The setting is technical; the argument is not
A full-text feed delivers the complete article inside the reader. A summary feed delivers a sentence or two and a link. The RSS specification accommodates both without preference; the <description> element accepts any length. What decides which a publisher chooses has nothing to do with XML and everything to do with traffic, attribution, and money.
The traffic-and-analytics argument for truncation is straightforward: if readers consume the piece inside their feed reader, the publisher's analytics platform never sees the visit. No pageview, no ad impression, no time-on-site signal, no retargeting cookie dropped. For a publisher whose revenue is tied to display advertising or whose editorial strategy depends on demonstrating audience scale to sponsors, a full-text feed is a literal drain on the numbers that pay for the operation. That is a real constraint, not a pretence.
The reader-experience cost is equally real. Subscribers who have chosen a feed reader as their primary interface have signalled a preference for a unified reading environment — one inbox, one scroll, no tab-switching. A summary feed forces them out of that environment for every article. Enough friction accumulated across a subscription list and readers stop opening the reader at all, or they stop subscribing to clipping-only sources. The feed becomes a notification service, which is a degraded version of what it could be.
Where platform defaults reveal the real logic
This is where Substack and Ghost diverge in ways that expose the decision for what it is. Substack publishes full-text RSS by default. The reasoning maps directly to its business model: Substack earns its 10% cut on paid subscriptions, not on advertising, so driving traffic back to a hosted web page produces no revenue for Substack and minimal advantage for writers on the free tier. A full-text feed is, for Substack's economics, cost-free generosity that earns goodwill with readers.
From the record
Platform defaults compared
- Substackfull-text RSS by default; no advertising revenue to protect; model is paid subscriptions
- Ghostfull-text RSS by default; self-hosted, publisher controls truncation if desired
- Legacy news publishers (e.g. pre-2020 NYT era)summary feeds by default; advertising and on-site traffic models drive the decision
The shifting stakes
- Pre-2013: Truncation debate centred on pageviews and ad impressions
- Mid-2010s: Some relaxation toward full-text as newsletters and subscriptions grew
- Post-2023: AI training crawlers introduce a new motive for withholding full content via feeds
Ghost, which the Ghost Foundation maintains as open-source infrastructure, also publishes full-text feeds by default — but from a different position. Ghost is a self-hosted tool where the publisher controls the stack entirely. Its default reflects an architectural philosophy: the feed is a first-class output, not an afterthought, and a publisher who wants to truncate can configure that themselves. The openness is structural.
Legacy media and large editorial operations have historically landed in the opposite default. The New York Times, for years the textbook example cited in feed-reader circles, famously limited its feeds to headlines and summary text. The reasoning was never hidden: the organisation's identity and revenue both depended on readers arriving at nytimes.com, where the full reading experience — and the advertising infrastructure around it — lived. That is not an irrational position; it is a coherent business decision stated honestly.
The AI scraping era adds a new variable
The truncation debate has a new dimension since large language model training became a mainstream concern. A full-text feed makes content straightforwardly machine-readable at scale; a summary feed plus a paywalled article is structurally more resistant to bulk ingestion, regardless of what robots.txt says. Some publishers who had relaxed toward full-text delivery in the mid-2010s are quietly reconsidering, not because the traffic argument has changed but because the opponent in the access calculus has shifted from the occasional ad-blocker user to crawler fleets ingesting at volume. Whether that reconsidering will manifest in feed policy or in other technical controls remains an open question — but it shows, again, that the technical setting is just the costume.
The feed specification will never answer the full-text question because the specification was never the right place to ask it. What the <description> element contains is a business decision; it always has been. Publishers who pretend otherwise are usually publishers who have already made their choice and would rather not defend it directly.
