How we test and rank

The short version

• We start from the job you need done, then test the tools people really choose between.
• We use the tools ourselves wherever we can, on the same tasks with the same inputs.
• Each ranking publishes its criteria, their weights and every score. The site computes the total from them, so nobody can nudge a number by hand.
• Every tool is labeled: tested hands-on, or scored from documentation.
• Each ranking links to its scoring sheet: the scale for every criterion, and the evidence and date behind every score.
• Prices and features carry the date we checked them.
• No company pays to be listed, to move up or to see a ranking before you do.

Choosing what to test

A ranking answers one question for one kind of reader, such as the best password manager for a family or the best meeting note-taker for a small team. We build the shortlist from what people actually compare: search results, the communities where people ask, and what AI assistants recommend today. Tools we considered and left out are listed at the end of the scores, with the reason.

Testing hands-on

Hands-on means we ran the tool ourselves. Each tool gets the same tasks, the same inputs and comparable settings, on the plan a typical buyer would use. We record the date, version and plan, keep the inputs and outputs, and report what failed as well as what worked.

Some tools cannot be tested fairly, for example when they are sold only to large companies or need hardware we do not have. Those are scored from official documentation, pricing pages and independent tests with a published method, and labeled Desk research on the page. A ranking is only marked Tested hands-on when every tool in it was.

How scores work

Each ranking chooses criteria that matter for its job and says what each one measures. A password manager is weighted toward security and recovery, a design tool toward output and speed. The criteria and weights are set before any tool is scored.

Every tool gets a score from 0 to 10 on every criterion. Before scoring, we write down what a 10, a 7 and a 4 look like on each criterion in terms we can observe, such as how many of 20 test logins a password manager filled. Scores follow that scale rather than an impression, so a re-check months later scores the same way.

The total is the weighted average of those scores, rounded to one decimal place. With these example weights, a tool that scores 9, 7, 8 and 6 gets 7.8:

Because every score is on the page, you can re-weight a ranking for your own needs. If price matters more to you than it does to us, the full score table under each ranking shows which tool would win.

Awards

Best overall goes to the highest total. Other awards, such as best free plan or best for teams, go to the tool that does that one job best, which can rank lower overall. A tie keeps the order we judged better for most readers, and the text says why.

Keeping rankings current

Software changes quickly. Prices carry a currency, the plan and the date checked. Fast-moving categories are re-checked at least every 90 days and the rest at least once a year. Each re-check is logged at the end of the page, even when nothing changed. A page shows Updated only after a change you would notice, such as a new price, score or order, and a wrong fact gets a dated correction at the top of the page.

Money and independence

No company pays to be included, to change its place or to review a ranking before it is published. We do not use affiliate links: links to tools are plain links. Free trials or review accounts, when we use them, are noted in the method at the end of the page.

What the labels mean

Tested hands-on

We ran the tool ourselves on the tasks described in the ranking.

Desk research

Scored from official documentation, pricing pages and independent tests with a published method, never written as if we had used it.

Editorial

Buying guides that explain how to choose, with no testing claim and no scores.

Ready to choose? See every ranking.