity and decide what they can live with in terms of analysis.  End-users I’ve spoken to say that this can take a little time and patience, but with some education, it does work.

Best Practice #3- Keep it visible.    I have spoken with a number of companies that understand that in order for text analytics to keep a seat at the (executive) table it is important to keep it front and center.  How do they do this?  Some companies distribute ongoing analysis while others distribute daily customer quotes from unstructured data sources.  Many find that the emotion and feeling in customer feedback can really capture the attention of senior leadership.  Other end-users have cited the ability to “tell the story” with your text data, and that story can be very powerful and provide visibility to the analysis.  

Best Practice #4 -Dig, Dig, Dig.  Dig deeper than just monitoring.  Many companies start off with social media analysis as the first part of their text analytics journey.  However, many social media analytics platforms are nothing more than listening posts that will give you some hint as to the buzz around your product and if it is positive or negative.  At the end of the day, that doesn’t provide you with much information.  They may be a good first step in getting your feet wet.  However, companies that are successful in utilizing text analytics tools believe that they are just touching the tip of the iceberg in the kinds of analysis they can perform with a text analytics tool.  They are integrating data sources (i.e. structured and unstructured) and digging deep into the data to determine the why around the what of certain issues.  This means visualizing the data as well as utilizing more sophisticated methods to analyze it.

Best Practice # 5- Actionable feedback requires a way to take action.   Text analytics can provide you with significant insight as to the “why” of a behavior.  However, you need to be in a position to make it actionable if you want to derive the most benefits from the technology.  For instance, you may determine that a group of customers are unhappy about a certain product or that a feature you provide isn’t meeting expectations and that is why they are no longer customers.  That’s great insight, but you need to be able to act on it.  Of course, action can come in a number of flavors.  It can be as simple as having a manual process in place to deal with the insights.  Or it can become more sophisticated.  For example, some are operationalizing a process by using tools that help to close the loop with their customers by routing comments to agents who can then reach out to these customers.  

Of course, many of these best practices are useful for any kind of advanced analytics.   However, as text data takes its place in organizations, it is essential not to forget them.  It is also important to remember that there are differences between analyzing and utilizing structured and unstructured data – a topic I will explore further in future posts.


Posted by Fern Halper, Ph.D. on February 14, 2013


Comments

Average Rating

Add your Comment

Your Name:(optional)
Your Email:(optional)
Your Location:(optional)
Rating:
 
Comment:
Please type the letters/numbers you see above