What You Need to Know to Climb the Big Data Beanstalk


Understanding the Basics

You can run, but you definitely can’t hide from the term “big data” these days.  Like everyone else, lately I’ve been inundated with all things big data.  The phrase is almost like that beanstalk – you know the one that Jack climbed? It started from an innocuous bean and grew into something gargantuan overnight.  And because of that rapid growth, the explanation of it has been a bit disjointed.  Lots of terms have gotten wrapped up in its description –  some of them technical, some of them jargon, and many of them tailored towards people with an analytic background.  Therein lies the problem.  So rather than give you a blog post about why you should care about big data, let’s take a step back and make sure we all understand some of the terms you’ll hear associated with it.  After all, it’s easier to understand algebra if someone taught you how to do arithmetic first, right?

All Data is Not Created Equal

By far and away two of the most common terms you’ll hear associated with big data are structured data and unstructured data.  I did a little internet research, and here are some definitions I found.

  • Structured data:  highly organized information, such that inclusion in a relational database is easily searchable by straightforward search engine algorithms
  • Unstructured data:  information that does not have a pre-defined data model or is not organized in a pre-defined manner

Making it Relatable

Crystal clear, right?  Let’s try a different approach.  Now to be fair, there are an untold number of types of both structured and unstructured data.  But given that I’m writing this for a company who focuses on harnessing information from customer interactions, I’m going to assume that most of those reading this are probably in a position where they would want to use big data to gain insights about their customers and the experiences their customers have with their company in order to improve their business.  When put in that context, the definitions shape up this way:

  • Structured data: information about customers and their experiences that contains transactional, demographical or historical data.  Often referred to as metadata, this can include information such as products purchased, date purchased, customer tenure, or call duration.
  • Unstructured data: information about the customer and their experience that is behavioral in nature or comprises sentiment.  This may include customer frustration, competitive mentions, technical issues requiring troubleshooting, or sales offers.  And just to confirm we’re all on the same page, behavioral data is data used to understand intent and predict future actions.

Structured data often tells a company the “who”, “what” “when” about a customer such as when they became a member, what services they subscribe to, their age and where they live and the last person they spoke to within the company.  This information is valuable, and can tell an organization a lot, but it’s the unstructured data the often holds the key to the “why.”   That’s because the unstructured data holds the unfiltered point of view of the customer.

Where is it Hiding?

So where does this unstructured data come from?  The truth is it can come from a variety of places.  Imagine any way a customer can communicate in “free form” and use his or her own words.  An easy way to think of it is this.  Picture the last survey you were given about a product or service you received from a company.  The questions at the beginning that asked you to rate things on a scale with 1 being bad and 10 being good generated structured data.  The questions at the end that asked you to describe how you felt or areas you thought left room for improvement generated unstructured data.  Other examples of sources of unstructured data would be chat, email, social media and the most dominant channel – the phone.

Two is Better Than One

Companies that want to best understand the business decisions impacting their customers and generate the most accurate predictive models are realizing the need for both kinds of data.  By using analytics tools to put structure to the unstructured data, the way interaction analytics does for audio, companies can fill in the gaps in the customer journey and have a true picture of what led up to a particular event such as a purchase decision or a churn.  This level of understanding allows marketing plans to be tailored, customer satisfaction programs to be developed, sales offers to be targeted and overall business operations to improve.

Final Thoughts

So now that you understand the basics, you’re ready to start climbing that big data beanstalk.  Don’t waste the potential of that unstructured data you’re generating.  By combining it with the structured data you’re probably already studying you have the potential to take your business transformation to a whole new level.

{Photo Credit: Westy48 from Flickr}

Categories: Choosing The Best