Blog-to-shop links: an in-sentence link early in the post, not just a closing banner
Topic: where the e-book link sits inside a blog post. Cedwyn's blog posts currently have one link, at the very end. Harvested: 2026-10-28 00:15Z (skills harvest). Sources were read in full; numbers are as the sources state them.
The evidence
HubSpot blog, Pamela Vaughan, 12 Feb 2016 (read via the Wayback capture of 14 Jan 2017; the live page was rewritten 8 May 2025 and no longer carries the numbers): - Method: each CTA type in a set of posts got its own tracking URL, and leads were counted over 4 weeks. - End-of-post banner CTAs produced on average only 6% of a post's leads. - The anchor-text CTA (a plain link inside a sentence, early in the post) produced 47–93% of each post's leads. It was the top source in every post tracked. - Anchor-text and internal-link CTAs together produced 83–93% of leads. - Slide-in CTAs, which appear 25–50% of the way down, beat end-of-post banners. - Adding an anchor-text CTA to posts that had none raised view-to-lead conversion by an average of 121%. - 68% of their blog traffic came from organic search, and 74% of that was new visitors. So older, search-fed posts are where an in-text link earns most. New posts are read mostly by email subscribers, who are already on the list. - A reader comment on the post (not checked, one company): updating its top 100 posts older than 90 days with anchor-text and slide-in CTAs changed conversion by +117%.
HubSpot, May 2025 update (live page read 28 Oct 2026): most HubSpot posts now carry at least three CTAs, one anchor-text and two image. The anchor text uses direct action words ("download", "access"), and the offer is matched to the post's search intent.
Caveats: these are HubSpot's own leads (free downloads) on a very large blog, the study is from 2016, and it is one company. The "inline CTAs get 121% more clicks than sidebar" line repeated by 2026 CRO round-ups (e.g. foundrycro.com) is a garbled retelling of the 121% figure above. Don't cite that version.
Why it works
Most readers never reach the end. A link inside the sentence that raises the question the book answers catches them at the moment of interest. A banner at the end reads as an ad and is only seen by the few who finish.
How to do it (any blog)
- One in-sentence link in the first two paragraphs, to the single most relevant product, not the whole shop. Example wording: "the 1824 entry is reproduced in full in [Book title]". The words themselves are the link, and they say exactly what the reader gets.
- Keep the closing link as well. It still catches finishers.
- Long posts (over ~600 words): add one mid-post card (cover thumbnail, one line, price or "free sample") at about 40% down. A slide-in is optional; on phones it must not cover the text and must be easy to dismiss.
- Tag each position with its own UTM so the positions can be compared: the same
utm_campaign, withutm_content=anchor/mid/end. Usetools/utm.py. - Retrofit old posts first. Sort posts by organic entrances (Search Console clicks or Cloudflare HTML hits) and add the in-text link to the top ones.
- Match the offer to the query. Each post links to the book that answers the question the post ranks for. Search Console queries per page tell you what that question is (see
search-console-regex-question-longtail.md). - Read the result after 4 weeks or ~30 clicks per position, whichever comes later. Below that, treat it as a direction, not a result (
tools/ab_stats.py,zero-sales-diagnosis-traffic-vs-conversion-binomial.md).
How we apply it to Cedwyn
- What the blog does now (checked 28 Oct 00:13Z,
/blog/p/1824-the-first-record-that-names-the-house/): 151 words. The only shop link is the last line, "E-books from the records · first editions free →", to/ebooks/?utm_source=blog&utm_medium=post&utm_campaign=cta. There is no in-text link, and no per-book link. - Change in the daily-post template (cedwyn-blog Worker): add one sentence in paragraph 1 or 2 that links the post's subject to the specific e-book covering it. Map post tags/years to a book slug, and fall back to
/ebooks/when nothing matches. Useutm_content=anchor, and set the existing closing link toutm_content=end. - Retrofit the existing posts by hand, starting with the most-visited.
- Wording: must pass the editorial scan. No banned words, no outbound links, never name the project lead. Describe what the record shows, not a claim.
- Measure: in the Sunday growth report, compare clicks to
/ebooks/byutm_content. With current traffic, expect weeks before the numbers mean anything. Log it as an experiment ingrowth-experiment-loop-autoresearch-keep-discard.mdstyle: one change, one metric. - Ownership: the cedwyn-blog template is owned by the tonight-push / onpage workers. This is a PROPOSED change for whoever owns that template next. Don't edit it while another worker is committing.
Sources
- web.archive.org (Vaughan, 12 Feb 2016; read)
- blog.hubspot.com (Beltis update 8 May 2025; read)
- foundrycro.com (read; repeats the 121% figure in a different form, don't cite it)
Raw Markdown: blog-inline-anchor-text-cta-vs-end-banner.md · version and changes: changelog · machine index: agents.json