2026 Astaxanthin Review – Rovina
The two chiefs
House chief · 7/10
doctor-signed ranking advertorial on the seller's own store · cold-to-warm paid traffic · buyers who tried astaxanthin and felt nothing · solution-aware, some product-aware · three tests (dose, source, delivery) resting on fat-solubility and a four-to-six-week build-up
Lead: Problem-solution lead on the reader's failed purchase. Grade B-plus: five short paragraphs, proof numbers folded into the reader's memory, the turn in nine words.
Offer: Two-for-one stated as a percentage, a daily cost, a 90-day guarantee, a rating and a customer count. No price in dollars, no reason for the discount, no bonus. The guarantee carries the offer.
Close: The strongest part: an offer block, a reasoned guarantee for the reader who did not tap, a callback close, a last button. The urgency is the reader's own lost months, and no deadline is invented.
Throughline: Letdown, vindication, growing confidence, relief of risk. It holds from the first line to the last and dips only in the research block.
Strongest
- the winner drawn in negative space: rivals' shortfalls reveal its features before its card (entry 18)
- rules that each carry a reason and a jab, preceded by their single cause (entries 12-13)
- a real limit admitted in the winner's card and turned into the case for the guarantee (entry 30)
- the guarantee argued as evidence about the seller (entry 43)
- the testimonial as the lead's buyer after the fix, placed before the offer (entry 39)
Weakest
- authority that cannot be checked: a doctor with no institution (entries 1, 6)
- scores and statistics with no method or source, two of them in conflict (entries 17, 28, 42)
- two rivals behind a headline about the year's top-rated brands, and 50 formulas claimed
- no disclosure and no disclaimer anywhere on the page
- one anonymous testimonial for 100,000 customers (entry 39)
Craft chief · REVISE
The first fifth sounds like someone who has stood where the reader stands; the cards and the research block sound like a label read aloud.
Strongest
- entry 13: three rules that each carry a reason and a jab, all open
- entry 18: the winner described through what the runner-up lacks, before the winner is shown
- entry 30: a real limit admitted and turned into the case for the guarantee, in 27 words
- entry 43: the guarantee argued as a bet, the best reasoning on the page
- entries 9-10: the replay and the turn, tighter than on the older ranking
Weakest
- entry 17: an orphan statistic that contradicts the page's premise
- entry 28: a column of perfect scores with no method
- entry 45: a leftover phrase, and no disclosure anywhere
- entries 34-35: science with nothing named
001 PASS·noted · good rhythm in entries 9-10 and 13; entries 34 and 38 run on parallel sentences of equal length. · 002 PASS · a complete movement, and the close gets real space: offer, button, trust line, a reasoned guarantee and a last paragraph (entries 40-45). · 003 PASS · plain words; the one Latin species name is glossed by its context (entry 13). · 004 PASS · the winner's name is an open loop held to 44% (entry 24); each rival card ends on a button; a hinge question opens the list (entry 15). · 005 PASS · the headline's upset is paid: two famous bottles are shown losing, and its verb describes what the page does. · 006 PASS·noted · an orphan statistic in entry 17 and five perfect scores in entry 28. For the guard: 100,000 customers, 50 formulas, 70 trials, 3,500 papers, 89 percent. · 007 PASS·noted · the capture shows the deck highlighted and the kicker in italics, both where a voice would lean; one button in capitals (entry 27). · 008 PASS·noted · the only-one claim is restated five times (entries 26, 33, 36, 38, 44). Cost: attention in the last third. · 009 PASS·noted · the narrator's account (entry 33) is a short proof anecdote with an expectation and a reversal; if the doctor is invented, so is the anecdote. For the guard. · 010 PASS · spoken register in the lead and in the trade-offs bracket (entry 30). · 011 PASS·noted · offer and buttons are in text; no picture has alt text; the hero repeats the byline's face (entries 1, 6). · 012 REVISE · entry 45 ships a phrase from a skincare template, and the page carries no disclosure or supplement disclaimer. Cost: trust, and exposure with the ad platforms. · 013 PASS · entries 8-10 are the reader's own history in the reader's terms. · 014 PASS · a compact arc: letdown, vindication, confidence, with the testimonial (entry 39) as the payoff. · 015 REVISE · em dashes in nearly every paragraph; setup frames at entries 10 and 33; a mirrored pair closes entry 43. Tier 1 in copy someone else wrote: flagged, not rewritten. The mirrored pair is earned. · 016 PASS · nothing on the first screen needs prior knowledge. · 017 PASS · emotion (the wasted bottle), curiosity (which product came ahead of the famous bottles) and stakes (money and months) are all present by entry 9. · 018 PASS · colour: the jab inside the first rule (entry 13); the winner described through its rivals' shortfalls (entry 18); the two bets (entry 43). · 019 PASS·noted · the first-person passage (entry 33) is smooth everywhere, with no odd phrase or rough beat; no corpus can exist for a persona. · 020 REVISE · about a fifth of this page's words recur on the store's category ranking (winner card furniture, research block, offer, guarantee, close furniture), and its lead and close recur from the older ranking. Cost: a reader who meets two of the three sees a template. · 021 N/A · not an email sequence. · 022 PASS·noted · the chant in entry 44 and the mirrored bets in entry 43 would be flagged by a mechanical pass and should stay. · 023 PASS·noted · the first screen names a ranking and shows the bag in the hero; the deal never appears in money terms, only as a percentage at 87%. · 024 N/A · no verb hands the product a decision the buyer owns. · 025 PASS·noted · 50 formulas claimed and three shown; a review score of five against 4.4 stars (entries 28, 42); value scored five here and four on the sibling page for the same daily price. Notes for the guard.
