The Pecta-Hecta-Byte Big Data Challenge (or Drowning In Your Data Lake)

Big Data can become Big Boring in a hurry.  The use (or abuse) of this overused term may have set the all time speed record in terms of velocity to meaninglessness.  Most of what we read and hear goes something like this, “There are 27 gazillion pectahectabytes of data out there, they are quadrupling every thirty-three milliseconds and if you just have the right combination of data scientists and software applications you will likely become the single smartest and wealthiest person since Bill Gates.”

So how do we parachute back down to planet earth, bring some sense and order to this mega-hypathon and truly drive value for companies & brands?  As a passionate Customer Experience executive and tireless marketer, I prefer to cut to the chase and focus the conversation on maximizing measurable, sustainable value.  Of course there is incredible potential to be derived from charting the multiple paths that will provide the right, properly weighted combination of survey-based data, transactional data and in some cases, such as in Telco, network data as well.

One Constant Nevertheless Remains: All Data, Big or Small, Starts Out as a Very Significant Cost Challenge.

Yes, every year it does indeed cost less to process data, less to store data and as the quantity of raw data continues to grow so does the potential to harvest golden nuggets from that very data which created all of this hype in the first place. Nevertheless, collecting data, ingesting data, aggregating data, discovering valuable insights from data, reporting on those insights, enabling the downstream usage of data driven insights – each of these crucial steps requires technology and labor, none of which come free.

Not only can their be significant cost involved, these actions each need to occur with machine-like regularity and precision and increasingly in real time before any value whatsoever can be extracted.  Despite what some vendors like to claim, we are not yet close to a magical point and click “Instant Win Big Data lottery” moment where everybody wins.  In fact, in absence of any publicly available ‘data on big data projects’ (sorry) it is likely that more than fifty percent of all Big Data projects undertaken to date have not met the lofty expectations with which they were conceived.  What I do know for sure is that more often than not, these types of real challenges are confronting our clients every day at The Customer Xperience Company.

So what to do?  Despite the Hype, the Promise and Potential are Indeed Immense.

Job 1.  Understand and Prioritize Your Business Problem(s), the Challenge(s) and/or the Opportunity. 

What are your organization’s top three to five business priorities?  It may turn out that leveraging data driven insights won’t help you solve for these priorities in a tangible, sustainable and cost effective manner.  Ok, if that is the case then it’s probably not worth the effort.  However, before you reach that decision there are a few important steps you need to take.

Job 2.   Define The Right Analytic Metrics Framework

To solve each of the top business priorities that you have selected, you need to go through the process of defining what data driven questions you would like to answer, how you would quantify those answers and ideally undertake this exercise at the single customer or at least individual customer segment level.  The more detailed your investigation of these potential uses cases are, the more reliable your analysis will be.  We often find that a small, cross-departmental team is best utilized for this effort so that multiple points of view and knowledge sets are brought to bear in providing the most objective situational analysis.

For illustrative purposes, here below I have provided some sample questions regarding the always mission critical “customer churn” use case.  These are just ten samples of the many questions that you might want to examine: 

• Are you in a business where there are sufficient data points that can truly help determine, in advance, if a customer is likely to churn?
• What revenue and profit results would be achieved if you actually diminished that churn?
• Do you have a reliable, dynamic segmentation schema set up from which to start identifying distinct, actionable segments and then asking differentiated questions about those segments or personas?
• Does each segment/persona have a measurable $ value assigned to it? 
• Are you facing churn equally across each of your customer segments and what is the $ value of that churn?
• What percentage of your customers are churning each week, month, quarter, year?
• Have you differentiated the way in which you interact with your at risk customers and how has that changed behavior and customer lifetime value (CLTV)?
• How, if at all, are you measuring CLTV and the impact that different offers, contact strategies and content have on it? 
• Are you seeing a drop off in CLTV when customers engage with certain channels or purchase certain products or services, which ones and what are the $ values of those declined?
• Are you measuring NPS by segment and is there a clear correlation between NPS and CLTV?

The types of questions I have listed above (and many, many more) all require hard-hitting analysis and insight development. They are indeed “Big Data” questions however they are substantive, actionable and, arguably, essential to drive incremental revenue and profits for your business.

Job 3.  Data Volume & Data Depth Inventory

You may not have the volume or depth of customer data required to answer these types of questions about your customers, if that is the case then you may want to hit the Big Data Pause button.  However, if you do have the data, then, in close collaboration with your IT colleagues, you will want to undertake some rapid analysis in order to determine the level of effort, steps required and technology required to enable a winning data driven strategy. That process might go like this:

Utilizing a simple grid table, in column 1. Write down your top three to five business priorities outlined in no. 1 above.  In the adjacent column, write down the data driven questions you would like to answer (see questions in no. 2 above).  Then, in a column immediately to the right of those questions you should list the categories or data types that you would need to access in order to quantify the answers you are seeking.   After that, and one more column to the right, you should seek to quantify the volumes of data currently or potentially available.  Then begin to document if you have the ability to collect, store, analyze and process that data and persist it over time.  As you think about that here are three critical qualitative “sanity check” questions you should be asking yourselves:

Data Completeness:  are you able to access a complete set of data regarding your prospects and/or customers?  Can you track their behaviors and content consumption prior to transaction even when their names are unknown to you and they exist solely as a bunch of cookies or device Id’s?  Are you able to capture behaviors on your web site, in your contact center, as well as all other applicable channels? 

Insight Creation:  Do you have the tools and skilled personnel to sift through all of the data and develop meaningful insights that can drive measurable incrementality?  Are you soliciting feedback through surveys and questionnaires and is that information properly weighted relative to the other data sets you have?

Insight Deployment:  Once created, actionable insights need to be used and that typically requires the ability to personalize and optimize content, offers, channel usage and achieve all of that quickly – even possibly in real time.  Do you possess a campaign planning and management platform and is your insight creation technology integrated with it?

Conclusion:

Big Data is not the point.  The cost effective creation of scaled, deployable, value creating insights is the goal one needs to keep focused on.  No company can afford for this to be a random exploration effort just because we have been inundated with a level of hype bigger than the ocean of data that apparently surrounds us all.  Like most big opportunities, there remains a dire need for process, precision, order and rational thinking.

Coming Soon: We are not just left brained.  The power and awesome potential of ideas and content has become increasingly elusive and is often about things that have little to do with algorithms.

Marc