Context Window:
Everything in your current conversation Claude's working memory for that session. More context = better output. Context resets when you start a new conversation.
Hallucination:
When Claude confidently states something false. Most common with specific numbers, dates, and citations. Rule: verify any specific factual claim before using it.
Prompt Chaining:
Using one output as the input to the next prompt. The basis of multi-step workflows.
RICE Framework:
Reach × Impact × Confidence ÷ Effort. A prioritization scoring tool. Useful for structuring the conversation; not a substitute for judgment.
Acceptance Criteria:
The specific, testable conditions that define when a user story is complete. Written in GIVEN/WHEN/THEN format. If a criterion can't be tested objectively, it's not a criterion.
Guardrail Metric:
A secondary metric you monitor during an experiment to ensure your primary improvement isn't causing harm elsewhere.
North Star Metric:
The single metric that best represents the core value your product delivers to users. When this moves, the product is working. When it doesn't, something is wrong regardless of other metrics.
Opportunity Sizing:
Structured estimation of how large a problem is before committing to solving it. The goal is distinguishing between meaningfully different magnitudes, not precision.
Activation:
The moment a new user first experiences the core value of your product. The most actionable early-stage growth lever.