“Not everything that can be counted counts, and not everything that counts can be counted.”
Albert Einstein

Einstein was a genius. He was way ahead of his time when he uttered these words, but he could as well have been speaking about customer data.

Today, we are faced with a deluge of data from many sources. Some of it happens in real time while others need to be processed in batches. Some of it is structured data generated/stored in our applications while others are unstructured, living in the realms of emails and social media. Some of it is generated in domains that we control while others happen outside.

This is a problem with customer data too, which is fragmented across multiple systems – Marketing Automation, CRM, Support/Help Desk, Community, and more – as the customer moves through various phases of the customer lifecycle. Each system is used by a different group to perform a specific (set of) function(s). This fragmentation ensures a view of the customer that is reminiscent of the six blind men and the elephant and is the main reason a Customer Success Automation platform needs a Data Engine at its core.
blog-dataengine-6blindmen1

So, what does the Data Engine need to do? To paraphrase Einstein, it needs to:

  • Make everything countable (to the extent possible), and
  • Focus on what counts

Make Everything Countable
To state the obvious, data needs to be brought together before it can be acted upon. To make this happen, the data engine must:

  • Include out-of-the-box connectors with common systems of record and engagement such as Marketing Automation, CRM, Support/Help Desk, Community, Engagement, and so on.
  • Have a sophisticated tracker that measures product usage in real-time or integrate with 3rd party trackers such as Google Analytics, Mix Panel, Omniture.
  • Have a standards-based integration interface that can be used to build custom integrations that work in real-time or batch mode.
  • Process structured and unstructured data generated in other systems, email, social media and other digital sources.
    Stitch together data generated across these systems to provide a unified view of the customer in near real-time.

Focus on What Counts
The next step is to make what can now be counted count.

More likely than not, the volume, variety and velocity of the data generated will be overwhelming – yes, this is a big data problem – and the data engine needs to be able to process this data into meaningful information. It then needs to glean pertinent insights from this information and present it in a manner that enables easy consumption.

For example, customers often leave traces – such as declining login trends – that indicate impending churn. Sometimes, their behavior is predictive of future churn. For example, Groupon found that mobile customers that didnt login during the first month were 80% likely to churn within six months.

But its not all about churn. As I indicated in my previous post, the power of negative churn can make the difference between success and failure. In most cases, it is likely that the data will contain insights that indicate future upsell, cross-sell and referral opportunities. This could include sentiment expressed in emails and support tickets, product usage trends, activities on social media, and more.

“The greatest value of a picture is when it forces us to notice what we never expected to see.”
John Tukey

The Data Engine is critical to unearthing these powerful insights about customers that can ensure the long-term business success of your company.  There are four components that enable the Data Engine to turn raw customer data into valuable customer insight:

  • Cohort analysis – Identify common patterns in customer behavior to identify cohorts of customers that are likely to act in similar manner.
  • Cluster analysis – Create clusters of customers using data-driven partitioning techniques designed to create clusters based on relevant variables.
  • Sentiment analysis – Track facts and opinions to identify potential issues with customer experience with automated sentiment analysis.
  • Propensity analysis – Identify the propensity to act in predetermined ways, including the propensity to buy model tells you which customers are ready to make their purchase and the propensity to churn model tells you which active customers are at risk.

blog-dataengine-6blindmen2

From Insight to Ingenuity
For many years, we have known that customer data contains many a gem of insight.  For the first time in our history, technology has progressed to the point where we are able extract these insights and act upon them.  Some of this is already happening – predictive lead scoring and behavior-driven online advertising are two areas where data science has made huge leaps.  These insights will unleash the ingenuity that is inherent to the human mind, and propel us to the next level of performance.  And Customer Success Automation is the next frontier.

The vast majority of us don’t possess anything close to the genius that was Einstein.  But now, we sure have an opportunity to look like him.