How we pick and rank

A person chooses which stores we cover. For each of them a script reads the store’s own public catalogue and adds a set number of products, taken in turn from each product type so one bestseller line can’t fill the page. Amazon products come the same way from Amazon’s product API, for brands we chose.

A creator can also add a product to their storefront by hand; those pages name the creator and say whether they’ve used it. No store can submit a product, and no store can pay to be listed.

The badge means all four of these are true — it is not a taste award and not paid placement:

• The store page is live and the product is actually in stock.
• Price and availability were verified against the store, not copied from a feed.
• The store publishes a rating for it, or the creator has used it.
• It is not the most expensive way to get the same thing (see the comparison block).

Every alternative shows a match percentage built from four observable facts, and you can open the breakdown on any of them:

• Same kind of thing — same category (35 points)
• Price distance from the product you’re looking at (30 points)
• Shared product-type words in the name (25 points)
• Same store, so sizing and shipping carry over (10 points)

We do not compare fabric, fit, or how things look side by side, because we don’t have that data.

Store links are affiliate links: if you buy, the store may pay Avahit a commission out of its own margin. That never adds anything to your price. It also never moves a product up a list — ranking rules are the ones above, and they don’t take commission into account.

Commission decides one thing, and only on Amazon: we don’t import product categories where Amazon pays less than 3%.

The most useful thing we can tell you is where our knowledge ends. Publishers like Tom’s Hardware say plainly which categories they measure and which they only judge by experience. Our answer is simpler: we measure nothing.

• No lab, no test bench, no sample units, no hours-tested counters.
• We never score a product ourselves — no ratings out of five, no verdicts.
• We don’t compare fabric, fit, durability or anything else you’d have to hold to know.

What we actually have is four things, all first-party or attributed:

• Our own price checks. Every product is re-checked automatically once a day; the history on each page is our own record, not a feed.
• Our own click counts. What visitors actually opened. Popularity, not quality — we never relabel it as “best”.
• The store’s rating, shown with the store’s name, the review count and the date we read it.
• Whether the creator owns the item — bought it, was gifted it, or hasn’t used it. Stated on every product page, and gifted items are called out separately because it changes how you should read an opinion.

Lists are rebuilt from the catalogue automatically, so they can’t drift out of date the way a hand-written article does. A product leaves a list when it goes out of stock, moves out of the price band, or the creator removes it. Each list shows the date of the most recent real price check — not the time you loaded the page.