Recognize: How to move from unknown to (varying degrees of) known
Imagine a customer registering across different channels such as mobile, web, and tablet with different identities. Add anonymous browsing behavior to this mix and we are presented with the first challenge of identifying the customer without invading their right to privacy. While traditionally name and address matching solutions have helped, the next step in the evolution is to build identity solutions based on interactions.
It is also important to understand the subtle objectives of marketing and analytics functions. Marketing will always communicate with a self-identified customer, whereas analytics wants an inferred customer identity to derive metrics such as life-time value (LTV). It is important to provide a mechanism to link the requirements of both these functions through identity solutions. Once a customer identity is established, the data can be enriched through third-party data providers such as Harte Hanks’ Global DataView.
Do you truly understand your customer?
Once you overcome the recognition stage, it is important to collate and curate every customer interaction. This could be each mouse click on a website or periodic interaction on a mobile app. technologies makes it easier and more cost-effective to store massive data sets and run artificial intelligence and machine learning algorithms at scale. With artificial intelligence and machine learning (AI/ML) technologies, we can distill billions of interactions to a few meaningful ones that matter, make educated estimates on customer intent specific to a context and moment in time, substantiate the intent through a specific treatment plan, and continually validate through a test and learn framework.
How do you effectively engage your customer?
Marketers have struggled to answer the question, “What is the true value and return on investment of content?” Personalization goes beyond getting a customer’s name right on a mass mailer. It should also help determine what media and content a customer wants to view. It is important to view customer engagement levels over time in addition to traditional outcome measures such as revenue and conversion to determine value of content. These metrics could include the number of visits and average time spent on a site visit. A higher engagement score has typically translated to a more profitable customer.
At Wipro, we have adopted a man-plus-machine approach for content personalization at scale. The creative element still needs human intervention but once the base template is created, we leverage our cognitive and AI platform to create variations of the base content template to effectively engage with the customer during a moment in time interaction.
Taking a longitudinal customer view to nurture customer relationships
While the traditional focus on nurturing customers has been on preventing attrition, it is important to take a longitudinal view of the customer journey across their lifetime. For example, look at every interaction from the time a customer anonymously browsed to when the customer registered, to specific customer responses and transactions from campaigns. Similar to the engage phase, it is important to look at lead indicators rather than just outcome metrics such as LTV and average basket size.
Deriving a customer’s need state at different points in the lifecycle will help build more robust personalization and drive better targeted solutions. There is no point in diluting margins by sending offers or promotions to a customer who values convenience over price.
With the shift toward digital interactions, we are presented with a unique opportunity to understand what customers want and more importantly what they don’t want. As events in customers’ lives can change their preferences over time, it is important to take a long-term view and treat personalization as a journey with monetization as an outcome. With the processes outlined in this paper, I hope you are able to jumpstart this journey