The blogs made the conversation.
The announcement made the list.

1The verdict

The cross-check across three launches gives one clean answer, and it is not about writing quality.

The split is not writing quality and it is not the channel. It is proximity to the offer, and whether the article reaches that launch's own opt-in page. Both halves repeat across three launches, three lists and three products. Open rate cannot tell them apart, and neither can revenue, because the educational sends carry the back catalogue and still bank $19,000 to $44,000 each.

2Naming a send, so the number can be traced

Every row in this report carries the exact Klaviyo campaign name, the campaign id and the send date. Not “the video”, not “BOGO”, not the date alone.

The reason is not style. Klaviyo campaign names are not unique, so a name on its own does not identify a send either. Three different Black Seed Oil sends carry the same name, and five Nattokinase sends are all called “Email 1”. Only the pair of name plus id is safe to quote.

The Nattokinase arc is the worst case: its five production sends are named “Email 1” in Klaviyo. They are identified here by send date and subject line (May 15 “A message for you”, May 19 “This is for you”, May 22 “Worth knowing”, May 26 “Be prepared”, May 29 “Thank you”). Renaming them is the cheapest fix in this report.

3How each launch is wired

No campaign links to an opt-in page. Every one links to a blog, and the blog carries the opt-in link. That second hop is where the launches differ, and it is read off the live articles.

Nattokinase 4-in-1 Cardio Complex · May 15 to Jun 8, 2026

Ethiopian Black Seed Oil · Jul 15 to Aug 21, 2026

Ancestral Protein Coffee · Sep 23, 2026, still running

4The cross-check

How to run it, and what it says. Four steps, all from data that already exists.

1. List the sends by exact name and id. GET /api/campaigns/ filtered to sent, with send_time. Never trust the name alone.
2. Find the blog each send links to. The body lives on the linked template, not on the campaign message, so follow campaign-messages → relationships.template and pull GET /api/templates/{id}/, then grep its HTML for /blogs/.
3. Find the opt-in page that blog links to. Pull the article from the Shopify Admin API and grep its body for /pages/. This is the hop that breaks.
4. Count sign-ups on that page on that day. For a Klaviyo-form page, sweep the Submitted Form metric and filter on page_url. For the coffee page, which posts directly, the form event does not fire at all, so count list joins by joined_group_at instead.

#4 Protein Coffee - Product Announcement Blog sent at 11:50 ET today and is still accruing, so its open rate, revenue and opt-in count are all partial. Its 237 opt-ins are counted from list joins, because the hand-coded form on that page does not fire the Submitted Form event that the other rows use.

5Comments and opt-ins move in opposite directions

The Black Seed Oil window, day by day, with the campaign that sent each day.

The two pure blog sends drew 73 and 85 comments and 23 and 13 opt-ins. #3 Black Seed Oil - Blog on Jul 30, the video, and #3 Black Seed Oil - Blog on Aug 6, the product announcement, drew no comments at all and 272 and 292 opt-ins.

6What the comments actually say

Comment volume is not the signal. Composition is. The share asking how to buy ranged from 0% to 43%.

The Microclot Scandal is the clearest case. It drew the fewest comments of the three nattokinase blogs (31), the highest ask-to-buy rate (19%) and the highest revenue of the three ($41,570). What Dr. Sumi Found drew three times the comments and less money.

Every article in the corpus, ranked by ask-to-buy

Threads are capped at 50 top-level comments by the comment system, so totals are exact and sampled composition is not. The cap shows in the Sampled column.

7Who actually opted in

Measured on the 845 Black Seed Oil submissions, the only completed launch with surviving event data.

A founding page is not acquisition. It is a way for the list you already have to tell you who intends to buy. Worth measuring, never worth reporting as list growth.

8What we cannot measure, and why

9The KPI set

Seven to add, one to keep, one to demote. Each computable from data that already exists.

10What to fix before Thursday

The coffee founding page opens Thursday Oct 8. Four changes, in priority order.

1. Point the remaining coffee blogs at the coffee page. Two of the three still reach the generic page and file to VdWm2S “On-Page SMS Subscribers”. That is a one-line edit per article.
2. Build the launch text. No flow triggers on RunR7E “Ancestral Coffee Launch SMS Opt-In”. Everyone who signed up today receives nothing.
3. Rename the campaigns that share a name, starting with the five called “Email 1”. A report cannot cite a send that cannot be identified.
4. Snapshot list membership weekly into the repo with joined_group_at. One call, and it survives the 98-day event window that erased Nattokinase.