Take
- Draw the winner in negative space: each rival card has a block of shortfalls measured against first place before first place is shown (entries 18, 23), so the reader knows the winner's features before its name.
- Silver, bronze, gold: the strongest rival first with real praise, the weakest second, the winner last and seen against the weakest (entries 16, 21, 26). CS-P1 is winner-first; this is the variant to test.
- A button to the winner under every rival card (entries 20, 24), and the name revealed first in a button.
- One cause, then the rules: the single fact about the molecule (entry 12) before three rules that each carry a reason and a jab (entry 13).
- Admit the real limit inside the winner's card and turn it into the reason for the guarantee (entry 30).
- The testimonial that mirrors the lead's failed buyer, placed directly before the offer (entries 39, 40).
- A reasoned guarantee section after the button, for the reader who did not tap (entry 43). Ours needs its own reason for 30 days (r7, b6).
- The treat-or-cure disclaimer spoken in body type at the point of doubt (entry 37).
- An early jump button on a winner-last page (entry 7).
- The close that returns to the reader's failed bottle from the first paragraph (entry 44).
Leave
- FABRICATION RISK: the doctor persona and everything that rests on it: byline, hero picture, three quotations, the first-person passage (entries 1, 6, 16, 21, 26, 33). CL-14.
- FABRICATIONS: letter grades and a perfect score table with no method (entries 16, 21, 28); 50 formulas tested with three shown (entries 1, 11).
- The orphan first-day statistic (entry 17).
- The 90-against-30-day argument in its present form (entries 18, 30, 43): our guarantee is 30 days, and the argument is aimed at brands like ours.
- Carrier-oil comparison and absorption superiority (entries 13, 18, 29): banned for us (b8).
- Crowd and review numbers with no source and in conflict with each other (entries 5, 28, 42).
- A page with no advertiser disclosure and no supplement disclaimer (entry 45).
- Five different button wordings (CL-16).
- An offer with no price in dollars and no reason for the discount (entry 40).
- Blocks cloned across the store's rankings: a fifth of this page recurs on the category ranking.
- Pictures with no alt text (entries 6, 14, 25, 32, 35).
Line by line · 45 entries
Every line is described in our words; the page's own text stays on the page (open it beside this table). Shaded rows carry the most pull.
| # | Beat | What the line does | Lever · reader | Why here | Craft | On our page |
|---|---|---|---|---|---|---|
| 1 0.2% | authority-stamp 2× h2 · 17 w | The page opens on an author block: a doctor's name with MD, the specialty gastroenterologist, a research-team credit for more than 50 formulas tested, and an update date in May 2026. | borrowed-authorityspecificity solution-aware | Authority comes first, above the headline, because everything below asks the reader to accept grades on trust. A gut specialist is a pointed choice for a page whose argument is about what is absorbed and what passes through. | FABRICATION RISK: a name, a degree and a specialty, with no hospital, licence number, link or signature. Treat the doctor as invented until shown otherwise. The capture shows a round portrait of a woman in a white coat with a stethoscope. The block credits a team as well, so two authors are claimed at once. The store's older ranking condemns a rival for this kind of persona. | rewrite-claim CS-P1 has an editor byline with a generated portrait and a press strip, and no doctor. CL-14 forbids an invented authority. r19 allows a longevity doctor's recommendation with a quote once the doctor is real or confirmed by David (BRAND.md lists a medical advisor persona). Until then the honest stamp is an editorial credit plus what was done: five products bought, labels read, prices worked out per day. r19 |
| 2 1.1% | deck quote · 17 w | A quoted kicker in a grey box above the headline says that even in 2026 most astaxanthin on sale is not made in a way that gets it into the cells. | borrowed-authoritycommon-enemycuriosity-gap solution-aware | It sits between the doctor's name and the headline, so the first claim the reader meets is heard as the doctor speaking. Naming the year makes an old product category feel like news. | Seventeen words, italic in the capture, one dash. No speaker is printed under it; the portrait above lends it a voice. The idea of getting into the cells is the page's refrain and is first sounded here. | rewrite-claim CS-P1 has no kicker above its H1. The idea is ours already: most of what people take never gets into the cell (p3, r5), and about 64% of unverified astaxanthin fails potency tests (c1). If used, it belongs in the intro, in the editor's voice, and not as a quotation from a person. p3r5c1 |
| 3 1.9% | hook-headline h2 · 14 w | The headline says the year's top-rated astaxanthin brands were compared and that a single 12 mg product came out ahead of the best sellers. | curiosity-gapcontrastspecificitypattern-interrupt solution-aware | The reader has probably bought or considered a best seller. Saying the best sellers lost puts the reader's default choice in doubt, and the lead then confirms the doubt from the reader's own experience. | Fourteen words. 'We' replaces the older page's 'I', and comparing replaces testing, a verb the body can support. Quotation marks round the last two words carry a sneer. The dose in the headline selects the winner's spec in advance. | keep-structure CS-P1's H1 names the organs (kidneys, brain, eyes, joints) and promises five ranked products; there is no upset in it. For a retargeting variant the upset could be lab-tested batches beating famous names, resting on c1, c2 and c3, with the organs kept explicit (CL-04). c1c2c3 |
