The Datasource described on this page is an illustrative example. The actual fields, structure, and naming will vary based on your brand’s specific data model and integration. Use this as a reference for the type of data needed, not as an exact specification.
Overview
The Navigation Datasource captures real-time browsing behaviour on the brand’s website and mobile app: products viewed, categories explored, time spent, search queries, and cart events. This provides intent signals that significantly improve recommendation relevance — what someone browsed yesterday is a strong predictor of what they want today.
Key Fields
| Field | Type | Description | Example values |
|---|
customer_id | String | Customer identifier (matched via cookie/login) | "C-2847391" |
session_id | String | Browsing session identifier | "SES-a3f2b1c9" |
viewed_products | Array | Products viewed in this session | ["SKU-12345", "SKU-67890"] |
viewed_categories | Array | Categories explored | ["Shoes > Sneakers", "Bags"] |
time_on_site | Integer | Session duration in seconds | 420 |
search_queries | Array | Search terms used | ["white sneaker size 40"] |
cart_items | Array | Products currently in cart | ["SKU-12345"] |
last_visit_date | Date | Most recent site/app visit | 2024-11-14 |
Connected Use Cases
| Use Case | How Navigation is used |
|---|
| Enriched Product Recommendation | Real-time intent signals for product selection |
| Drive to Store / Web / App | Browsing intent as store-visit trigger |
| 1:1 Personalised Digest | ”Still interested in…” block based on browsing |
Data Requirements
| Requirement | Details |
|---|
| Update frequency | Real-time or near-real-time (within hours) |
| Minimum fields | customer_id, viewed_products, last_visit_date |
| Format | JSON via Datasources API or JavaScript tracker |
| Volume | One record per session or aggregated per customer |
Integration Notes
Navigation data is typically collected via:
- JavaScript tracker on the website (similar to analytics tools)
- Mobile SDK for in-app behaviour
- Server-side events from the e-commerce platform
Navigation data has a short shelf life — browsing from 7 days ago is far less relevant than yesterday’s session. Workflows should prioritise recent navigation signals (last 48 hours) over older data.
