Before I trust a marketing trend, I ask who benefits from it.
Not a topic. A standing move, applied to any metric, platform claim, or industry consensus — including my own past ones.
Three things, rarely together
Most people in marketing thought leadership have one of these. The combination is what makes the method more than an opinion.
Trained to read discourse critically
A sociology background focused on media, audience behavior, and public discourse — including a thesis on how narratives get constructed and whose voice gets amplified or erased.
Builds the systems, not just evaluates them
Hands-on development of the AI systems in question — not commissioning a vendor, not reviewing a pitch deck. Direct exposure to what actually happens inside the build.
Accountable inside real governance
Shipping ideas across legal review, brand compliance, and multi-market stakeholders — so any critique has to survive contact with reality, not stay academic.
This method only stays credible if it's sometimes turned on my own past decisions — not only on industry narratives or competitors. If every application concludes "the establishment is wrong, I called it," it stops being a method and becomes a shtick. At minimum one self-critique stays in rotation for every 2–3 external-facing applications.
Claims run through the method — verdicts included
A running, honest record. Not everything survives contact, and that's the point — a method that always confirms the interesting conclusion isn't a method.
Reddit's citation share proves it's the platform to chase for GEO right now.
SurvivesCitation share is a snapshot of one company's retrieval engineering, not durable authority. Flagged the instability before Reddit's citation share in AI search results fell sharply — a predictive call, not a reaction.
"Cost per unit of intelligence" is the right way to judge AI ROI in marketing.
SurvivesMeasures efficient consumption of a vendor's product, not business outcome. What actually correlated with results when shipping an AI comms system inside enterprise governance was the governance layer, not spend efficiency.
Paid social is the default channel for driving conversions — including for enterprise events.
Self-critiqueMy own belief, not an industry claim I disagreed with. Confidently invested based on the standard modern-marketing playbook; couldn't trace a single conversion when asked directly. Shifted to ABM — conversions became traceable immediately. The self-critique that anchors the whole method.
Cutting human content moderation for automation is a clean efficiency win.
Open questionNo lived operator proof point yet on this one — running as a genuine question to the field rather than a confident claim: what's the escalation path when it's not a normal day? Will update this entry if and when there's a real example behind it.
AI shopping agents choosing on structured data over brand means brand no longer matters.
Not earned yetGenuinely interesting reframe in progress — brand may not be dying, it may be moving up a level to platform-gatekeeper trust. Held back from publishing until there's a concrete example to anchor it, per the method's own standard.