From Jonathan Hart's published writing at Relatively Objective.
- Agentic AI
- Autonomous systems capable of using tools and reasoning to achieve a complex goal — without step-by-step human instruction. The foundation of Harbinger's decision layer.
- Sequential Association Rule Mining
- A data science technique used to predict the next likely action in a customer's journey based on previous behaviors. Invented and operationalized by Jonathan Hart.
- Transferable Demand
- The measure of how much a customer's preference for one product feature can be satisfied by a different one. Core to Harbinger's Shelf Science module.
- Value Stream Mapping (VSM)
- A lean methodology adapted for data flows — identifying where information loses value between collection and execution. Harbinger's Leak Detector visualizes this in real time.
- Evolutionary Algorithms
- A class of optimization algorithms that evolve toward better solutions over successive generations — without requiring manual retraining. The basis for Harbinger's Neural Optimization module.
- P.I.P.E. Score
- Harbinger's proprietary prioritization framework: Purpose, Importance, Potential, Ease. A four-dimensional scoring system for determining which projects belong in your immediate roadmap.
- Live Tribes
- Harbinger's term for dynamically formed consumer micro-segments. Unlike static personas, Live Tribes update in real time as behavioral data evolves — no human assumption required.
- Kaleidoscope
- Harbinger's self-service data extraction and dashboarding platform. Pulls in owned, earned, and physical-world data streams and merges them into a unified science-ready dataset.