Why ranking fails
A ranking asks one question of every option at once — "how good is it?" — and answers with a number. The number has to fold everything you care about into one figure, so it hides exactly the trade-offs you needed to see: the option that is best on price and worst on support ends up in the middle, next to the one that is average at everything, and the ranking cannot tell you which of those you would rather live with.
Most comparison sites, Versus among them, compare two things at a time for this reason. Five options means ten pairs, and the notes are yours to keep.
The elimination method
People who decide well between many options rarely rank them. They remove them. The method is short:
- Name the one thing that would rule an option out on its own — over budget, no delivery to your country, no support in your language. Remove everything that fails it. Do not look at anything else yet.
- Look at what is left, together. With one fewer option the picture changes: a strength that made one option stand out may be shared by everything remaining, and stop mattering.
- Remove the weakest of what is left on the thing that now matters most. Look again.
- Stop when one remains, or when two remain and you would be content with either — at which point pick the cheaper one and stop thinking about it.
The step that matters is "look again". Each removal changes what the remaining options are being compared against, and a comparison that was written for five does not stay true for three.
How it works on compare⁄io
Compare up to six things — three on the free plan, six on Pro. The result page opens with Decision Elimination Mode across the top: every item is there with a remove button. Take one out and the verdict card names the winner among the items that remain, the attribute table narrows to them, and the follow-up chat knows which ones you have ruled out — ask it to compare what is left and it answers about those alone. Change your mind and put it back. Get down to one and the page says so.
The chat answers from the research the comparison was built on — the live web search that ran when you asked for it — rather than from a fresh guess, and the pages it used stay listed under the result throughout.