Good choices start with clear information.

This page explains how rate.red checks information and scores products, and what a score cannot tell you, so you can verify any ranking before relying on it.

Last updated

What we assess in each category

Each ranking has its own criteria, weights and rubric (the condition behind each score level), published on that ranking page. These are the topics we usually look at. They are a guide, not a fixed list.

  • Hosting: how clear the renewal terms are, backup schedule, management workload, and the resources the provider states.
  • Headphones: noise cancelling and battery life as the manufacturer states them (figures compared in the same mode, such as ANC on for both), foldability, and supported codecs.
  • Home and living: mounting method, footprint, bulb socket and electrical specification.

Evidence levels

Every score carries a label saying what kind of evidence it rests on.

Evidence levels and the label shown next to a score
LevelMeaningLabel shown
specScored from specifications or documents the manufacturer publishes, using the published rubric. The team did not use the product.Assessed from manufacturer information, not tested
testedThe team used the product and recorded the method, equipment, conditions and where the sample came from on the review page.Tested by the rate.red team
mixedA ranking that contains both kinds of score. The evidence label appears on each individual score.A label on each score

A spec score is not a test result. It says nothing about sound quality, comfort, real website speed or service stability, and we do not mark such scores up as a review in the structured data (schema.org) of the page.

Heart ratings and ranking order

Comparison guides use1–3 hearts for editorial interest, supported by each product’s stated rationale. These are not customer stars or team test results.

  • — Strong editorial interest within this comparison
  • — Interesting for specific needs or users
  • — A narrower fit with more trade-offs

Recommendation order is separate from the heart count. Missing evidence is not assigned0or1heart. Existing criterion ratings retain their original scope and conditions, rather than becoming overall quality scores.

Numeric scoring used in legacy structured records
  1. Each criterion is scored from 1.0 to 5.0 in steps of 0.5 using that criterion’s rubric. Every score has a note or a source.
  2. Weights are whole numbers that add up to 100, and each ranking has between 2 and 8 criteria.
  3. Total score = the sum of (weight × criterion score) ÷ 100. A made-up example to show the formula, not a real product: criterion A with weight 60 scores 4.0 and criterion B with weight 40 scores 3.0, so the total is (60 × 4.0 + 40 × 3.0) ÷ 100 = 3.6.
  4. Rounding: the total is kept in full and shown with one decimal, rounding half up (4.25 is shown as 4.3). The order is calculated from the unrounded value.
  5. When totals are equal, the score on the highest-weighted criterion decides (if weights tie, the criterion listed first). If that is equal too, products are ordered by product code (slug) alphabetically, which says nothing about quality.
  6. A ranking with scores is never ordered by hand. Manufacturers, shops and commission cannot set its order. Read more in the editorial policy.
  7. Price is not part of the score because prices change quickly. If a price is shown, it always names the shop, the date it was found and the source.

When information is missing

If a product has no score for any one criterion, we do not rank it. We do not fill in zero or substitute an average. A specification the manufacturer does not state is shown as “not specified”; we do not guess numbers.

Use-case shortlists and scored rankings

Category pages and the earlier numbered lists (01, 02, …) are ordered by use case, with the reasoning explained on the page. They have no scores and are not test results, sales figures or popularity rankings. A page under “Rankings” that shows scores is calculated from the criteria above. Each page says which kind it is.

See the rankings

Where sample products come from

A hands-on review always says where the tested product came from: bought, loaned or received free. A spec score has no sample product. This appears in the method section of the product page. The policy on free and loaned products is in the editorial policy.

Check dates and updates

The “check date” is the day we opened the original source and compared it with the page. It is not the day the manufacturer changed anything, and it does not promise the information is still current afterwards.

The “updated” date is the most recent recorded change to the content.

When the scores, weights, criteria or products of something already published change, the system does not accept the change without a reason, and records the date and reason in that page’s “edit history”. When a product is taken out of a ranking, for example by moving it to draft or deleting it, the system records that with the date in the edit history of every published ranking that lists the product. For other substantial changes to the content, the editors add a note themselves.

A year appears in a heading only when the content was really updated in that year. Some information may be older than you assume, so check the date on each page before deciding.

Limitations

  • rate.red is not a laboratory. We have no specialist measuring equipment, and any test result comes from the team using the product by the recorded method.
  • Spec scores do not measure sound quality, comfort, real-world speed or service stability.
  • Manufacturers measure under different conditions, such as battery life with ANC on or off, or at different volume levels, and some state no measuring conditions. We compare figures in the same mode (for example ANC on for both), but finer conditions such as volume level or codec may differ, and we put the conditions we know in the note beside the score.
  • The model sold in Thailand may differ from the one the manufacturer’s page describes. Check the specification of the listing you are buying.
  • We compare only the products in that ranking, not everything on the market.

Writers and reviewers

Each ranking names its writer, editor and any experts, with the scope of each person’s work. An expert who only reviewed the how-to-choose section is not presented as having chosen or endorsed products. A page without a named person is produced by the rate.red editorial team. The list and scopes are in the team section.

Prices, sellers and links

Shop links point to manufacturer pages or to Shopee and Lazada search results. We do not display live shop prices or confirm stock. Buyers should confirm the model, delivered total, warranty and returns.

Right now these links do not earn rate.red any commission. Details are in the editorial policy.

Read the marketplace buying checklist

Languages

Thai is the main edition and some pages exist only in Thai. An English page is published only when the text and the data of every product on it are fully translated. Some sources describe overseas variants, so check the specifications of the actual Thai listing before ordering.

Found a mistake

How we fix information that is wrong or out of date is described in the corrections policy, which also says that there is no email address for reports at the moment.

Edit history of this page

  • — Rewritten to match score-based rankings: added evidence levels (spec and tested), the score formula, tie-breaking, handling of missing data and limitations. The earlier text said there were no scores, which no longer holds.
Explore the shortlists