Methodology

How I rate vendors: the Papa Score

The Papa Score rates research peptide vendors on two axes instead of one ranked number. This page is the full rubric. If a vendor review links here, this is what "See how scoring works" means.

Why two axes instead of one score

A single number can't answer the question you actually have, which is whether a vendor is right for you. Two vendors can land on the same total for opposite reasons: one with excellent lab documentation and a brutal returns policy, another with thin paperwork and genuinely good support. Averaging those into one figure hides the only thing that would help you choose.

So the score separates what is proven about the product from what it is like to be the customer. Read the shape, not just the total.

The policies

• Scores are based on publicly verifiable documentation and my own ordering experience, not on a vendor's marketing copy repeated back as fact.
• Anything I haven't tested is excluded, not scored as zero. If I haven't placed a real order, the ordering, shipping, and packaging points leave the rubric entirely. If I haven't sent a vendor a question, the responsiveness points leave too. A gap in my research is a fact about me, not a finding about the vendor, and it should never be quietly converted into a penalty. That's why scores are shown out of 85, 90, 100, or 105, rather than always out of 105.
• Each finding is counted once. If a vendor's certificates name another company, that's scored in the one subcategory that owns it. It doesn't also get deducted from three other places for being the same problem in a different hat.
• Marketing puffery is not scored. Founding-year boasts, "trusted by thousands," and "number one" banners are how nearly every vendor here talks, and they say nothing about the product. What is scored is what a vendor says about the product itself. A product claim needs a document behind it: one I cannot verify earns no credit, and one a document contradicts costs points, such as advertised purity testing that no published document contains.
• I don't score traits that every vendor in this market shares. Almost none of these companies names a legal entity, most have a domain under two years old, most sell all-sales-final, and many use a mailbox address. Deducting equally from everyone just lowers the whole scale without telling you anything. Those characteristics get explained once, where they belong, instead of twelve times as a penalty.
• Affiliate relationships never affect scores, and no vendor can pay to change a score. I only review vendors I have an affiliate relationship with, but the relationship doesn't buy a better number.
• Pricing is reported because it's useful to know, but it is never part of the score. A vendor doesn't score higher for being cheap or lower for being expensive.

Axis one: Evidence Quality, up to 46 points

What is documented and independently verifiable about the product itself. This is the half that separates a vendor who tests their inventory from one selling an unverified guess in a vial.

• COAs published (0 to 12). Whether certificates of analysis are posted where a buyer can see them before ordering, not just available on request after the fact, and how much of the catalog they actually cover.
• Independent lab (0 to 10). Whether the testing lab is named, independent from the vendor, and whether its own records can be checked. A lab running a public verification portal scores higher than one whose reports can only be taken on faith.
• Purity documented (0 to 10). Whether purity is stated with a method (HPLC, mass spectrometry) and whether the tests on file actually measure it. An identity-and-quantity assay does not, even sitting next to a purity guarantee badge.
• Batch matched (0 to 8). Whether certificates are genuinely lot-specific and current, rather than one submission covering the whole catalog once and never repeated.
• Test panel breadth (0 to 6). One point for each of these that appears on the certificates a vendor actually publishes: purity by chromatography, identity or quantity, endotoxin, sterility or a microbial screen, heavy metals, and one further screen such as fentanyl, water content, or counter-ion content. A test the vendor advertises but does not publish earns nothing, and a test that appears only on the lab's own copy of a certificate does not count. Vendors whose certificates come from two labs with different panels are scored on the mix.

Axis two: Buyer Experience, up to 45 points, or 60 with a real order

What it is like to actually be their customer.

