Claim
The tone, intelligence, and values of the content you publish select the audience you get. Negative content breeds negative audiences; thoughtful content attracts thoughtful people. The audience you build is downstream of the editorial choices, not orthogonal to them.
Mechanism
Each post is a self-selection filter: it rewards the readers who resonate with that voice and quietly drives away the ones who don't. Iterated over hundreds of pieces, the comment section, share patterns, and inbound advertiser inquiries calibrate to the same band as the content. Brand-safety and advertiser-fit are emergent properties of editorial discipline.
Conditions
Holds when:
- The content output is consistent and high-volume enough for the selection effect to compound (Kyra was 50 people producing daily across PAQ, Bad Canteen, Greatness).
- The publisher cares about advertiser sentiment ratios, not just raw reach.
Fails when:
- The platform algorithm overrides editorial choice (e.g. ad-driven recommendation that pushes engagement bait regardless of what you publish).
Evidence
"If the content you publish is negative, you will breed an audience fuelled by negativity. Haters in the comments. Low sentiment ratios. If your content is positive, uplifting, inspirational, aspirational, you’re likely to receive the same response from the audience."
— Devran Karaca, What I learnt in the first 12 months of starting a media company, LinkedIn Pulse, 2018-02-26
Signals
- Comment-section sentiment ratios converge to the editorial tone within months.
- Advertiser conversations get easier (or harder) in lockstep with the audience's perceived quality.
- Publishers who chase virality see audience drift; publishers who hold the editorial line see audience fit.
Counter-evidence
Algorithmic distribution can decouple audience from editorial intent (e.g. a thoughtful post that gets pushed to a hostile cohort by recommendation systems). Karaca's claim assumes a publisher building its own subscriber base, not riding pure recommendation.
Cross-references
- (none in current corpus)