Signal-to-Activation Time
Why this concept exists. "Real-time" is nebulous. Sales conversations and marketing collateral use it interchangeably across orders of magnitude — from sub-100ms edge decisioning to 30-minute audience refreshes. Signal-to-Activation Time replaces the word "real-time" with a measurable axis (source.real-time-clarity-md).
Definition.
- Signal — an event implying, based on user context and historical patterns, that a particular response should be delivered.
- Activation — the marketing response (page change, email, push, SMS, POS update, etc.).
- Signal-to-Activation Time — the interval between signal timestamp and activation readiness.
The taxonomy. Discrete tiers, each with characteristic technical approaches and order-of-magnitude cost differences:
- latency-tier.intra-page (< 100 ms)
- latency-tier.inter-page (100 ms – 10 s)
- latency-tier.intra-session (10 s – 5 min)
- latency-tier.inter-session (5 min – 30 min)
- latency-tier.intra-day (30 min – 6 h)
- latency-tier.inter-day (6 h – 12 h)
- latency-tier.daily (12 h – 3 days)
- latency-tier.weekly (3 days – 7 days)
- latency-tier.temporal (calendar-driven)
Cost rule of thumb. Moving an activation up one tier (e.g. Intra-session → Inter-page) typically incurs an order-of-magnitude cost increase. Moving down imposes equivalent client-side or batch-side complexity.
Why this is the agent's primary framing tool. When a user says "we need real-time personalization," the agent's first job is to translate that into a tier. The architectural recommendation depends on it.