The pattern that captures behavioral events from a server-side endpoint as the primary source, with client-side capture as the fallback for events the server can't observe directly. Resilient to ad blockers, browser tracking limitations, and consent state changes that affect client-side firing. Increasingly the default rather than the upgrade path, though many implementations are still in transition.
The pattern that extracts intent from text, voice, and other unstructured customer interactions: search queries, support tickets, chat transcripts, product reviews. Increasingly important as customer behavior moves into conversational interfaces and away from clickstream events. Requires NLP or LLM processing to convert unstructured input into structured signals downstream recipes can use.
The pattern that captures different signals for different consent states without breaking the whole pipeline when someone withdraws consent. Defines which events fire under no-consent, partial-consent, and full-consent states, and how each state degrades the available recipes. Required under GDPR and increasingly under CCPA, CPRA, and the patchwork of US state privacy laws.
The pattern that detects absence: customer disengagement, the pause before churn, the things that don't fire an event but should have. Models for retention, win-back, and predictive churn depend on negative signals more than positive ones, but most stacks are built around positive event firing and treat absence as a query problem rather than a capture problem.
The pattern that combines disparate intent signals (content engagement, product activity, third-party intent feeds, sales conversations) into a single composite score the rest of the stack can decision against. Aggregation rules carry the editorial judgement: which signals count, how recency is weighted, whether velocity matters more than volume, whether account-level signals override individual-level. The signals are loud and noisy; the aggregation is where the meaning lives.