| 4 3.0% | deck h2 · 19 w | The deck says each product was graded on three points that settle whether the ingredient does its job or leaves the body unused. | specificityopen-looploss-aversion solution-aware | It stands under the headline so the upset has a method attached before the reader can doubt it. The three things stay unnamed until the lead has replayed the failed bottle. | Nineteen words with a dash before the consequence; highlighted in the capture. Its closing image repeats the discovery line of the older ranking. | keep-structure CS-P1's subhead has the same count structure (five softgels, three tests, one winner) and adds the organs. It has the move. |
| 5 3.9% | proof-social fine · 9 w | A small coloured line claims more than 100,000 customers have moved to the top pick. | social-proofconsensus solution-aware | It sits under the deck and above the picture, inside the first screen, so the ranking's winner has a following before it has a name. The verb implies those buyers left the bottles the reader owns. | Nine words. A round number with a plus sign and no source. The winner is still unnamed, so the crowd is attached to a position and not to a product. | swap-proof CS-P1 shows its count in the sidebar only (2,847 verified customers, the four-digit form CL-11 asks for), and its comment thread is placeholder text marked as a launch blocker. A count in the first screen needs a real number from the store. r17 covers panel statistics once our panel has run. r17 |
| 6 4.1% | image img | The hero picture, with no alt text. In the capture the woman from the byline sits at a consulting desk under a poster of the digestive tract, three packs in front of her, the seller's bag nearest the reader. | borrowed-authorityidentification solution-aware | It comes straight after the crowd claim, so the first screen ends on a face. Three packs tell the reader how short the list is, and the bag's place in front hints at the result. | The room and the portrait have the over-clean look of a generated picture, and the face matches the byline photo. The gut poster ties the specialty to the absorption argument. | drop CS-P1's hero is our bag alone. No doctor picture without a real doctor (CL-14). A lineup of the five packs would do the second half of this job. |
| 7 5.6% | cta cta · 4 w | A full-width dark button offers to show the top-ranked product; it jumps down the page to the winner's card. | easecuriosity-gap solution-aware | It stands at 6% of the page, before a word of body copy. It costs nothing to offer: whoever taps lands on the card with the scores and a buy button, and whoever does not has been told a winner exists. | Four words and an arrow. It names a rank and not a product, so the suspense holds. | keep-structure CS-P1 has no early jump button; its first button comes after the winner's card. Our winner is the first card, so no jump is needed. If the order changes to winner-last, add one. |
| 8 6.1% | failed-solutions p · 5 w | A five-word opener assumes the reader has already taken astaxanthin. | mirroringidentification solution-aware | First line of the body. The page is written for the buyer who is about to give up on the ingredient and says so at once; everyone else reads on as an onlooker or leaves. | Short, second person, softened by one adverb. It risks losing the reader who has never tried the ingredient. | rewrite-claim CS-P1 opens on the reader who was told to take a multivitamin (r3), which fits p7: most of our market has never heard of astaxanthin. This opener suits only a retargeting or lookalike audience. p7r3 |
| 9 6.3% | failed-solutions p · 38 w | The reader's history is replayed: read about a very strong antioxidant, 6,000 times vitamin C with over 70 studies; bought some; two months of taking it; little to notice. | mirroringspecificityloss-aversionborrowed-authority solution-aware | The proof numbers are placed inside the replay, as things the reader once read, so they are granted without being argued. The letdown comes last in the paragraph, where the eye stops. | One paragraph that speeds up: a long sentence of credentials, then three fragments and a trailing ellipsis. Tighter than the older ranking, which needed two paragraphs for the same job. | keep-structure CS-P1 states the 6,000 figure (m3) as a fact in its intro and has no replay. For our senior reader the replay is years of vitamins with nothing to show (r3, r23). r14 gives our study count of 60+. m3r14r3r23 |
| 10 7.7% | villain-reveal 2× p · 15 w | A one-line setup promises the part that goes unsaid. The next line clears the ingredient and blames most products for failing to deliver it. | vindicationcontrastcommon-enemycuriosity-gap solution-aware | It comes directly after the letdown. The reader is asking whether the ingredient is hype; the page answers no before the doubt hardens, and turns brand choice into the only open question. | A setup line ending on a colon, then two sentences of three and six words. The setup is a throat-clearing frame; the payoff is clean and quotable. The share of failing products has grown from some, on the older page, to most. | rewrite-claim CS-P1 has the fact in its method box (c1) and no turn. Ours: the bottle was never verified; about 64% of unverified products fail third-party potency tests (c1), and ours is third-party lab tested every batch (c3). Never blame what she already takes (p1 and p2 stay banned). c1c3 |
