Relevant first. Measured. Kept exactly.
Research completed 2026-10-27T15:31:23.598Z. Ten primary posts were inspected before the eight new pieces were written. Seven are LinkedIn posts; three are Substack articles.
Sabrina Ramonov’s supplied method governs the workflow: choose a topic, study successful examples, keep the selected opening line unchanged, and remix the substance with the author’s own perspective. The 70/20/10 word split is Jet’s additional instruction.
The selected headline is the source post’s complete opening line. It is the first copied line in each new draft, including punctuation and emoji. No separate invented headline replaces it. The card headline and the copied opening both use the source headline verbatim. The headline appears once in the copied text.
Why these four openings
Sheila
Direct fit to a service-business buyer awaiting a reply; 1,408 reactions and 266 comments exceed the measured customer-support comparison (80 and 34). Use the text-caption opening exactly, never the video anecdote or its earnings/conversion claims.
Body mix: Timely first reply + tangible consequences of waiting + continuity of customer context.
Primary source. Alternatives: •Shep Hyken · hyken-repeat; Sabrina Ramonov · sabrina-roi.
Egelyn
The strongest observed calendar/owner-dependence example in this sample: 2,248 reactions and 984 comments. Its opening contains no borrowed biography or results. It fits the owner-supplied limited-time admin context.
Body mix: Owner dependence + one unsupported absence exposing a handoff gap + removal of unnecessary repeated work.
Primary source. Alternatives: Ilse Noppen · noppen-delegation; Sabrina Ramonov · sabrina-stages.
Angelo
For delegated decision authority, Braun (121 reactions, 152 comments) exceeds Noppen (15 and 15). Hormozi has larger counters but focuses on check-in speed, a different management problem, and is not used to replace the better-fitting authority hook.
Body mix: Learned escalation + actual decision rights + supported development of specialists.
Primary source. Alternatives: Ilse Noppen · noppen-delegation; Alex Hormozi · hormozi-checkins.
Jed
A directly relevant measured LinkedIn AI/marketing example with 887 reactions and 401 comments. Explicitly an evergreen 2023 source. The newer Substack articles guide substance; their counters are not combined with LinkedIn or treated as a global ranking.
Body mix: AI output versus buyer relevance + assumption checking + a response path and actual funnel evidence.
Primary source. Alternatives: Sabrina Ramonov · sabrina-multiply; Sabrina Ramonov · sabrina-stages; Sabrina Ramonov · sabrina-roi.
The measured sample
Topic relevance and truthfulness first; compare observed engagement within the same platform and relevant topic. Reactions and comments remain separate. No combined cross-platform score or causal hook score. “Strongest relevant” means an editorial choice within this stated comparison. It is not a global #1 claim, a verified organic-reach ranking, or proof that the hook caused engagement.
| Source | Published | Reactions / likes | Comments | Reposts / restacks | Observed UTC |
|---|---|---|---|---|---|
| Justin Welsh · welsh-calendar | 2026-09-23 | 2248 | 984 | Unknown | 2026-10-27T15:27:45.644Z |
| Alex Hormozi · hormozi-speed | 2026-04-07 | 1408 | 266 | Unknown | 2026-10-27T15:27:44.989Z |
| Shirley Braun , Ph.D., PCC · braun-delegation | 2026-09-01 | 121 | 152 | Unknown | 2026-10-27T15:27:45.355Z |
| Justin Welsh · welsh-ai | 2023-01-20 | 887 | 401 | Unknown | 2026-10-27T15:27:45.630Z |
| Alex Hormozi · hormozi-checkins | 2026-08-19 | 2574 | 298 | Unknown | 2026-10-27T15:30:03.312Z |
| Ilse Noppen · noppen-delegation | 2026-10-01 | 15 | 15 | Unknown | 2026-10-27T15:30:03.075Z |
| •Shep Hyken · hyken-repeat | 2021-08-25 | 80 | 34 | Unknown | 2026-10-27T15:30:03.320Z |
| Sabrina Ramonov · sabrina-stages Substack | 2026-10-03 | 572 | 40 | 54 | 2026-10-27T15:30:02.460Z |
| Sabrina Ramonov · sabrina-roi Substack | 2026-09-19 | 122 | 7 | 6 | 2026-10-27T15:30:02.361Z |
| Sabrina Ramonov · sabrina-multiply Substack | 2026-09-26 | 130 | 5 | 5 | 2026-10-27T15:30:02.810Z |
Checks on the new drafts
- 8 of 8 copied first lines match their primary source headline character for character.
- 4 LinkedIn posts: 140 problem words, 40 solution words, 20 call-to-action words.
- 4 Substack newsletters: 280 problem words, 80 solution words, 40 call-to-action words.
- The title/opening line is counted once, inside the problem. Counts use whitespace-separated tokens.
- Original bodies use illustrative situations and owner-supplied voice context; no borrowed client results, earnings or conversion rates.
Limits that matter
- Purposive sample, not an exhaustive platform ranking.
- Seven main-post LinkedIn counters were read from fresh HTTP 200 primary responses; Substack header counters were independently reopened this turn. Counts can change.
- Unknown impressions, reader retention, exposure denominators and organic/paid distribution; engagement rate, paid-ad status and hook causality cannot be established.
- Comments can include low-value or automated responses. A visible counter does not prove qualified buyer attention.
- Two selected sources are older than 90 days; the AI-marketing reference is from January 2023. None is labeled currently trending.
- A video post's written caption is not its spoken video opener. The Sheila draft preserves the written caption opening for a text post; it is not represented as a reproduction of the video's first seconds.
- Cross-platform transfer is an editorial experiment; LinkedIn source metrics do not establish future Substack performance.
- No Zuporta publication or performance outcome has been observed.
Higher counts did not automatically win
- Inspected reference: 1,183 likes are higher than other inspected Substack sources, but an AI design-demo opening is a poor fit for these four operating problems. High attention alone does not make a transferable hook.
- Inspected reference: Relevant missed-call topic, but no usable main-post total exposed. Not selected as a proven attention example.
- Inspected reference: Profile/discovery counts were inconsistent with later primary post responses. Only stable main-post URLs and their own counters govern the selected evidence.
The team collection is publicly accessible by link. Private source records and raw snapshots are not included in this site. Selecting, copying or downloading does not establish social publication or measured results.