How We Score Software
Every score on this site comes from the same six criteria, weighted the same way, applied to every product. Here are the weights, what each one measures, and what we refuse to do.
Why we publish this
A review score is worthless unless you know what produced it. Most software "reviews" score on undisclosed criteria, which makes the number decorative — it cannot be argued with, checked, or compared across articles.
So we publish the rubric. You can disagree with our weights, recalculate using your own, and reach a different answer. That is the point: the underlying criterion scores are the useful output, and the overall number is just one way of combining them.
The six criteria
Weights sum to 100 %. Each criterion is scored 0–10; the overall score is the weighted mean, rounded to one decimal place.
Value for money — 25%
Total cost at 5, 10 and 25 seats over 12 months versus what the tool actually delivers. Headline price is ignored; we model the tier you will realistically land on.
Ease of use — 20%
Time from signup to first genuinely useful outcome, and how much of the team can operate the tool without training.
Feature depth — 20%
Depth in the tool's primary job, not the length of its feature list. Capabilities locked behind enterprise tiers are scored as absent.
Integrations & API — 15%
Native connections to the tools a small business already runs, plus API quality, rate limits and webhook reliability.
Support & docs — 10%
Channels available on the plan a small business actually buys — not the enterprise tier — plus documentation quality.
Room to grow — 10%
What breaks between 5 and 50 users: permissions, performance, reporting and the size of the next pricing jump.
Why value carries the most weight
Value for money is weighted at 25 %, more than any other criterion, because for a small business a software decision is primarily a budget decision. A tool that is 15 % better and 200 % more expensive is not a better tool for a ten-person company — it is a worse one.
Crucially, we do not score value against the headline price. We model the 12-month cost at the tier that includes the capability you actually need, at 5, 10 and 25 seats. That figure is frequently double the advertised one, and modelling it changes the ranking more often than any other part of our process. The reasons are catalogued in our guide to SaaS pricing traps.
How we gather evidence
Reviews are built from four sources, and we say which one a claim rests on:
- Live trial accounts. We sign up, import realistic data and run the core workflow. Where a judgement comes from extended hands-on use, we say so.
- Vendor documentation and pricing pages. The authoritative source for tier contents, limits and prices. Every price we publish carries the date we last checked it.
- Operator interviews. Structured conversations with people running the software in real businesses, focused on what broke after month three.
- Public support material. Community forums, changelogs and status histories — useful for reliability and for finding limits vendors do not advertise.
Editorial independence
We earn money from display advertising and affiliate commissions, both disclosed on every page where they apply. Neither influences scoring. Specifically:
- No vendor sees a review before publication.
- No vendor can buy a score, a ranking position, or the removal of criticism.
- Products with no affiliate programme are included and recommended when they are the right answer — Wave and Obsidian are current examples.
- Every recommended product carries a "what doesn't work" section and a "skip it if" line. No exceptions.
Full detail on the commercial side is in our affiliate disclosure.
How often we re-check
- Pricing: quarterly, per category. Every figure on the site shows its verification date.
- Scores: annually, or immediately after a material product or pricing change.
- AI tools: quarterly, because the category moves fast enough that an undated review is misleading.
Corrections
We get things wrong. When a reader or vendor identifies a factual error, we verify it and correct within 48 hours, and we note material changes on the page rather than editing silently. Report an error through our contact page.
Questions about our scoring
Why does a product with a lower score sometimes come first in a roundup?
Because the overall score and the recommendation answer different questions. The score is a fixed weighting applied identically to every product. A roundup ranks for that article's specific reader — an article for five-person teams may put a lower-scoring tool first because setup time and free-plan quality matter more at that size than the criteria that lifted the other product's average. Where we do this, the article says so explicitly and the individual criterion scores are always shown so you can check the reasoning.
Can I recalculate scores with my own weights?
Yes, and you should if your priorities differ. Every review publishes the individual criterion scores in a breakdown chart. Multiply by your own weights and you have your own overall figure. If integrations matter far more to you than price, our ranking is not your ranking.
Do vendors get to respond before publication?
No. Vendors can correct factual errors after publication and we act on those quickly, but no vendor previews a review or influences a score. Pre-publication review is how software coverage becomes marketing.
How current are the prices on this site?
Every price shows the date it was verified, on the page, next to the figure. We re-check each category quarterly. Between checks, vendor pricing pages are authoritative — and if you find a discrepancy, telling us is the single most useful thing you can do for other readers.