| 11 8.8% | turning-point p · 37 w | The team says it compared over 50 formulas with the clinical research and found that three factors decide the outcome, and that nearly every product fails one or more. | earned-authorityspecificityopen-loopconsensus solution-aware | It bridges blame to criteria. The reader now needs a way to tell bottles apart, and the claim that almost all fail raises the odds that the reader's own bottle did. | One sentence of 37 words, the longest in the lead, with a dash before the sting. The 50-plus figure repeats the byline; three products are ever shown. | keep-structure CS-P1 has this in its second intro paragraph: five products bought, labels read, ranked on three things. Its sample is small and true. Our failure rate is c1. c1 |
| 12 10.7% | mechanism h2 · p · 49 w | A heading about what buyers learn after three wasted bottles, then the reason behind the tests: the compound is fat-soluble, so the amount, the origin and the delivery must all be right or too little gets to the cells. | reason-whyloss-aversionnew-mechanism solution-aware | The reason comes before the list of tests. A reader who accepts one fact about the molecule will accept three rules derived from it. The heading prices the lesson at three bottles, more than the reader has lost so far. | The heading runs on a colon into the paragraph. Three clauses repeat one adverb, a drumbeat. The closing phrase about a costly dud is carried over from the older ranking's deck. | swap-mechanism On our page this slot is The Cell Shield: fat-soluble, so it settles into the cell membrane and spans it (m2), where vitamin C and vitamin E cover only part of the cell (p3). CS-P1's intro has it in two sentences. r31 lets us say take it with a meal that has some fat. No claim that one carrier oil absorbs better (b8). m2p3r31r5 |
| 13 13.0% | ranking-criteria 4× p · 102 w | An intro line, then three numbered rules, each with a reason and a jab: the full 12 mg against bottles of 4 to 6 mg, natural microalgae against a synthetic petrochemical form, an oil softgel against dry powder. | reason-whycontrastspecificitycommitment solution-aware | Criteria come before any brand, so the reader agrees while nothing is at stake. Each rule carries its own enemy, which means a rival card later needs only one line to fail a product. | The best-built block on the page. Each rule runs label, dash, claim, reason, jab, in about 30 words. The jab in the first rule is the freshest phrase: a dose large enough for the label and too small for the result. The explanations are open, where the older ranking hid them behind toggles. | rewrite-claim CS-P1's method box has five open rules with reasons. Ours rest on b7 (12 mg against the 4 to 6 mg brands), c4 and r16 (natural microalgae against petrochemical synthetic), c1 and c3 (lab tested) and b9 (one softgel a day). Their third rule is a delivery-oil argument and stays out (b8). CS-P1 lacks the jab inside each rule. b7c4r16c1c3b9 |
| 14 17.2% | image img | A three-by-two picture with no alt text follows the three rules. Its content was not captured; by position it most likely illustrates them. | ease solution-aware | It stands after the criteria and before the hinge question, where a scanner who skipped the text can still pick up the three rules. | No alt text, so the picture says nothing to a screen reader or on a slow connection. | keep-structure CS-P1 has no picture for its method box. A plain three-row graphic of our tests would serve the scanner, with alt text written as customer copy. |
| 15 20.3% | ranking-criteria 2× h2 · video · 9 w | A question asks which brands pass all three rules. A short wide video strip sits beside it, and a two-word title opens the rankings. | open-loopcuriosity-gap solution-aware | At 20% the rules are agreed and the reader is ready to judge. The question makes the list the answer to something the reader now wants to know. | Seven words in the question. The video has no caption or alt text, and its content was not captured. | keep-structure CS-P1 goes from the method box straight into the first card with no hinge line. One question would do. |
| 16 21.4% | ranked-entry 3× h2 · p · img · 41 w | Silver medal, second place: Double Wood, nicknamed for its Amazon best-seller status, grade A-minus. A quotation over the doctor's name calls it the nearest thing to a real rival and sound on the fundamentals. A square product picture follows. | candorborrowed-authoritycontrast solution-aware | The list starts at second place and not at the bottom. Generous praise for the rival the reader is most likely to own builds the trust that the later markdowns spend. The winner is held back. | The nickname reduces the brand to where it is sold. The quotation opens by giving credit, in a spoken register. A high grade for the runner-up makes the grading look honest; no method is shown. | keep-structure CS-P1 opens each rival card with a tagline that credits and then qualifies, in the editor's voice, with three score bars. It has the move, and its facts were read on the rivals' own pages (CL-13). It has no quotation from a named person and needs none. |
| 17 26.5% | ranked-entry h2 · 4× p · 22 w | Four plus points for the runner-up: the full 12 mg, a line about working from the first day for 89% of users, an olive-oil softgel, third-party testing. | candor solution-aware | The concessions come before the shortfalls, so the shortfalls read as fine distinctions between two good products. That is the only ground the winner can claim against a rival that passes every rule. | ORPHAN STATISTIC: the first-day figure has no source and contradicts the page's own premise that the compound needs four to six weeks. It looks like a leftover from another product's template. | keep-structure CS-P1 lists three checked plus points for each rival. Keep ours. |
