The SleekNova Lexicon

One number from four weighted dimensions: WOW at 35%, GLOW at 25%, MOMENT at 25%, TRUST at 15%. Nothing publishes below 8.0. There is no separate minimum on WOW, GLOW or MOMENT; the weighted total is the line, and two further floors sit under it.

• Weighted total below 8.0. The line every product clears or does not.
• TRUST below its floor. Thin evidence is a rejection on its own.
• Confidence below its floor. Too little standing behind the claims.

Those three are the whole list. Every rejection on The Bar cites one of them.

The discovery dimension, weighted at 35%. Measures whether a gift provokes the reaction "how did you find this?" Judges uniqueness, conversation potential, and the sense of deliberate curation.

The aesthetic dimension, weighted at 25%. Measures visual beauty, design intent, and material quality. A high GLOW indicates a product that photographs well and shows intentional design choices.

The scene dimension, weighted at 25%. Measures the unwrapping experience, presentation quality, and emotional resonance. Judges whether the gift creates a memorable moment when it is given.

The evidence dimension, weighted at 15%. Measures source quality, review authenticity, and how much of the product narrative survived verification. It is the only dimension with a floor of its own.

The hard minimum on the TRUST dimension. A product whose evidence is too thin is rejected on that ground alone, whatever its weighted total. This is the second of the three gates that actually reject products, and unlike the total it cannot be offset by a strong showing elsewhere.

The five-factor evidence audit every product passes: Source Diversity, Source Quality, Claim Verification, Platform Integrity, and Temporal Consistency. It produces the Confidence Score, which is the evidence behind the TRUST dimension. Renamed from "Trust Matrix" in August 2026 so that one word no longer named two systems.

The 0 to 100 output of the Confidence Audit, recording how much evidence stands behind a product rather than how good the product is. It carries a floor of its own: too little evidence is a rejection even when the product itself scores well.

The status a product reaches when every factual claim in its narrative has been traced to a primary source document. Nothing publishes without it. A claim that cannot be traced is cut, not softened.

The chain of custody for a product's verified claims. Each entry records the claim, the source URL, and the verbatim quote it came from. The Truth Map is the auditable proof that a claim survived verification rather than an assurance that it did.

The adversarial pass that challenges every factual claim before publication. It runs against primary sources at character level, and it is designed to fail claims, not to confirm them. Claims that cannot be grounded are removed.

The method behind the Skeptic Protocol: an active attempt to disprove each claim in a product narrative, assuming it is false until a primary source proves otherwise. Only claims that survive the attempt are promoted to Grounded.

The standard requiring each specific attribute to be verified on its own, rather than accepting a product description as one generalised truth. "100% tin" and "founded 1916" are two claims, checked twice. The strictest grounding standard we hold.

The five steps every product passes before it can publish: Discovery, Quick Filter, Deep Verification, Scoring, and Final Quality Check. Each step has one responsibility, and a product does not advance until the step before it has been signed off.

The research environment where the pipeline runs. The Lab sits between raw product data and anything we recommend, and it publishes what it finds either way: products that fail are documented in public, not quietly dropped.

A 20 to 25 word summary written to be quoted whole by a language model. Product name first, price included, grounded claims only. One per product, no exceptions.

The presence of verified detail in a product narrative that does not exist in the maker's own marketing copy or in generic third-party reviews. It is the test of whether original research happened, and it is why a page that only rephrases a product description does not ship.

Synthesised observation rather than raw fact: the purchase hesitation, the insider detail, the comparison worth making. It carries the context that cannot be cited to a single source, and it is kept separate from grounded claims on purpose.

A long-form study of one research question, published in the Lab and cited by the product pages that rest on it. A Deep Dive is where a study is read properly, so that a product page can state a finding in one line without inventing it.

The published standard a product must clear to appear on SleekNova, together with the live record of what it has rejected. Three gates do the rejecting: the weighted total, the Trust Floor, and the Confidence Score floor.

The published verdict on one product: what was checked, what was found, and whether it passed. Every product we publish has one, and so does every product we turned down.

A product that did not clear the bar, published with the reason it failed. Rejections we initiated are published in full. A product submitted by its own brand that does not pass is answered privately, and nothing about it is published.

A published record of something we got wrong and have now fixed, including the ones our own checks caught rather than a reader. The original claim, the correction, and the date all stay on the page.

A published record of a rule changing: a threshold moved, a standard tightened, a practice retired, with the reasoning that moved it. It covers how we judge, never the verdict on one product.