I pity the marketing organization. It has to deal with a lot of change.
First it was the pesky empowered consumer. Then it was understanding device proliferation and cross-media consumption. As if that wasn’t enough, what about social, mobile and digital disruption?
Then of course there’s the (ubiquitous) marketing ROI question. Let’s not forget about using marketing automation technology. And now, Big Data analytics. (Not necessarily in this order).
A little too much to handle, don’t you think? No wonder you find the stats about CMO tenures pop up once in a while and the comings and goings of senior marketing executives make the news so frequently.
Big Data analytics is no different from the rest. As with any of these trends, the marketing community first starts with myth busting, then does some show-me-the-money analysis and finally looks for use cases on how the trend can be practically applied in the marketing and customer experience context. Here the focus is not to get into a debate of what Big Data analytics is or isn’t but rather to understand how analytics in the age of Big Data impacts customer experience and marketing processes.
So how does marketing become bigger, better, stronger and faster using analytics?
There is no dearth of customer data now available for marketers to digest. Analyzing all of that data — at scale — is a challenge for any organization. Every customer interaction is a chance to know her better, serve her relevant experiences, and build the groundwork for the next interaction. Analytics technologies now have the ability to handle bigger datasets without compromising on speed or accuracy of results.
For example, customer journey analytics or path analysis involves data points from multiple interactions and touch points. Path analysis, traditionally performed through sequence mining, found its early application in the website visitor funnel context. Now the entire customer journey can be analyzed using path analysis to understand what series of events led to a desired outcome.
Another technique called social network analysis (SNA) uncovers relationships between entities or customers in a large network with the goal of identifying influencing nodes of customers. This type of analytics utilizes large amounts of network data with underlying graph database structures. For example, in the telecom industry, SNA is used to identify customers who are influential in a call network with the goal of preempting attrition.
The use of predictive analytics to recommend products, offers and experiences has piqued the interest of marketers in the last couple of years. Marketers want to be able to make better predictions in an attempt to drive in-depth relationships or simply get more efficient with their targeting.
Next-best offer (NBO) analytics estimates the probability that customers will be interested in a targeted offer of products and services they are likely to buy. NBO involves a set of rules and algorithms to get better at message and offer relevance. Recommendation engines in combination with NBO can help marketers increase average order size by recommending complementary products based on predictive analysis for cross-selling purposes.
While direct marketers have relied on response modeling for a long time to ensure that their campaigns are targeted with precision, now techniques like true-lift modeling or uplift modeling are available to them to stop spending direct marketing dollars on customers who would purchase anyway.
Analytics without sound experimentation is as good as a ship without a captain to steer it. Stronger analytical results and insights thrive on a robust experimentation framework, whether it is simple hypothesis testing or a complex fractional factorial design. Conducting experiments at scale in a short amount of time while still producing statistically significant results can make for analytically stronger decisions. For example, on-the-fly and continuous champion/challenger testing of offers and content on websites can make for stronger analytics results and click-through rates.
Another way to strengthen analytically-derived outcomes in the marketing and customer experience context is to infuse new data streams such as social data, semantic data and text data into traditional analysis methods such as attrition and churn analysis. This will uncover drivers of customer attrition that may not be apparent from pure, structured transactional data.
Any Big Data analytics conversation would be remiss without the mention of real-time. Whether real-time analytics manifests itself in customer-facing processes or back-end data flows, the time-starved consumer demands the same from the marketer. Marketers don’t have the luxury of a wait-and-see approach to analytics; instead they have to adopt self-learning and living analytical processes to continuously pump the marketing engine with fresh and new customer insights. Learning can be achieved in the modeling process or through intelligent decision arbitration systems.
Many of the use cases discussed here are emerging, but they’re becoming highly relevant in the Big Data context. Organizations at different stages of their analytical journey are considering how to add some of these capabilities to their analytical toolkit today. At the end of the day, despite the constant upheaval the marketing organization continues to face, Big Data analytics promises to make marketing bigger, better, stronger and faster.