| 18 28.5% | ranked-entry h2 · 3× p · 84 w | Three shortfalls measured against the still-unnamed winner: olive oil where the winner uses organic coconut oil, an Amazon-first buying experience with thinner support, and a guarantee shorter than the winner's 90 days. | contrastopen-loopreason-whycandor solution-aware | The winner has not been shown, yet each line reveals one of its features. By the time its card arrives the reader knows its oil, its sales channel and its guarantee, and has been told why each matters. | The heading measures against first place and not against the rules. The oil point admits the rival's choice is fine, which is honest and leaves the point empty. The guarantee point carries a reason (the weeks needed to build up) and is the only one with force. | rewrite-claim CS-P1 marks rivals down against the tests, and its winner is already known. The part that transfers is a reason in every rival cell (house law P-13). Ours: every batch third-party lab tested (c3), one softgel a day (b9), sold by the maker only (r29). Not transferable: the oil comparison (b8) and the point about guarantee length, since ours is 30 days. c3b9r29 |
| 19 32.3% | ranked-entry p · 40 w | The verdict calls the runner-up first-rate and correctly dosed, the one to buy if the winner were sold out, and puts the gap down to oil, buying experience and the time allowed to judge. | candorcontrastscarcity solution-aware | It ends the card on praise, so the button below, which leads to the winner, follows a fair-minded sentence. The sold-out remark is the page's only hint of scarcity. | Forty words in three sentences, with no insult. The generosity is the persuasion. | keep-structure CS-P1 ends each rival on a short verdict tag. It has the move in compressed form. |
| 20 33.8% | cta 2× cta · 5 w | A button under the rival's card leads to the winner's product page, still naming it only by rank. | curiosity-gapease solution-aware | It stands at 34%, where a reader who owns the rival has just been told it is second best. No card on this page links to the product it reviews. | Four words and an arrow, split over two lines of markup. The winner's name is withheld once more. | keep-structure CS-P1's rival cards carry no button; the sticky bar and sidebar cover it. A plain link to our top pick under each rival card is a cheap test. |
| 21 34.5% | ranked-entry 3× h2 · p · img · 36 w | Bronze medal, third place: Doctor's Best with the AstaReal source, nicknamed for its 6 mg dose, grade C-plus. The quotation credits a trusted name and source and says the catch is printed on the label. A square picture follows. | contrastborrowed-authoritycuriosity-gap solution-aware | The weak card is second in reading order, straight before the winner, so the winner is seen against a C-plus. The nickname does the markdown before the reader has read a word. | The quotation is two short sentences and ends on a small open loop about the label. Naming the ingredient supplier shows knowledge of the category. | keep-structure CS-P1 ranks only 12 mg products (CL-13, same species), so it has no underdosed card. b7 covers the class. b7 |
| 22 39.7% | ranked-entry h2 · 3× p · 12 w | Three plus points: natural microalgae from the AstaReal supplier, an oil softgel, and a reputable brand that is easy to find. | candor solution-aware | The page concedes source and delivery so the single failure, the dose, stands alone and unmissable in the next block. | Ten words across the three points and no glosses; the card is in a hurry. | keep-structure CS-P1 lists three checked plus points for each rival. Keep ours. |
| 23 41.0% | ranked-entry h2 · 2× p · 67 w | One shortfall: 6 mg a serving is half the research dose, and taking two doubles the cost against a single 12 mg softgel. The verdict: clean, trustworthy, and an introductory dose sold at the price of a whole one. | specificitycontrastloss-aversionanchoring solution-aware | These are the last words before the winner's card. The reader arrives at first place having just done a sum in which one softgel beats two. | The arithmetic is simple and checkable. The closing phrase of the verdict is among the best lines on the page. | rewrite-claim CS-P1 has the equivalent sum, aimed at this very seller: four softgels a day to reach 12 mg and a bag that lasts 15 days, read from its product page on 2026-09-14. Ours is one softgel (b9, b7). b9b7 |
| 24 44.0% | cta cta · 6 w | The first button to name the brand, with the 90-day guarantee in its label. | risk-reversalloop-closed product-aware | It stands at 44%, between the third-place verdict and the winner's medal. The name appears first in a button, so the reveal and the ask are the same line. | Four words, a dash and an arrow. A different label from the other button wordings on the page. | swap-offer Ours carries the noun and the 30-Day Money-Back Guarantee (CL-09), in one button vocabulary (CL-16). |
| 25 44.7% | image img | A small square picture with no alt text stands ahead of the winner's heading. Its content was not captured; by size and slot it is most likely a seal or award badge. | anchoring product-aware | It separates the two rival cards from the winner's card, so the reader sees a boundary before the medal. | No alt text. If it is an award seal, nothing on the page says who gave the award. | drop CS-P1 marks its winner with a ribbon in text. No seal that implies an outside award. |
