# How to choose comparison criteria before you name products > A way to pick criteria from the reader’s job so a shortlist does not collapse into brand familiarity. - Canonical: https://listedpress.com/blog/how-to-choose-comparison-criteria - Markdown: https://listedpress.com/blog/how-to-choose-comparison-criteria.md - Site: Listed Press (https://listedpress.com) Published: 2026-11-03 ## How do you choose criteria for a comparison? List the decisions the reader cannot avoid, turn each one into a criterion you could score without liking the brand, and drop any criterion that every serious option will pass. Name products only after that list is stable. ## Criteria are decisions, not features “Has an API” is a feature. “Can a support lead launch it without an engineer” is a decision. Features proliferate until the table is unread. Decisions stay tied to the person in the introduction. Write decisions as questions the reader would ask a colleague: Who sets it up? What does it replace? What breaks if we are wrong? Each question that survives becomes a criterion. ## Throw out criteria everyone passes If every product you might list has SSO, SSO is not a criterion for this page. It is table stakes, and it can be one sentence in the introduction. A criterion that does not separate options is decoration. The same goes for “easy to use” and “great support.” They feel like criteria and score like moods. Replace them with a check a stranger could repeat: time to first ticket, or whether support is chat, email, or a forum. ## Weight them, or admit they are equal A reader weighing a help desk for ten people cares more about setup than about a sandbox environment. If you do not say that, your ranking secretly used a weight and hid it. Either assign weights that add up in public, or say the criteria are unweighted and the reader should sort for themselves. The criteria weighter on this site does the arithmetic. It will not pick your weights. If moving a weight by ten points changes the winner, the page should say the result is sensitive, not pretend it is a law. ## Name the source of each score A score needs a trail: docs, a pricing page, a trial, a quote from a user with a role. “We think” is not a source. When the source is your own use, say what you used it for and for how long. If you cannot source a criterion, cut the criterion. An empty cell is more honest than a guessed 4 out of 5. Readers forgive missing data. They do not forgive a table that invented it. ## Freeze the list before you fall for a product Write the criteria in a note that does not contain product names. Then add names. If a product you like fails a criterion, the criterion stays. Changing the criteria to save a favorite is how a shortlist becomes an advertorial with a grid. You can add a criterion later if a reader’s job demands it. Mark it as an addition and rescore everyone, including the products that looked good under the old set. ## Hand the criteria to the reader Put the list near the top, in words, not only in a table header. A reader who disagrees with the weights should be able to ignore your ranking and still use the facts. That is the difference between a shortlist and a verdict they did not ask for. End by saying which single criterion would change your own pick. That sentence teaches the reader how to use the page when their job is slightly different from the one you wrote down. ### How many criteria are enough? Three to six decisions the reader cannot skip. More than six usually means two criteria are features in disguise. ### Should criteria be yes/no or a score? Yes/no when the fact is binary, such as “needs an engineer.” A score when the fact is a range, such as time to first ticket. Do not score a binary fact from 1 to 5. ### Can I change criteria after publishing? Yes, if you rescore every product and say what changed. Silent edits to the ruler make old comparisons unreproducible. ### What if two products tie? Say they tie on these weights, and name the criterion that would break the tie. A forced winner is less useful than a tie with a lever. ## Elsewhere on this site - [Today](https://listedpress.com/) - [Trending](https://listedpress.com/trending) - [Categories](https://listedpress.com/categories) - [Tags](https://listedpress.com/tags) - [Search](https://listedpress.com/search) - [Pricing](https://listedpress.com/pricing) - [Get on a shortlist](https://listedpress.com/submit) - [Get published](https://listedpress.com/get-editorial-backlinks) - [Editorial publications](https://listedpress.com/editorial-backlinks) - [About](https://listedpress.com/about) - [FAQ](https://listedpress.com/faq) - [Free tools](https://listedpress.com/tools) - [Blog](https://listedpress.com/blog) - [Contact](https://listedpress.com/contact) - [Privacy policy](https://listedpress.com/privacy) - [Terms of use](https://listedpress.com/terms) - [Disclaimer](https://listedpress.com/disclaimer) - [Editorial policy](https://listedpress.com/editorial) - [llms.txt](https://listedpress.com/llms.txt)