Personalization is at the heart of today’s consumer brand companies. It defines the experience of the brand in the eyes of customers and prospects and drives the underlying economics of the business. And personalization at scale is not just a phenomenon outside the brand; it describes a fundamental shift in how operations happen inside the brand. The value of the brand is measured not just in profit and loss, but in reach and engagement, customer retention and loyalty — driving every action in the business through the lens of customer lifetime value.
In order to fuel transformation, brands need a collection of capabilities forming an engine for customer data that can power their entire enterprise — from analytics to customer support to finance to marketing to compliance. One way to think about the customer data engine for a given brand is “every system across the organization that touches customer data.” This broad view allows a holistic look at the ecosystem of connected infrastructure required to deliver the promise of personalization. From talking with leading brands, there are seven common steps to building a successful customer data system.
This diagram shows the seven steps in the middle. Read on for a detailed explanation of each.
Step 1 - Collect All The Data
The first step in a customer data engine is getting access to all the data about the customer across the brand. This includes all the first-party data available across all channels, online and offline, including the full historical dataset. It may include data licensed from third-party providers or other data collected from resellers or partners. This is the foundation for a complete customer view, and needs to be up-to-date with the latest changes in customer profile information, transactions, behaviors and interactions.
Having this data centrally collected allows for flexibility and iteration that adapts to the changing needs of a business. Typically this requires a broad set of connectors to different systems — including real time streams, API-based connections, and the ability to ingest large volumes of raw historical data. This data should be available in its full form and not require complex ETL (extract-transform-load) to be ingested. This is because, in order to build a complete customer view, it is important to have access to all of the information about customers, not just a sub-set. This collection of data represents the raw ingredients necessary to power your customer data engine.
Step 2 - Resolve Customer Identity
The second step is to take that raw data and identify the building blocks used to understand the customer, the first of which is customer identity. This is the linking key that unlocks the full view of a customer, and it must be refreshed every day so that it’s up-to-date and accurate. This key allows the rest of your customer data to be attached to the right customer — so transactions, visits, support experiences, and interactions can be accurately assigned to the same person. It is critical that this phase include the ability to identify and match customers using only first-party data owned by the brand, particularly given the changes in the third-party data ecosystem driven by increased privacy regulation and changes in how browsers and phones manage cookies and mobile advertising IDs.
The identity resolution step is foundational to everything that happens downstream — if identity is inaccurate or not up-to-date it means that every other calculation and action will suffer from an incomplete view of the customer.
Step 3 - Build the Customer Profile
Once the data is linked together, the next step is to build a complete customer profile. This profile should provide the single view of the customer that powers all parts of the brand — from analytics to customer support to marketing to finance. Because the business changes all the time, customer data changes as well — which means that flexibility is a must. It should be easy to add new attributes to the profile, add in new data sources about the customer, and create brand- or role-specific views of the information that meet the changing needs of the brand.
Change management and change control is crucial, including the ability to validate changes before they are used in production, since changes can have a material impact on the quality of analytics and marketing campaigns that power the business. Self-service capability is also critical, as the brand needs to be able to respond quickly and cannot be locked into a fixed view.
Step 4 - Generate Insights & Predictions
Next it’s time to put the data to work. This capability is made up of two key pieces: what insights you derive from customer data at face value and what machine learning and AI can predict about customer activity based on historical data. Insights should be generated at the customer and segment-level, including brand and channel behaviors, product preferences, revenue sizing, and recommended actions.
Because personalization is a business driver, it’s also important to visualize and monitor customer-centric metrics and KPIs that highlight shifts in customer economic indicators, identifying risks and opportunities. Additionally, predictive models can help marketers identify which segments, personas, or audiences have the highest growth opportunity, affinity for specific products, or churn propensity, all of which can be applied to improve campaign ROI and customer lifetime value.