| 26 46.0% | ranked-entry 4× h2 · p · img · 44 w | Gold medal, first place: the seller's Sunglow 12 mg softgel, grade A-plus. The quotation over the doctor's name says it alone met all three rules, in the amount, origin and form that the studies used. A product picture follows. | loop-closedexclusivityborrowed-authority product-aware | The winner comes last of three, at 46%. The reader has met its features in the rivals' shortfalls and has just seen a C-plus, so the card confirms what the reader has already worked out. | The quotation restates the rules in order. FABRICATION: grade and quotation rest on the unverified doctor. The product name has gained a word since the older ranking. | keep-structure CS-P1 puts the winner first with a ribbon and a callout that explains the top spot. Winner-last is the variant to test: rivals first, then The Cell Shield as the only one passing all three tests (b7, c3). b7c3 |
| 27 50.9% | cta cta · 3 w | A three-word button in capitals asks the reader to try the product now. | easeurgency product-aware | It stands directly under the winner's picture and above the scores. A reader who came by the jump button at entry 7 meets a buy button before any argument. | Capitals, three words, the bluntest label on the page. It states no price, guarantee or noun. | swap-offer Ours carries the noun and the guarantee (CL-16, CL-09). CS-P1's first button follows the winner's card, which is the better order for a reader who has read nothing yet. |
| 28 51.5% | ranked-entry 10× p · 24 w | A five-row score table for the winner only: dose, source, delivery, value at 33 cents a day on the current offer, and verified reviews, each scored five out of five. | specificitysocial-proofanchoring product-aware | Only the winner gets a table. The rivals had a letter; the winner has a letter and five rows, so the reader sees more measurement where the seller wants belief. | FABRICATION: five perfect scores with no method, and the review row contradicts the 4.4 stars printed in entry 42. A column of perfect scores reads as an advert. | keep-structure CS-P1 scores every product on the same three bars and gives itself 7.5 on cost per day. That is the stronger form: one table for all, one honest weak score. |
| 29 54.5% | ranked-entry h2 · 5× p · 61 w | Five plus points: the full 12 mg in every softgel, natural microalgae and no synthetic copy, an organic coconut-oil softgel, a clean label, and the lowest cost on the list at about 33 cents a day on the two-for-one offer. | reason-whyspecificityanchoring product-aware | After the scores come the words behind them. The order is the order of the rules, then the label, then price last, where it is remembered. | Each line is label, dash, gloss. Five lines against the older ranking's nine. The daily price depends on an offer the reader has not yet been shown. | rewrite-claim CS-P1's winner card has five plus points of one line each. Ours: the full 12 mg in one softgel (b7, b9), every batch third-party lab tested (c3), natural microalgae from closed bioreactors (c4, c5), the 30-Day Money-Back Guarantee, free shipping. We cannot claim the lowest daily cost, and we make no carrier-oil claim (b8). b7b9c3c4c5 |
| 30 58.3% | ranked-entry h2 · 2× p · 36 w | Two trade-offs: it is sold online only, and its results build over four to six weeks and so reward daily use, which a bracket gives as the reason the guarantee runs 90 days. | candorobjection-preemptrisk-reversalreason-why product-aware | Slowness is the strongest objection to a supplement. The seller raises it, inside the winner's card, and answers it in the same line, before the reader can set it against the price. | The second trade-off is a true cost to the buyer, unlike the older ranking's boast about selling out. The bracket is a spoken aside. Twenty-seven words that do more than the score table. | rewrite-claim CS-P1's list of things worth knowing has three honest limits and none about time. Ours: energy in weeks 1 to 2 (r7) and eye strain relief in 14 to 28 days (b6), both inside the 30-day guarantee; kidney labs are re-drawn at 8 to 12 weeks (r10), outside it. CS-P1's FAQ asks the reader to allow 12 weeks, which a reader will set against a 30-day window. Say plainly what the window covers. r7b6r10 |
| 31 60.5% | cta cta · 4 w | A button invites the reader to claim the top-ranked product. | commitmentexclusivity product-aware | It stands at 60%, at the foot of the winner's card, after the trade-offs have been answered. The verb treats the product as a prize won. | Four words. A fourth button wording. | swap-offer CS-P1 closes its winner card with a full block: the noun, the per-bag price, free shipping and the 30-Day Money-Back Guarantee. Ours is stronger. |
| 32 61.4% | image img | A wide banner with no alt text separates the winner's card from the narrator's account. Its content was not captured. | ease product-aware | It divides the page in two: above it, a ranking a scanner can use; below it, the reasons, for the reader who wants them. | No alt text and no caption. | page-furniture CS-P1 moves from the last rival card to the timeline under a plain heading. No banner is needed. |
