Editorial
Our rating methodology
The seven criteria behind every score, how each is weighted, and what a given number actually means.
Last updated 29 August 2026
Every product we test is scored out of 10 on seven criteria. The overall score is a weighted average, and we publish the weighting so you can disagree with it.
The weighting
| Criterion | Weight | What we are judging |
|---|---|---|
| Features | 25% | Does it do the job the category exists to do, without workarounds |
| Value for money | 20% | Total cost at realistic usage, including renewal, against what you get |
| Ease of use | 15% | Time from signup to first useful result, for the intended buyer |
| AI quality | 15% | How the AI performs on hard cases, not demo cases |
| Integrations | 10% | Whether it connects to the tools that category's buyers already run |
| Customer support | 10% | Response time and usefulness, tested with a real question |
| Security | 5% | Certifications, data handling, and what happens to your data if you leave |
What the numbers mean
- 9.0 and above — we would recommend it without hesitation to its intended buyer
- 8.0 to 8.9 — strong, with trade-offs we name on the page
- 7.0 to 7.9 — works, but something better usually exists at the price
- 6.0 to 6.9 — only worth it in a specific situation we will describe
- Below 6.0 — we would not buy it
Category weighting can differ
Security carries more weight in accounting software than in a logo generator. Where a category uses adjusted weights, we state it on the category page.
What the score does not include
Commission rates, affiliate relationships, advertising spend, and vendor requests. None of these are inputs.