Ever wondered how advanced analytics enrich customer segmentation? We asked Charlotte Tison Pierron-Perles, Big Data and Analytics Lead at Capgemini Consulting.
“Today, marketing segmentation needs to go beyond classic customer segmentations based on sociodemographic data. It should encompass a broader scope of information and provide a dynamic view of the customers, adapted to the company’s marketing needs. To achieve this goal, big data and analytics are not only buzzwords, they provide key innovative assets by enabling us to capture data from many different sources, work on various types of data, and provide dynamic results."
More Touchpoints Means More Data
The growing number of customer touchpoints and interactions is generating more and more information: CRM data, web logs, call center recordings, social media, email conversations, etc.
Big data technologies enable us to leverage these huge amounts of data by offering capacity to capture and manage:
- Internal information, by breaking silos between the different divisions of the company to leverage the whole internal assets of the organization beyond usual marketing data. There is great synergy to create through the joining of marketing data with datasets from others departments (e.g., sales, operations, customer service).
- External data, by aggregating information from different sources and ecosystems outside the organization, e.g., social networks, blogs, forums, open data. Strategic alliances with other companies can also be made to share data (through third-party specialists) with an eye toward uncovering common business opportunities. The data covers many different types of customer characteristics and usages, both explicit (e.g., information entered by a customer on a web form) and implicit (e.g., visited pages and time spent on a website).
Better Technology Means Better-Used Data
Analytics technologies, on their side, now offer the capacity to analyze these new amounts of data and provide relevant insights through:
-
Ability to leverage many formats of information that used to be complex or impossible to take into account — e.g., deep learning algorithms to understand the content of an image or video and the meaning of text data.
-
New algorithms for segmentation such as network mining and density-based clustering (as opposed to classic K-means).
-
Enhanced hardware and software performance on large amounts of data : segmentation jobs that used to run overnight in 10 hours can be optimized to run in five minutes.
-
Capacity to compute dynamic segmentations in real time, thanks to new technological solutions, e.g., Spark Streaming and Flink.
Get Real
Data and analytics are often used to run one-shot studies (proof of concepts) in an exploratory mindset. It is now possible to have as many segmentations as business questions to answer. But their real value lies in the industrialization of dynamic segmentations updated automatically in real time with a link to both marketing and IT. This can be done by integrating segmentation into the service catalog offered by a company’s data lab in addition to its exploratory role.