Methodology

How Espresso Machines currently produces guides and product candidate lists.

What this is

This is a description of the current publishing pipeline. It is not a claim of independent laboratory testing, in-home use, or expert certification.

How a guide is planned

Each guide starts as a planned row: title, slug, intent (for example buying, comparison, or problem), content cluster, primary product family, and launch priority. Those fields decide what kind of page it is and which product family the candidate search should stay inside.

How the article is written

A draft is generated from that plan with AI assistance, then published by the operator. The draft is not allowed to invent ASINs, prices, ratings, or firsthand testing. See Editorial policy.

How product candidates are chosen

When a guide asks for products, the system sends catalog search queries to Amazon and keeps candidates that pass relevance and public-display filters. Eligible items can appear as an early recommended option and as later product groups. Ranking inside those filters is not a lab ranking and is not labeled as “tested.”

An Amazon API result is not evidence of a tested recommendation. Candidate sourcing is separate from hands-on product testing. This site does not operate a test lab.

Limits

• Catalog coverage can be incomplete or change after publication.
• A highly ranked accessory can appear before a core machine if the search slot asks for that accessory.
• Stored prices are catalog values as saved, not a live checkout quote.
• Guides can be wrong, dated, or too narrow for a specific room or job.