In brief
- Every product is scored 1–10 on six criteria, each with written definitions for what a 10, 8, 6, 4 and 2 look like.
- Criteria are weighted by product type — security counts most for security products, ease and value for business software.
- The overall score is calculated from the weighted criteria by code. Nobody types it in.
- Each score shows its evidence and a confidence level, the date it was set, and when it will be re-checked.
- Money never touches the score: no paid placements in rankings, no vendor previews, and affiliate status plays no part in the formula.
The six criteria
Each criterion answers one question. Editors score against the anchor definitions below; scores between anchors (for example 7.5) mean the product sits between the two descriptions.
Features & fit
Does it do the core job well for small and mid-sized teams?
What we look at: Coverage of the category's must-have features; Depth of features on affordable tiers; Fit for teams of 1–250 people; Product roadmap and pace of improvement.
| Score | Label | What it means |
|---|---|---|
| 10 | Category-defining | Covers every core and advanced need in the category with depth that competitors measure themselves against. |
| 8 | Excellent | Covers all core needs and most advanced needs; gaps are minor or niche. |
| 6 | Solid | Covers core needs; advanced features are missing, shallow or locked behind expensive tiers. |
| 4 | Limited | Misses at least one core need most buyers expect. |
| 2 | Inadequate | Does not reliably do the core job. |
Ease of use
How quickly can a non-specialist get value from it?
What we look at: Time to first useful result; Clarity of navigation and defaults; Onboarding, templates and in-app help; Admin effort to keep it running.
| Score | Label | What it means |
|---|---|---|
| 10 | Effortless | Most users are productive within an hour with no training; admin is near zero. |
| 8 | Easy | Productive within a day; occasional help docs needed for advanced features. |
| 6 | Manageable | Needs deliberate setup or a short training session; some confusing areas. |
| 4 | Steep | Needs a dedicated admin or consultant to set up and maintain. |
| 2 | Hostile | Routinely blocks users from completing basic tasks. |
Value for money
Is the price fair for what a small business actually gets?
What we look at: Usefulness of the free plan or entry tier; Cost at 5, 10 and 25 seats; Renewal and tier-jump surprises; Add-ons needed for core features.
| Score | Label | What it means |
|---|---|---|
| 10 | Exceptional | Generous free plan or entry tier; transparent pricing with no meaningful upsell traps. |
| 8 | Good | Fair pricing; key features available on mid tiers; renewals predictable. |
| 6 | Average | Priced in line with the market; some important features gated behind costly tiers or add-ons. |
| 4 | Poor | Expensive for what it delivers, or significant renewal increases and hidden costs. |
| 2 | Very poor | Pricing is opaque or out of proportion to the value delivered. |
Support & reliability
Will help be there, and will the product stay up?
What we look at: Support channels and hours on affordable tiers; Quality of documentation and community; Published status history and uptime commitments; Consistency of patterns in public user feedback.
| Score | Label | What it means |
|---|---|---|
| 10 | Outstanding | Fast, expert, multi-channel support on every tier; excellent documented reliability. |
| 8 | Strong | Responsive support on paid tiers, strong docs and a good reliability record. |
| 6 | Adequate | Support is slow or limited on lower tiers; documentation fills most gaps. |
| 4 | Weak | Support hard to reach; recurring reliability complaints. |
| 2 | Absent | Little or no support; frequent outages. |
Security & privacy
Does it protect your data and give you control over it?
What we look at: MFA/SSO and role-based access on affordable tiers; Encryption and independent audits or certifications; Data handling, retention and AI-training policies; Security incident history and transparency.
| Score | Label | What it means |
|---|---|---|
| 10 | Best-in-class | Strong defaults, independent audits, SSO on accessible tiers and transparent incident handling. |
| 8 | Strong | MFA, encryption and admin controls as standard; audits published; minor gaps such as SSO only on top tiers. |
| 6 | Acceptable | Basic controls in place; limited admin visibility or unclear data policies. |
| 4 | Concerning | Missing basic controls such as MFA, or an unresolved pattern of incidents. |
| 2 | Unsafe | Material security or privacy risks for business data. |
Integrations & ecosystem
Does it connect to the rest of your stack?
What we look at: Native integrations with popular tools; API availability and documentation; Automation-platform support (Zapier, Make and similar); Import/export and data portability.
| Score | Label | What it means |
|---|---|---|
| 10 | Ecosystem hub | Huge native marketplace, excellent API and effortless data portability. |
| 8 | Well connected | Native integrations with the tools most small businesses use, plus a documented API. |
| 6 | Adequate | Key integrations exist but others need an automation platform or workaround. |
| 4 | Isolated | Few integrations; exporting data is awkward. |
| 2 | Closed | No meaningful integrations or exports. |
Weights by product type
Weights are fixed per pillar so that every product in a category is judged on the same basis. Each row adds up to 100%.
| Pillar | Features | Ease | Value | Support | Security | Integrations |
|---|---|---|---|---|---|---|
| Business Software | 25% | 20% | 20% | 10% | 10% | 15% |
| AI Tools | 25% | 15% | 15% | 10% | 20% | 15% |
| Security | 25% | 15% | 15% | 10% | 30% | 5% |
| Web Hosting | 20% | 15% | 20% | 20% | 15% | 10% |
- Business Software: Business software lives or dies on adoption, so ease of use and value carry extra weight alongside features.
