Having a contextual understanding of a customer often refers to having situational awareness of a customer journey. It is about leveraging data readiness to compile and analyze enough signals about a customer to understand real-world relationships – what the customer cares about, the customer’s place in a household or business, how the customer interacts with a brand.
In addition to real-world, situational context that comes from recognizing and understanding customer signals, another aspect of context is the context of the data itself. Contextual metadata – the data universe context – is just as important for truly understanding and engaging with customers with a relevant customer experience (CX). Contextual metadata helps marketers and data engineers understand, trust and responsibly use the data.
Here, we’ll break down what each means, why they matter, and how they work together to power smarter marketing.
Situational Context: Understand the Person Behind the Data
Situational context is gleaned from collecting and analyzing signals about a customer and a customer’s relationships with other people, with brands, products, and even to themselves over time. These signals reveal intent, affinity, and preferences. Examples include social media activity (likes, reviews, etc.), loyalty program participation, purchase history, store visits, search behavior and subscriptions.
All of this real-world, situational context helps marketers deliver a relevant CX. Putting signals together yields a deep, personal customer understanding that marketers use to deliver a next-best action, to understand whether the customer is more likely to churn or convert, to know the preferred communication channel, etc.
Customer context is dynamic and situational – it changes as the customer’s behavior and environment evolve.
Data Context: Understand the Data Behind the Customer
Metadata provides another layer of context to customer data, providing an understanding of data as it relates to other data. This data context enables data engineers to better understand what matters for marketers, which helps them build and deliver data products for downstream business users
This additional data context helps evaluate whether data is ready for business use, i.e, complete, accurate, timely, actionable, trusted, and compliant. It also helps determine whether data is semantically clear.
Metadata that provide data context include elements such as data lineage, data quality, timelines, meaning, compliance and usage – anything that helps us understand the data itself. For instance, whether data feeds are complete, missing, or delayed. Or whether there is data drift, i.e., unexpected changes in volume, accuracy or timeliness. Data may become more reliable or less reliable; either way, this type of contextual understanding about the underlying data is a key element of data trustworthiness – which underlines data readiness.
Some elements of data context may not even relate directly to the customer. For instance, context may include how much it costs to acquire a piece of data. Or how often the data is uploading, an important metric for calculating the cost to keep the data reliable.
Semantic understanding means providing additional detail about data so that people, applications, and AI can understand meaning, constraints, and usage for that data. This may include textual definitions, allowable values or rules, data types and relationships. For example, semantics provides a common understanding for the value of a number – does the number “5” refer to a size, an amount, a unit of currency, etc.? If the number denotes a purchase amount, a comparison can be made between purchases in dollars and euros provided a known exchange rate.
The data context provided by metadata lets data engineers and data agents do their job reliably and consistently. But this context also lets business users understand the meaning, relationship, and usage of data for better insights and CX.
Why You Need Both: The Power of Combined Context
There is ample crossover between situational context and data context that underscores the importance of both. For instance, imagine trying to predict churn based on email engagement – only to find out the customer opted out of email months ago. That’s a situational signal that changes how you interpret the data.
On the one hand, you have a situational signal about a customer (channel preference). On the other, that data point is important context about the data source itself, letting you know that you can’t make conclusions or calculations based on how often that customer interacts with emails. Likewise, if a significant percentage of customers have opted out of email, perhaps you don’t calculate a churn prediction based on email interactions.
In this and similar examples, the real-world, situational context for the customer (channel preference) is tied back to an element of the data universe. Conversely, there are many elements in the data universe context that tie back to situational context. For instance, a data engineer training AI models needs to make sure the model understands the meaning of individual data elements, e.g., this number represents a shirt size, this number represents a purchase in dollars. These and other semantic understandings are crucial for the AI model to make the correct recommendations or predictions for both direct and indirect interactions with customers.
What Does This All Mean?
Marketers well understand the value of capturing situational signals that reflect customer behavior and intent. Doing so is foundational to providing a relevant CX. Making this context clear for data engineers and developers lets them create visualizations, APIs, applications, and models that handle the whole customer context.
Capturing metadata and developing a deep understanding of data as it relates to other data is just as important for extracting the optimal value from data. No longer just an issue for IT and data engineers, it is key to ensuring data is trustworthy, timely and well-understood.
The Redpoint Data Readiness Hub bridges the gap by providing both types of context, helping marketers see top-level signals without having to dig into raw data. It automates the translation of context into actionable insights, validating recommendations and predictions through the delivery of accurate, relevant experiences.
The reason context is king is because the best marketing decisions come from understanding both the customer – and the data that represents them.
