23 August 2026
In plain English
Four companies now sell experimental drugs called peptides (short protein fragments used for recovery and longevity) online across America, each with a doctor's review and a real pharmacy behind the order. A comparison site ranked them and updated it three days ago, prices and all.
Two of those four already pay us commission.
We hold a forecast saying money from peptide sellers stays a small slice of our income, because they pay far less per customer than weight-loss clinics do. These four do not pay like peptide sellers. They sell year-long bundles at $139 a month, which is weight-loss-clinic money.
We missed it because we were watching the wrong thing — our own peptide website, which has had no visitors for 27 days running. It arrived through the commission ledger, not the traffic report.
Signal detail
The first evidence in 27 readings that tests this row rather than confirming it. The forecast is priced on one mechanism: peptide affiliate CPA is 10–30× below GLP-1 affiliate CPA because vendors pay 10–20% on an $80–200 order. Healing Maps, fetched at source (HTTP 200, 1,227,136 bytes, updated 2026-08-20), scores four national programmes — and two are already in our Katalys table. Bodybuilding Health+ sells "Sermorelin + NAD+ + GLP-1, the men's stack" at $139/mo on 12-month bundles; its Katalys programme is named GLP-1 & Longevity and it took the table's second-highest EPC $133.33 on one $400 conversion this week. Strut Health (oral Sermorelin, ~$99/mo) sits at 15 clicks, 0 conversions. Bodybuilding.com's own store carries a Peptides category and a BB HEALTH+ section. The peptide-versus-GLP-1 revenue boundary this row assumes is clean is not clean inside our own payout table, and a 12-month bundle is not vendor-commission economics. The cut is 0.05 and not 0.15 for one honest reason: nothing attributes that $400 to a peptide rather than a GLP-1 purchase. Unchanged underneath: bestpeptideforthat 0 clicks, 25 pages, 51 impressions, 27th consecutive zero; 48/88 → 25/51 is 07-24 rolling out, not decay.
Up on competitive pressure, not on a pathway. The legal side was checked and is unchanged: Federal Register silent at day 31 (count 5, none a 503A bulks action) and EDGAR 0 hits on the Hims CIK since 08-13. What is new is four operators already selling what Hims would sell, one carrying BPC-157 at ~$116/mo. Held to +0.02: the source is SECONDARY and affiliate-disclosed, it cannot establish those protocols are lawful, and the FTC/Utah/California complaint pending in N.D. Cal. cuts the other way harder than a listicle does. 130 days out. peptide-safety-incident: holds 0.95, sixth day at ceiling. Falsifier queried, negative; the top-5 field is unchanged since 08-16. openFDA `meta.last_updated` reads 2026-08-12 for the third day, so `total 2` is a stale index read and is reported as one.
CMS byte-identical for a third day, day 51, no enrolment figure. Sixty-nine items arrived after yesterday's run read the digest; every Bridge item among them is coverage volume — failure case F1. Augur. The category is building a white market in public: Healing Maps' stated methodology excludes "research use only" suppliers with no prescriber and calls them unregulated chemical suppliers rather than telehealth. A gray market that publishes an inclusion bar is organising itself ahead of its regulator. Premortem — rotation: `medicare-glp1-bridge-uptake` (the only pending row never premortemed). It is 2027 and CMS reports 2.4 million. What had to be true? The $50 copay beat the paperwork: attestation moved into the prescribing software, Amazon fulfilment absorbed the friction NPR documented, and the 3.8M denominator was a 2023 artefact against a 2026 pool coded far wider. The latent condition is visible now — a national awareness wave running four months ahead of the first enrolment figure. Implication: 0.88 is this row's ceiling, not its floor. Stoic. One actionable item, and it is not a signal: the operator decision on `peptide-safety-incident`, day 22. Noise the operator cannot act on: the HIMS share-price cycle, the Barclays cut, the Visa chargeback penalties, the Zoetis holder churn.
Resolutions & upcoming
Nothing resolved. Nothing pending resolves inside 14 days — next horizons are 2026-12-31 (three rows) and 2027-03-31.
Two operator items unchanged. `resolutions.tsv` still says "all four nominated peptides were recommended" — seven were nominated, six recommended, over two days — and Ra may not edit that file. `peptide-safety-incident` has sat 22 days on a threshold with no rubric.
How Barque got smarter today
- The instrument was pointed at the wrong surface for 27 days. `peptide-affiliate-share` was re-confirmed daily off a click count that is zero and has told us nothing since 07-28, while what finally moved it sat in `first-party-pulse.md` the whole time — a partner named GLP-1 & Longevity. Proposed for the Evolution Log: when a forecast turns on a revenue ratio, read the payout table's partner NAMES every run, not just its totals. Zero cost, no network.
- A falsifier query answered a different question than it asked, and that was the value. It searched for a peptide platform reaching affiliate-network standards. Still no. But it surfaced the page showing two such platforms are already in our network. The 2026-08-02 rule is earning more than it was written to earn.
- One self-inflicted fetch failure, named rather than absorbed. The first dispatch sent the Federal Register URL with literal `[` and `]`; curl exited 3, zero bytes. Mechanical fix, a command not an intention: always send `%5B`/`%5D` in `-f urls`, because `agent-fetch` passes the string to curl unquoted. Cost one dispatch; would have cost a day-31 UNVERIFIED unnoticed.
- A partial primary reported as partial. PubMed `esearch` for the HRT/Alzheimer item returned 200 and 4,213 hits but rejected `sort=date`, so the eight ids are not the newest eight. Item stays UNVERIFIED; no order written off an unordered list.
- Yesterday's UNSIZED watch item is sized, and it is not a vertical. `peptides for brain fog` 390 at KD 0, `epitalon for sleep` 110 at KD 0, five terms null-volume at KD 5–17. Small, uncontested, cheap — a page order on a surface we own, which is what the threshold discipline is for. No order defects this cycle. Two reroutes stand, both already re-filed (08-02, 08-04); nothing new to re-file.