Claim
With web traffic to the open web down 46% in three years and ChatGPT CTR to external sites at 1.3% versus Google's 29.2%, the strategic choice is no longer whether to fight for clicks but whether to produce content made for the platform or the article. Platform-first content builds brand influence that accumulates across feeds, Reddit threads, and AI responses, even when it generates no measurable referral traffic.
Mechanism
Last-click attribution cannot see influence that accumulates off-site. A brand that shows up consistently in AI responses, Reddit discussions, and social feeds builds preference before the buyer ever visits a URL. That preference is invisible to standard analytics but it is what decides which vendor gets shortlisted. Platform-first content is designed to feed that process rather than to drive URL visits.
Conditions
Holds when: a buyer's consideration journey includes AI answers, social feeds, and community discussions before visiting any vendor site. Holds for most B2B and considered-purchase B2C categories.
Fails when: the category is transactional and URL-visit intent is strong (branded search with high purchase intent). Also fails when the product has no discoverable surface in platform conversations.
Evidence
SparkToro and Datos data: web traffic for the open web down 46% in three years. ChatGPT CTR to external sites: 1.3%. Google CTR: 29.2%. Fishkin frames brand influence that compounds across feeds, Reddit threads, and AI responses as invisible to last-click models but real in buyer decisions.
Signals
- Brand mentions in AI answers increase without a corresponding referral traffic increase.
- Reddit and community discussions reference the brand without linking to the domain.
- Customers cite the brand as something they had heard of without a traceable source.
Counter-evidence
The 46% traffic decline is an aggregate; categories and individual brands vary. Some domains grew traffic in the same period by capturing AI-adjacent queries. Platform-first content requires a longer feedback loop and is harder to justify to stakeholders working with last-click ROI models.
Cross-references
- Web traffic fell 46% in three years; clicks were always the wrong proxy for attention, and the fix is a measurement reframe, not a content overhaul (Rand Fishkin, measurement reframe)
- 58.5% of US Google searches end without a click, create standalone value where audiences already are (zero-click strategic framing)