- AI Tools: AI tools handle sensitive prompts and documents, so security & privacy (including training-data policies) is weighted heavily.
- Security: A security product that is not itself secure is worthless, so security & privacy is the single biggest weight.
- Web Hosting: Uptime and help when things break matter as much as features, so support & reliability is weighted up.
How the overall score is calculated
The overall score is the weighted average of the six criterion scores, rounded to one decimal place:
Overall = Σ (criterion score × criterion weight) ÷ 100
The calculation runs in our site code at build time, from the sub-scores stored with each review. It validates that every sub-score is between 1 and 10 and that each pillar’s weights total 100%. Editors set sub-scores only — there is no field for an overall score, so it cannot be adjusted by hand to favour a product.
Worked example: HubSpot CRM
Here is the live breakdown for HubSpot CRM. Each row shows the sub-score, its weight and the points it contributes to the overall score.
| Criterion | Weight | Score | Points | |
|---|---|---|---|---|
| Features & fit | 25% | 9.0 | 2.30 | |
| Ease of use | 20% | 9.3 | 1.90 | |
| Value for money | 20% | 8.0 | 1.60 | |
| Support & reliability | 10% | 8.6 | 0.90 | |
| Security & privacy | 10% | 8.4 | 0.80 | |
| Integrations & ecosystem | 15% | 9.4 | 1.40 | |
| Overall | 100% | Weighted for business software products | 8.8 | |
What scores mean
| Overall score | Label |
|---|---|
| 9.0+ | Outstanding |
| 8.0–8.9 | Excellent |
| 7.0–7.9 | Good |
| 6.0–6.9 | Fair |
| 0.0–5.9 | Weak |
A lower overall score does not mean a product is wrong for you. Our “best for” labels and “Not for you if” sections show where each product fits.
Evidence & confidence
Every review lists the evidence behind its score. Each type of evidence contributes points; the total sets the confidence level shown on the review.
| Evidence | What it involves | Points |
|---|---|---|
| Hands-on testing | An editor used the product on a real or trial account against our test script. | 3 |
| Vendor demo | A guided demo or briefing with the vendor, cross-checked against documentation. | 1 |
| Documentation review | Public product docs, help centre, security pages and changelogs. | 1 |
| Pricing check | Current public pricing pages, plan limits and renewal terms, recorded with a date. | 1 |
| Verified user signals | Recurring themes in verified-buyer reviews on independent platforms. We look for patterns, never quote star ratings as our score. | 1 |
| Practitioner input | Feedback from consultants, accountants or IT providers who deploy the product for clients. | 1 |
| Confidence | Points needed | Meaning | Reviews today |
|---|---|---|---|
| High | 4+ | Hands-on testing plus at least one corroborating source. | 0 |
| Medium | 2+ | Multiple independent desk-research sources; no hands-on test yet. | 112 |
| Low | 0+ | Limited evidence — treat the score as provisional. | 0 |
Re-review cadence
- Business Software: re-checked at least every 6 months.
- AI Tools: re-checked at least every 3 months.
- Security: re-checked at least every 6 months.
- Web Hosting: re-checked at least every 6 months.
Reviews are also re-checked when a vendor changes pricing, launches or removes a plan, or has a significant security incident. Pricing is recorded with the date it was checked, and every review shows when it was last verified and when it is next due.
Independence rules
- Affiliate commissions never influence scores. Scores are set before commercial status is checked, and the score formula has no input for it.
- Sponsored placements and advertising are always labelled and are kept separate from rankings, scores and “best” lists.
- Vendors can correct factual errors (pricing, features) but cannot see, approve or change a score before publication.
- Editors do not review products from companies they own shares in or have worked for in the past two years.
- We do not accept payment, free upgrades or gifts in exchange for coverage.
- When a score changes, the date and reason are recorded in the review's changelog.
Read more in our editorial policy and affiliate disclosure.
Methodology changelog
- v2.0 — Moved to six weighted criteria with per-pillar weights, evidence-based confidence levels and computed overall scores. Added Security & privacy and Integrations criteria.
- v1.0 — First published scoring framework: ease of use, features, value and support.
Frequently asked questions
Why publish the methodology?
So you can check our work. A score is only useful if you know what it measures. Publishing the rubric, weights and evidence rules lets readers, vendors and other reviewers hold us to the same standard on every review.
Is the overall score an average?
It is a weighted average of the six criterion scores, using the weights for that product's pillar. It is calculated by code from the sub-scores — editors cannot type in an overall score.
Can a vendor pay for a better score?
No. The formula has no commercial input, vendors cannot see scores before publication, and affiliate status is checked only after a score is set.
What does “Researched” mean?
The score is based on desk research — documentation, dated pricing pages and other sources — not an extended hands-on test. Such scores carry Medium confidence. Hands-on tested reviews are labelled and carry High confidence.
How often are scores updated?
Every review is re-checked on a fixed cadence (every 3 months for AI, 6 months for other categories) and sooner when a vendor changes pricing or ships a major release. The next review date is shown on every review.
Why do weights differ between categories?
Because what matters differs. Security & privacy is the most important criterion for a password manager, while ease of use matters more for a project management tool. Weights are fixed per pillar so every product in a category is judged the same way.
Spotted something we got wrong? See how to contact the editorial team. For a shorter overview, read how we research and test.