How We Evaluate Products

SexToysDB is a spec-first catalog, not a panel of anonymous five-star reviews and not a medical advice site. This page states how products get into the database, what “evaluation” means here, where automation helps, and where we stop.

Canonical policy detail lives on Editorial & Sourcing Policy. This post is the plain-language version you can link from guides and product pages.

What “evaluate” means on this site

  1. Ingest facts — dimensions, materials, waterproof rating, connectivity, battery minutes, noise in dB when the manufacturer states a number, price bands, and retailer offer URLs.
  2. Normalize — units into millimeters, grams, minutes; structured columns instead of one-off marketing blurbs where possible.
  3. Derive — fixed formula indices (Satisfaction, Durability, Stealth, Value, Travel) and care-oriented cards (lube compatibility hints, porosity/safety status, charging type) from those facts. See how catalog scores work.
  4. Explain briefly — short grounded copy so a human can scan intent without a wall of interchangeable essay sections.
  5. Price-check — live offers and history when coverage exists. See how deals and price history work.

We do not mean: a double-blind lab, a clinical trial, or “our staff used this for six weeks and ranked orgasms.”

Standards we hold ourselves to

Specs before story

If a fact is in the database, it should appear as a table cell, chip, or compare row, not only as a paragraph. Prose must not contradict typed fields.

No invented measurements

  • Noise stays blank without a stated dB figure.
  • We do not turn “quiet” / “powerful” adjectives into fake numbers.
  • Non-USD offer prices are not FX-converted into pretend USD deals.
  • Formula indices are not published as schema.org AggregateRating or star reviews.

Body safety without medical theater

Material and porosity guidance stays in hardware/material language and cites third-party context where needed (see materials guide and are sex toys regulated). We do not diagnose, treat, or promise therapeutic outcomes.

Consistent rules, no pay-to-win scores

Affiliate commissions do not change specs, indices, facet membership, or deal math. Placement is not sold as a rating boost.

Index only what earns it

Thin or flagged product pages stay out of the public index until data and quality gates pass. Empty brand/category shells are not a growth strategy.

AI-assisted disclosure

We use language models as writing assistants to turn dense spec sheets into readable short descriptions, intros, pros/cons, and FAQs under constraints:

  • Grounding — only facts supplied from our product record; no invented features.
  • No hype superlatives as product truth (“revolutionary,” “the best,” etc.).
  • No medical/therapeutic claims.
  • Human review — pages that fail checks are flagged and corrected before we treat them as sound catalog entries.

AI does not set Durability/Stealth/Value scores, does not decide deal %, and does not replace manufacturer documentation as the source of truth. Longer interchangeable essay blocks are being reduced over time in favor of structure and short grounded text — quality of unique data over word count.

Internal drafting standards for commercial listicles and hardware posts also live in our article guidelines (repo: docs/article-guidelines.md): hardware metrics over wellness advice, cite authorities on materials, prefer live catalog embeds over pasted prices, and do not scale thin “best X” URLs for their own sake.

What we don’t claim

Claim type Our stance
“Lab tested in our facility” Only if we actually ran and documented that test — default is manufacturer specs + formulas
“Doctor recommended” / ED / pelvic therapy cures Out of scope without proper medical review process; usually avoided entirely
Guaranteed lowest price on the web We show our offers and history only
User star average from thin affiliate reviews Not fabricated; not mapped from formula indices
Every quiet marketing adjective = whisper-quiet facet Facets like whisper-quiet and long battery life need real typed data

How you should read a product page

  1. Offers + price insight — what retailers we track and whether a drop vs list/history is real.
  2. Typed specs + free-text leftovers — the decision meat.
  3. Derived indices + care cards — fast compare signals; open the methodology when unsure.
  4. Short prose + FAQ — orientation only; if it conflicts with the table, trust the table and tell us.
  5. Compares / facets / guides — waterproof, app-controlled, materials and cleaning posts when the job is education rather than one SKU.

Corrections

Wrong dimension, material, or dead offer URL: contact the SexToysDB Editorial Team. We would rather leave a field empty than publish a confident guess.

Related reading