| 33 62.8% | turning-point h2 · 3× p · 91 w | Under a heading repeating the all-three-rules claim, the narrator admits expecting the top Amazon seller to win, reports that popular bottles each fail one test, and names the seller's product as the one with no compromise. | candorearned-authoritystory-transportcontrast product-aware | It comes after the card and not before it. The scanner has the result; the reader who continues gets the reviewer's change of mind, which answers the suspicion that the ranking was fixed from the start. | The first person singular appears here for the first time, after a page of 'we'. The opening promises straight talk, a frame that adds nothing. The middle paragraph names the three failures in three fragments and then sums them up. | rewrite-claim CS-P1 has no first-person passage; house framework 4 is a zero-narrator page (L-01). The content transfers without the 'I': what the label reading found, rival by rival. Ours rests on b7, c4 and c3. b7c4c3 |
| 34 67.0% | proof-study h2 · 2× p · 59 w | A research heading, then counts: over 70 human trials in peer-reviewed journals and 3,500 papers. The studies are said to cover the skin's moisture and elasticity, eye fatigue from screens, recovery after exercise and everyday oxidative stress. | borrowed-authorityconsensusspecificity product-aware | The science arrives at 67%, after the ranking. The reader came already believing in the ingredient, so proof is reassurance placed near the offer and not an argument placed at the top. | Hedged verbs throughout: the research examines a role in supporting. Nothing is named: no journal, no trial, no year. Two sentences of 27 words each, flat. | swap-proof CS-P1 has no research section; its claims sit in the intro and the FAQ. Ours: 60+ clinical studies (r14), human trials on wrinkle depth, elasticity and hydration (b1), the double-blind trial in screen workers (b3), the joint science (b4). x3 asks that Harvard-level language be paired with the actual journals. Sell the eye relief and do not dramatize the screen problem (p4). r14b1b3b4x3 |
| 35 69.9% | image img | A wide banner with no alt text sits inside the research section. Its content was not captured. | ease product-aware | It splits the counts from the sentence that ties them to the product, so the tie-in at entry 36 starts a fresh paragraph under a picture. | No alt text and no caption. If it shows a chart or a journal page, the text never refers to it. | swap-proof House law L-05 asks for a real PubMed, NIH or CDC screenshot with a verified identifier on doctor-led pages. On a ranking page a real abstract screenshot beside our study claims (b1, b3) would be proof a reader can check. b1b3 |
| 36 71.5% | differentiation p · 26 w | One sentence ties the research to the product: the studied results came from 12 mg in an absorbable form, which this product is said to deliver and most bottles to miss. | borrowed-authorityexclusivitycontrast product-aware | It follows the counts at once, so the reader cannot credit the studies to astaxanthin in general. The science is made conditional on the three rules. | Twenty-six words and a dash. The logic is a leap: studies of a dose are presented as studies of a product. | rewrite-claim Ours is narrower and true: 12 mg is the dose the human trials used (b7), and the natural, microalgae-derived form is the one in the positive studies (c4). CS-P1's method box says this. We claim the dose and the form, never the studies themselves. b7c4 |
| 37 72.6% | disclosure p · 36 w | In body type, the narrator states that none of this is a claim to treat or cure; it is an antioxidant, the research indicates support for the cells, and that depends on the molecule arriving there. | candorobjection-preempt product-aware | It is placed right after the research and before the explanation of how the molecule works, at the point where a careful reader starts to wonder whether the page is overclaiming. It buys belief for the offer that follows. | Three sentences; the second has three words. It reads as a person drawing a line. The last sentence bends the disclaimer back into a selling point. | keep-structure CS-P1 has this in three places: the first line of the things worth knowing, the kidney FAQ and the footer. It does not have it as a spoken paragraph in the body. Worth adding beside the organ claims. |
| 38 74.2% | mechanism h2 · 3× p · 107 w | A heading says all three rules must hold together. The molecule dissolves in water and fat, lies across the whole cell membrane and guards both zones; the rules let it build up over weeks; missing one undoes the rest. | new-mechanismreason-whyexclusivity product-aware | The deepest explanation is held until 75%, after the winner is known, so it reads as the reason the winner won. Placed earlier it would have been a lecture before the reader cared. | The middle paragraph is three parallel sentences on what the research uses. The section ends on a fifth restatement of the only-one claim. The biology can be pictured: a molecule lying across a wall. | swap-mechanism This is The Cell Shield slot: the molecule spans the entire membrane (m2) where ordinary antioxidants cover part of the cell (p3). CS-P1 gives it one sentence in the intro and one in the winner's body. A short section after the cards would let the reader who stays learn why. Explain first, then the noun (CL-07). m2p3b7c4 |
| 39 79.0% | proof-social h2 · img · quote · p · 47 w | A heading, a square picture (not captured) and one quotation from an unnamed verified customer: a cheaper astaxanthin had done nothing and seemed not to suit them; after a few weeks on this one, recovery and skin improved. | social-proofidentificationloop-closedvindication product-aware | The testimonial is the reader of the opening paragraphs, a few weeks later. It closes the loop the lead opened and stands directly before the offer, so the reader sees the outcome and then the deal. | One testimonial on a page that claims 100,000 customers. No name, age, date or caption. The voice is smoother than a real review, with a dash and a finished metaphor. | swap-proof CS-P1's comment thread is four placeholder cards marked as a launch blocker. When real words arrive, choose one that mirrors our opening (years of vitamins, nothing to show) in a senior voice (house law P-10). No invented customer. r17 covers panel statistics once our panel has run. r17 |