• Community sentiment (0 to 10). What independent, non-affiliate sources say, weighed by how independent they really are. A rating only counts once I've confirmed the profile belongs to that exact storefront, top-level domain included. Vendor-hosted testimonials are not community sentiment and score nothing.
• Refund policy (0 to 8). Whether a clear, reasonable policy exists in writing, and whether the policy page and the product pages agree with each other.
• Research-use labeling (0 to 8). Whether products are labeled and marketed for laboratory research use, without dosing instructions or consumer framing that misrepresents what's legally being sold.
• Site transparency (0 to 7). Whether the site says who runs it, where it ships from, and how to reach a real person rather than a form and nothing else.
• Catalog consistency (0 to 6). Whether listings, specs, and the vendor's own stated policies stay consistent with each other, rather than the homepage promising one thing and the policy page another.
• Responsiveness (0 to 5). I send every vendor a question as a prospective customer before I review them. I score whether a person replies, and how fast: 5 points for a human reply within 24 hours, 3 points for a human reply after 24 hours but within 7 days, 0 points for no human reply, or one that takes longer than 7 days. Automated acknowledgments do not count as a reply. This line scores the reply, not its content. If a reply reveals something about a vendor's testing or policies, that is scored once, in the category it belongs to. A vendor that hasn't replied yet scores 0 until a reply arrives, and the score is updated when it does.
• Checkout and payment (0 to 5), shipping speed (0 to 5), packaging and cold chain (0 to 5). Scored only from an order I placed and received. Excluded entirely otherwise. Shipping speed is judged against the delivery window the vendor states: an order that arrives inside that window scores full marks, and one that arrives late loses points in proportion to how late it was.

Flags

Some findings don't belong in a number. If a vendor's published certificates disagree with the testing lab's own records, that matters whether their score is 40 or 80, and burying it as a few lost points would be the wrong way to tell you.

So those are reported as flags: named, specific, and shown beside the score rather than subtracted from it. A flag is never also a deduction, because the underlying fact was already scored once in whichever subcategory owns it.

Flags marked integrity are a separate matter. Those cover a vendor misrepresenting documentation: publishing another company's lab reports as their own, or publishing results that differ from what the lab actually recorded, or advertising testing that no published document contains. Any integrity flag sets the band to Not Recommended regardless of the score, because that finding is categorical rather than incremental. An excellent lab doesn't offset it.

Bands

The total is normalized against whatever that vendor's applicable maximum was, then mapped to a band:

• 82% and up: Papa's Pick. The vendors I'd point a friend to first.
• 68 to 81%: Recommended. Solid, nothing that would stop me ordering.
• 54 to 67%: Solid with Caveats. Good enough, with specific gaps worth knowing first.
• 40 to 53%: Mixed Record. Real strengths and real problems, in roughly equal measure.
• Below 40%: Limited Verification. Not an accusation. It means too little of what this vendor claims could be independently confirmed for me to tell you much either way.
• Not Recommended. Reserved for vendors carrying an integrity flag, and unreachable by a low score alone. Being under-documented is a gap. Misrepresenting a document is a different thing, and only the second one earns this label.

A vendor that hasn't been scored under this rubric yet shows "Not Yet Scored" instead of a band. That's not a bad score, it's an honest gap: I haven't put in the verification work required to publish a number I'd stand behind.

Revision history

This rubric was revised in September 2026. The previous version used five weighted categories and scored untested items as zero, which penalized vendors for the limits of my own research and let a single finding cost points in three places at once. It also branded thinly-documented vendors "Not Recommended," which conflated a lack of evidence with evidence of a problem. The current version fixes all three. No vendor had been published under the old rubric.

Changelog

• September 18, 2026: Responsiveness now scores whether and how fast a person replies, on a fixed 24 hour and 7 day scale. The content of a reply is scored elsewhere, so no vendor is penalized twice for the same finding. All reviewed vendors were rescored.
• September 18, 2026: Added a test panel breadth line (0 to 6) to Evidence Quality, so a full screen is scored directly. Panel breadth is judged only from tests that appear on published certificates. All reviewed vendors were rescored.
• September 18, 2026: Shipping speed is now scored against the delivery window a vendor states, not against a fixed number of days. Ascension was rescored under it. Nura's order already met its stated window.