The Opposite Start
Map the consensus, invert its assumptions, and build from the strongest gap.
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 8
- Confidence
- 99%
The Opposite Start uses AI as a research and differentiation engine rather than a ghostwriter. First, it searches platforms such as X, Reddit, LinkedIn, and the open web for recent discussion of a defined topic. It then clusters recurring claims to expose the dominant narrative—the content red ocean. Instead of drafting from that consensus, the method identifies audience-relevant omissions and inverts the prevailing story through six lenses, including reframing the mechanism, surfacing tension or hidden costs, changing the category, and changing the story's hero. Candidate angles are compared, and the strongest defensible departure is expanded into a brief containing hooks, supporting arguments, counterarguments, evidence, stories, and closing lines. The creator then authors the content from that differentiated starting point.
Origin
Extracted from Marketing Against The Grain, where Kieran Flanagan demonstrated a Claude Code skill that researches current narratives, creates six inversions, recommends one angle, and generates a supporting brief.
Core principles
- 01Research the dominant narrative before generating an angle.
- 02Treat consensus as a map of where competition is concentrated.
- 03Invert a narrative through several lenses rather than accepting the first opposite.
- 04Prefer an underexplored angle that remains relevant to the intended audience.
- 05Use AI to prepare the brief while preserving human authorship.
How to run it
- 1
Define the assignment
Specify the subject, audience, and content context so the research can distinguish a relevant gap from novelty for its own sake.
Pro tip Narrow the audience enough that the system can ask what the dominant narrative misses specifically for them.
Watch out A vague topic produces generic clusters and weak inversions.
- 2
Map the current narrative
Search recent posts and articles across several platforms, then summarize the claims, benefits, benchmarks, and stories appearing repeatedly.
Pro tip Use a recent time window for launches or fast-changing subjects.
Watch out Do not treat one platform or one loud account as the whole consensus.
- 3
Cluster the consensus
Group similar claims into dominant narrative clusters so the crowded angles become explicit before ideation begins.
Pro tip Name each cluster as a concise claim rather than a broad topic.
Watch out Starting generation before this step allows the model to reproduce the same consensus invisibly.
- 4
Find audience-specific omissions
Ask what each cluster ignores, understates, or assumes about the intended audience. Turn those omissions into possible sources of tension.
Pro tip Look for lost authority, hidden costs, changed incentives, or an overlooked protagonist.
Watch out An omission is not valuable unless it matters to the audience.
- 5
Apply six inversion lenses
Flip the core mechanism, expose the real debate, reveal costs, recategorize the issue, construct a counter-position, and change who serves as the hero of the story.
Pro tip Generate at least one candidate per lens before judging any of them.
Watch out Do not stop at a literal opposite that lacks evidence or practical significance.
- 6
Select the strongest angle
Compare candidates for differentiation, audience relevance, evidentiary support, and a plausible path to truth. Recommend the angle with the best combined profile.
Pro tip Favor an angle that surprises immediately but becomes convincing when explained.
Watch out Novelty without defensibility becomes empty provocation.
- 7
Build the evidence brief
Develop hooks, pro arguments, counterarguments, statistics, illustrative stories, and closing lines around the chosen angle.
Pro tip Include the strongest objection so the eventual content can earn trust rather than merely assert a contrarian view.
Watch out Verify every statistic and source before publication.
- 8
Create in a human voice
Use the brief as intellectual scaffolding, then write or speak the final content using firsthand judgment, experience, and storytelling.
Pro tip Add examples only the creator could credibly tell.
Watch out Letting the model draft the finished piece can erase the differentiation gained during ideation.
In the wild
For content about GPT 5.5 for marketers, the system found that most commentary emphasized productivity, autonomy, benchmarks, and upgrades. It inverted that consensus into a governance argument: marketing leaders who fail to make their teams AI-native may lose authority over AI as budgets and decisions move to IT. It then supported the angle with C-suite statistics, hooks, counterarguments, a story, and closing lines.
→ A crowded product-launch topic became a differentiated episode brief about organizational power rather than model capabilities.
A leadership creator maps the usual remote-work arguments and finds that most focus on productivity versus office attendance. Applying a hidden-cost and changed-hero lens, she develops an angle about junior employees losing informal access to decision-making rather than workers losing productivity. She validates the claim with interviews and writes the article herself.
→ The creator enters a saturated debate with a specific, defensible, audience-relevant perspective.
Common mistakes
Prompting for the finished post first
A direct drafting prompt begins inside the model's consensus and often reproduces the same angle offered to other users. Map the narrative before asking for any content structure.
Choosing novelty without evidence
A surprising inversion is not automatically useful. Require relevance, supporting proof, and a plausible line to truth before selecting it.
Outsourcing the final voice
Using AI for both ideation and finished prose can turn a differentiated brief back into generic content. Preserve the creator's experience, judgment, and storytelling in the final work.
Is it for you?
Best for
It is best for marketers and creators addressing crowded, fast-moving subjects where conventional AI prompts yield interchangeable ideas.
Not ideal for
It is not ideal when factual reporting requires a neutral consensus account or when the creator lacks time to validate a contrarian premise.
From the transcript
“And that's why I call that the opposite start. So I don't start in the red ocean of ideas, I go all the way over…”
“I think that one of the core strengths of this skill is that it basically builds a cluster of what is the common narrative first,…”
“Actually, here's all of the things that current narrative misses for your audience, and then it clusters them into those categories and then says, This…”
From the episode
Use AI for Ideas, Not Content (Here’s How)