| 40 84.4% | offer h2 · img · p · 27 w | A question heading asks whether the reader is ready for a dose that works. A product picture follows, then the offer: buy one, get one free, stated as up to 67% off. | commitmentanchoringease most-aware | The deal is held until 85%, after proof and testimonial. Earlier, a daily figure stood in for it. The reader meets the terms only once the product is wanted. | The heading reuses the older ranking's closing phrase. The offer line gives a percentage and no price in dollars, and no reason for the discount. | swap-offer Ours: 1 bag $29.99, 3 bags $59.97, 5 bags $89.95, free shipping, read from 29next at build time (CL-10). CS-P1 states the per-bag price of the 5-bag tier in its button block and names launch pricing as the reason. x4 lets the 5-bag tier be called the lowest price ever. x4 |
| 41 88.2% | cta cta · 6 w | The main offer button: try the brand without risk for 90 days. | risk-reversalease most-aware | It follows the deal at once, at 88%. Offer, button and trust line form one block, so the reader who is ready never has to scroll for the terms. | Six words. The same label as the older ranking's last button. | swap-offer CS-P1's buttons carry the noun and the 30-Day Money-Back Guarantee (CL-09). Check before traffic: they land on the PDP's Try Before You Buy box (ob=c), which BRAND.md describes with a monthly lane and $6.99 shipping, while the line under the button says one-time purchase and free shipping. |
| 42 88.5% | proof-social p · 9 w | A one-line trust strip under the button: the guarantee, a rating of 4.4 stars labelled excellent, and the 100,000-customer count. | risk-reversalsocial-proofconsensus most-aware | It sits directly under the button, where the eye rests before a tap. Each item is a fact the reader has met before, now gathered at the point of decision. | Nine words, separated by dots. The rating is below the perfect review score in entry 28, which makes it the more believable of the two. No source for either. | swap-proof CS-P1's proof bar has three items (30 days, 12 mg in one softgel, lab tested every batch) and its sidebar has a rating with a four-digit customer count (CL-11). Both numbers need to be the store's real ones before traffic. c3b7 |
| 43 89.5% | guarantee h2 · p · 118 w | A long argument: the compound needs four to six weeks, so a fair trial needs over a month; this brand gives 90 days, most rivals 30; a short guarantee bets on the buyer quitting, a long one on staying. | risk-reversalreason-whycontrastcommon-enemy most-aware | It comes after the button, for the reader who did not tap. The hesitation at this point is about wasting money again; the argument says the seller carries that risk for three months and that rivals will not. | The strongest reasoning on the page and its longest paragraph, 113 words. It opens by promising a tell, poker language that flatters the reader as someone who can read sellers. Two mirrored sentences about the two bets close it. No word on how a refund is claimed. | rewrite-claim CS-P1 states its guarantee in the FAQ and the buttons with no reason attached. Ours is the 30-Day Money-Back Guarantee, bag kept (CL-09), so this argument is aimed at brands like ours. Our reason has to differ: the first changes are due inside the window (energy in weeks 1 to 2, r7; eye strain relief in 14 to 28 days, b6), and the bag is kept either way. r7b6 |
| 44 94.6% | close h2 · img · p · 53 w | The close, under a heading and a banner picture, tells the reader to choose the product that passes every test, chants the three specifications, blames anything less for the last bottle's failure, and warns against losing more months. | loss-aversionloop-closedurgencycommitment most-aware | Last paragraph. It returns to the failed bottle of the first paragraph and names it as the cost of choosing wrong again. The urgency is the reader's own lost time and not a deadline. | Three sentences, a three-item chant and a last line of six words with no product name. The chant drops third-party testing from the older ranking's four items, to match the three rules. | rewrite-claim CS-P1's close says the difference is the full 12 mg in one softgel, lab tested and shipped free, and that it shows on the next labs. A callback would sharpen it: return to the cabinet of vitamins from the opening (r3, r23), then the chant (b7, b9, c4, c3). The next-labs promise should go to the claims reviewer (r10, r11). b7b9c4c3r3r23 |
| 45 99.0% | cta cta · p · 16 w | The last button repeats the 90-day trial. Under it a line lists the guarantee and two stock phrases, one of which belongs to skincare photography. | risk-reversalease most-aware | It is the final element, at 99%. The page ends on the guarantee and carries no disclosure or supplement disclaimer after it. | LEFTOVER: the last phrase of the line comes from a skin page's template. In the extracted rows this page has no advertiser disclosure and no regulatory disclaimer anywhere. | swap-offer CS-P1's last button is followed by one line (one-time purchase, 30-Day Money-Back Guarantee, free shipping) and by a footer with the supplement disclaimer and the ad disclosure. Ours is complete where theirs is bare. |