We all want to build experiences around customer needs. To do that, we need to not only listen to what customers say, but also the intent driving their actions.
With a nearly infinite number of individual goals, how can we organize Intent signals into a practical and reliable framework?
We have developed such a Customer Intent taxonomy for Intent ONE, our data framework. It unifies millions of intent signals from searches, conversations, reviews, and first‑party data into a single, queryable view of customer intent. This taxonomy was proven effective across Customer, B2B, Financial Services, and Healthcare use cases. It remains a work in progress and will change as we help more organizations, so please be in touch with any feedback or suggestions.
Customer Intent Taxonomy
Observations are broken into multiple intent signals that are, in turn, classified into six broad categories. Each intent category has its own taxonomy and sub-classification approach. An observation can be a literal ‘utterance’, what someone says (email, chatbot, search, social post), or observable behaviors (user interaction, engagement, ad clickthrough, etc.)
Type of Customer Intent:
- Aspirational: The goal behind someone’s intent. They know the problem, not the solution. Example: “I want to sleep better.”
- Informational: Looking to understand a topic or get answers, not tied to a specific brand or product. Example: “How do heat pumps work?”
- Commercial: Making decision and comparing options. Example: “Best stroller for small cars” or “HubSpot vs Pipedrive.”
- Transactional: Ready to take an action or complete a task without needing an existing account or relationship. Example: “Purchase a product”, “Book an appointment.”
- Relational: Needs that involve an existing relationship with a brand. Example: “Help me import contacts,” “Refund a duplicate charge,” “Cancel subscription.”
- Testimonial: Expressing opinions, experiences, or feedback: “I love this mattress—highly recommend.”
Starting from a solid base
This classification model follows the Search Intent taxonomy, defined by Andrei Broder at AltaVista in 2002, which categorizes intent into Navigational, Informational, and Transactional (later expanded to include Commercial Intent) intents. While this classification works wonders in the field of Search and SEO, it fails to capture the totality of customer intent across their journey. We added the Aspirational, Testimonial, and Relational Intent to cover the whole customer life cycle.
The Navigational Intent category was removed, because it acts as a default value when there is insufficient context and doesn’t accurately represent intents applicable to organizations. We want to know the why behind an action, not the where. An Intent behind an action can be understood entirely differently depending on the context.
Built to break silos
This approach connects disparate sources in a way that has never been done before, rather than creating a classification model that only works for advertising, SEO, product, support, and so on. The goal is to consolidate all intent signals into a single, shared model across the organization for an accurate 360-degree view of the customer.
Biased towards the customer journey
Every classification is biased, but some biases are more useful than others. Classifying the totality of Human Intent is next to impossible. Customer Intent, however, follows patterns uncovered through years of research in behavioral economics and psychology. This top-level taxonomy closely matches the recognized customer journey phases that practitioners are familiar with.
While the model looks linear, customer behavior is far from it. People typically jump from one Intent to another back and forth, often simultaneously. This model considers each observation carries multiple intents.
Working on multiple levels
Let’s classify all intents into broad categories first, and refine them further as new data becomes available. A multilevel hierarchical taxonomy makes for a more resilient and reliable system. We can extract different types of data based on the class or reclassify Intent at a later step, such as during entity reconciliation. We also need to allow for tailoring for the specific needs of an organization. While the top intent classes are shared, further classification becomes more tailored to the organization.
Conclusion: A new foundation for connected experience
This model has been proven effective across various industries, including Customer, B2B, Financial Services, and Health organizations. A robust intent classification model helps an organization create a connected experience that works in the customer’s terms.
- It enables every team to focus on what people want to achieve
- Deliver an intelligent, responsive experience that meets people where they are
- Accelerate the pace of innovation
- Bring all customer data into one shareable view
By sharing findings, we can all benefit. This is an open discussion, so please feel free to reach out to share your feedback and learnings. Together, we can build the next generation of intelligent customer experience!
