Decisions
144 rules
Twelve areas, twelve rules each. Every rule is an IF / THEN / ELSE with a primary metric, a guardrail and the LAKA levels it operates at.
49 of 144 rules · clear filters
- BIZ-001 Define value before volume IF the program has no defined qualified business event Baseline
- BIZ-002 Traffic is diagnostic, not the north star IF organic traffic increases AND qualified conversions or margin do not Baseline
- BIZ-003 Separate forecasts from actuals IF search volume, CTR, conversion, or revenue is estimated Baseline
- BIZ-006 Count assisted value carefully IF an informational asset rarely receives last-click conversions but appears in qualified journeys Baseline
- BIZ-007 Split branded and non-branded demand IF brand familiarity materially changes intent, CTR, or conversion Baseline
- BIZ-010 Resolve competing goals explicitly IF traffic, lead quality, margin, brand, retention, and learning objectives conflict Baseline
- BIZ-011 Allocate by marginal value IF multiple valid initiatives compete for limited resources Baseline
- BIZ-012 Use economic stop rules IF expected remaining incremental value is below expected cost OR risk exceeds tolerance Baseline
- AUD-001 Require an audience-task pair IF a candidate query cannot be mapped to a plausible audience and task Baseline
- AUD-002 Treat funnel modifiers as heuristics IF a word such as how, best, review, price, or buy suggests a journey stage Baseline
- AUD-011 Explicitly exclude wrong audiences IF queries attract jobs, students, freebie seekers, unsupported geographies, or unrelated senses Baseline
- SEM-001 Start from parent concepts IF an offer lacks a mapped semantic territory Baseline
- SEM-002 Expand through subtypes IF a parent concept contains meaningful narrower types with distinct tasks Baseline
- SEM-003 Normalize synonyms and paraphrases IF different phrases express the same sense and dominant task Baseline
- SEM-004 Disambiguate homonyms and polysemy IF the same wording can refer to different concepts, entities, or tasks Baseline
- SEM-005 Model modifiers as facets IF terms vary by price, material, size, audience, platform, location, time, or other attributes Baseline
- SEM-009 Triangulate demand evidence IF a candidate appears in only one tool or source Baseline
- SEM-010 Do not equate zero tool volume with zero demand IF a query has low or zero reported volume but appears in first-party or customer evidence Baseline
- CLU-001 One dominant intent, one owner IF a dominant intent cluster is approved Baseline
- CLU-002 Merge same-task candidates IF sense, task, answer structure, result class, and next action are compatible Baseline
- CLU-003 Split different tasks IF queries share vocabulary but require different completion conditions or document forms Baseline
- CLU-004 Use result overlap as evidence, not law IF current top results overlap strongly for two queries Baseline
- CLU-005 Detect cannibalization by behavior IF multiple owned URLs receive the same cluster’s impressions with unstable dominance or divided value Baseline
- CLU-008 Inspect equity before deletion IF a page appears useless or redundant Baseline
- CLU-012 Test uncertain cluster boundaries IF merge/split evidence is inconclusive Baseline