In the first part of this blog, I had been looking at marketing data housed in different data silos as one of the key pain points for marketers who want to establish a powerful customer segmentation practice. In this continuation of that theme, I would like to cover why it is so difficult to derive insights from the data once you hold it in your hands. The nature of this data can be a nightmare for marketers across the globe.
Firstly, marketing data is huge in volume. Marketing is among those capabilities within an organization that can really know Big Data when they see it – and can tell how it feels to deal with it on a daily basis. It is also one of the top concerns of CMOs of organizations of all types trying to figure out how to leverage customer information in order to generate superior insights and use it as an asset to drive business results. This Big Data challenge that we see in Marketing is also the reason why there is so much attention paid to the plight of the Global Marketing organization from both vendors and analysts alike.
So, with Big Data in your hands, you will almost certainly run into performance issues. To get around it, IT and business often have to compromise on which data to use and at what level of granularity. In customer segmentation, many organizations are actually working with predefined reports and aggregated data. But that of course is not what I, as a marketer, would want. What I would like is to slice and dice the data and have a high degree of flexibility in how I do this. Sometimes I need all customers who’ve had > x amount of revenue in a specific product category over the last two years; sometimes I need to drill down to the details of a specific purchase. Yes, I really do want to have the full transparency on customer interactions to understand them better! And I’ll need the result back really fast no matter how “big” Big Data gets. What I want is to “play” with data and not schedule queries to run overnight (that is so yesterday!) or have to ask IT to create an additional view or cut of the data. I want to do this on my own, I’d like to become an empowered marketer!
Naturally, more and more data is unstructured – Social Media is, in fact, among the driving forces for the explosion of “Big Data”. In order to make logical sense out of tweets, wallposts, reviews, blogs and others, I’ll need to understand what is being said via the social channel, and text mining capabilities can listen for and group sentiment around conversations. As an example, if there is any negative sentiment about my brand, I certainly want to list everyone who has the terms/phrases in question posted to various social channels. I may also want to combine this with structured information – like the Klout score or the number of followers of an individual / influencer, so that I can focus on the those whose influence really have an impact on my customers.
Another notion of Big Data is current vs. historic data. A customer has a memory like an elephant. The customer sums up all of his customer experiences into a perception – so in any situation, I’ll need to know what interaction history that customer is taking into account when building that sum. I’ll need to be able to look at not only the current situation, but also trends based on historic data. Is this an “A” customer that just turned into a “B” customer – or the other way around? If so, how come? Are there avenues to explore growing this customer strategically? Historic data is also the key to predictive analytics, awkward as it may sound – I need the interaction history to predict future behavior.
Finally, having all of this data available in a timely manner is critically important. You need these insights and understandings at your fingertips – when an event happens and where that event occurs. Often that truly means real-time. For instance, if I as a marketer am preparing to launch my cross-selling campaign and before that launch, one of the customers within my target group orders the product I am about to promote I’d prefer to avoid the cost of the campaign and the risk of puzzling a customer by promoting a recently purchased item.
In the next part of the “Segment Like Never Before” series, I’ll be highlighting the relationship between IT and marketing in the context of customer segmentation. If you have any comments or experience to share on customer segmentation, I’d be really pleased to hear more from you!