Cloud Security Compare

Cloud security scoring methodology

Last full review:

Every platform is assessed against the same seven criteria using public vendor material. Scores are editorial assessments, not hands-on test results. The weights are fixed before reviewing vendor outcomes and are applied identically to every vendor.

Criteria and weights

The seven criteria, their weights and what we look for.
CriterionWeightWhat we look for
Agentless coverage and time to value18%How much of the estate (VMs, containers, serverless, PaaS, AI services) is visible without installing anything, across how many clouds, and how fast.
Runtime protection depth16%Detection and prevention inside running workloads: sensor technology, blocking, investigation, response.
Risk prioritization and attack paths18%How findings are correlated into exploitable paths and ranked, so teams fix the few things that matter.
Code-to-cloud and AppSec14%SCA, SAST, IaC, secrets and image scanning, and tracing a cloud risk back to the code and owner that caused it.
AI workload security12%Inventory and posture for models, AI services, pipelines, agents and MCP servers, plus runtime detection for AI-specific threats.
Ecosystem and integrations12%Integrations, partner and marketplace reach, platform breadth around the CNAPP, and market footprint.
Pricing transparency10%Whether a buyer can model cost from public material before talking to sales.

The weights sum to 100%. Overall total = sum of (criterion score x criterion weight) / 100. Totals and rankings are computed from the shared scores, not entered by hand. Rankings use the exact weighted total; displayed totals have one decimal, with matching displayed totals marked as tied.

Scoring scale

Each criterion is scored from 0 to 5 with a written reason. These bands guide judgement; they are not a measured capability-to-decimal formula. A small score gap is not a measured performance difference.

The published score bands.
ScoreMeaning
4.5 to 5.0Leading. A buyer comparing on this criterion would shortlist it.
4.0 to 4.4Strong. Covers the criterion well with a named gap.
3.5 to 3.9Adequate. Present, but narrower or tied to an add-on or another product.
2.0 to 3.4Partial. Exists in limited form or only as a published unit without prices.
0 to 1.9Minimal. For pricing, 1.5 means quote only.

How sources are chosen and dated

We use official vendor product pages, technical documentation, pricing pages and release announcements relevant to each criterion. Product-specific documentation is preferred to a general platform claim. Each cited source has a link and a checked date recording when that source was read, not when the capability was introduced. Older evidence retains its actual checked date rather than inheriting a newer review date.

Vendor claims are attributed to the vendor, not presented as our measurements. Where a detail cannot be established, we say so. A failed or missing page is not evidence that a feature does not exist. Source links and checked dates appear with the vendor facts in the comparisons.

When a score can change

A previously published score can rise or fall only when a cited source supports a documented capability change relevant to that criterion. The reason must explain the change and link to its evidence. The same standard applies to every vendor and to both increases and decreases.

A newly read description of an existing capability, unmeasured breadth or uncertainty alone does not justify changing a score. Without evidence of a capability change, we retain the previous score and reason. Facts and checked dates can be refreshed without moving a score. The full review on 4 November 2026 retained the September scores after unsupported movements were reverted.

Buyer-situation picks and limitations

Buyer-situation picks apply separately published, fixed situation weights to the same criterion scores, with exact ties preserved. Compliance, small-team efficiency and rollout speed are proxies in these shortlists, not tested certifications or deployment guarantees. Public documentation establishes what a vendor describes, not how well a product will work in your environment.

See all scores and sources or read the comparison guide. To report an error, contact us with the claim and a public source.