Independent reference. Not affiliated with any zero trust vendor. Updated Q1 2026.
ZeroTrustCost
Vendor framework

Zero trust vendor evaluation - vendor-neutral framework

A framework for evaluating zero trust platforms without falling into the vendor-marketing trap. This page does not list specific dollar pricing, those numbers are negotiated, regional, and rarely match list. Instead it covers vendor categories, what each is best at, deployment complexity, lock-in risk, and how to map vendors to your scenario. For pricing, request quotes.

Categories

Six vendor categories cover the zero trust market

Most platforms fall into one of six archetypes. Each is strong in one or two pillars and weaker in others. Recognising the archetype is the first step in evaluation.

Identity-centric platforms

Identity
Primary
Network
Secondary
Device
Secondary
Workload
Not core
Data
Not core
Deployment
Moderate. Heavy on directory integration, SSO migration, conditional access rule design.
Best for
Heterogeneous estates not committed to a single productivity suite. Multi-cloud SaaS-heavy organisations.
Lock-in risk
High - identity is the deepest integration point in any zero trust stack. Switching IdPs is the most expensive vendor migration.
FedRAMP
Major identity vendors typically offer FedRAMP Moderate at minimum, FedRAMP High variants exist for federal estates.

Microsoft-suite platforms

Identity
Primary
Network
Secondary
Device
Primary
Workload
Secondary
Data
Secondary
Deployment
Lowest if already on M365. Bundle integration is largely click-through. Highest if migrating from non-Microsoft identity.
Best for
Microsoft 365-centric organisations. Mid-market and enterprise already paying for E3 / E5.
Lock-in risk
Very high - bundle economics make leaving expensive. M365 commercial dependence creates indirect identity lock-in even if technically portable.
FedRAMP
Multiple FedRAMP authorisations including High tiers for government estates.

Network-centric / SSE platforms

Identity
Not core
Network
Primary
Device
Secondary
Workload
Secondary
Data
Secondary
Deployment
High. Network platform implementations require connector deployment in every private app environment, parallel VPN running, and policy migration from IP-based to identity-based rules.
Best for
Large enterprises replacing complex on-premise proxy and VPN infrastructure. Heavy private-app estates.
Lock-in risk
Moderate. ZTNA can be migrated more easily than identity, but SSE bundles (ZTNA + SWG + CASB + FWaaS) create platform-level lock-in once policy depth grows.
FedRAMP
Most major SSE platforms have FedRAMP Moderate authorisations; some have High.

Lightweight ZTNA platforms

Identity
Not core
Network
Primary
Device
Not core
Workload
Not core
Data
Not core
Deployment
Lowest. Modern ZTNA-only platforms deploy in days for SMBs, weeks for mid-market.
Best for
SMBs and developer-centric organisations. ZTNA replacement of legacy VPN without need for full SSE.
Lock-in risk
Low. Standards-based, easy to swap. Pricing is transparent.
FedRAMP
Variable. Most consumer-focused ZTNA platforms do not pursue FedRAMP. Enterprise variants increasingly do.

Endpoint-centric platforms

Identity
Secondary
Network
Not core
Device
Primary
Workload
Secondary
Data
Not core
Deployment
Moderate. EDR agent deployment, exclusion tuning, MDR onboarding for managed-response variants.
Best for
Estates where endpoint visibility is the priority concern. Organisations with mature SOC operations.
Lock-in risk
Moderate. Agent-based architecture creates operational dependency. Switching requires removing one agent and deploying another, with parallel-run period.
FedRAMP
Major endpoint vendors typically have FedRAMP Moderate; High tier available from select vendors.

Cloud-native / CNAPP platforms

Identity
Not core
Network
Secondary
Device
Not core
Workload
Primary
Data
Secondary
Deployment
Moderate. API-based deployment, cloud account integration, IaC scanning hookups.
Best for
Cloud-native organisations. AWS / Azure / GCP-heavy estates with significant container or Kubernetes footprint.
Lock-in risk
Low to moderate. Most CNAPP platforms support all major clouds and standards-based output formats.
FedRAMP
Variable. CNAPP is a newer category, FedRAMP authorisations are still growing.
Scenario fit

Which category goes where

Map vendor categories to the scenarios most CISOs face. Multi-vendor combinations are normal and often optimal, single-vendor zero trust is rare outside the Microsoft ecosystem.

ScenarioIdentityNetworkDeviceWorkloadData
100-user SMB on M365Microsoft suiteLightweight ZTNAMicrosoft suite (Defender)Native cloud toolingMicrosoft suite (Purview)
500-user mid-market, Google WorkspaceIdentity-centric platformLightweight ZTNA or SSEEndpoint-centric platformCNAPPCASB layer
2,000-user enterprise, Microsoft-centricMicrosoft suiteSSE platform or Microsoft EntraMicrosoft suite or endpoint specialistCNAPPMicrosoft suite (E5) or specialist DLP
Federal contractor, CMMC L2FedRAMP-High identity platformFedRAMP-authorised SSEFedRAMP endpoint platformFedRAMP CNAPPFedRAMP DLP / classification
Cloud-native dev-led estateIdentity-centric platformLightweight ZTNACloud-managed MDMCNAPP (primary investment)Cloud-native KMS + selective DLP
Heavy regulated data (healthcare, finance)Identity-centric platform with PAMSSE with microsegmentationEndpoint-centric with MDRCNAPPSpecialist DLP + classification (heavy investment)
Decision framework

Eight dimensions to score every vendor

  • Pillar coverage depth. For each of the five pillars, score primary / secondary / not core. Bolt-on modules count as secondary at best.
  • Deployment complexity. Average rollout duration for organisations your size. Reference customers of comparable scale are essential here.
  • Existing-stack fit. Native integration depth with your IdP, SIEM, ITSM, HR system. SCIM provisioning, SAML / OIDC SSO, OPA policy, OpenTelemetry export.
  • Lock-in risk. Standards-based config export, contract-level data extract rights, switching cost as percentage of new platform first-year licensing.
  • FedRAMP / compliance authorisation. FedRAMP Moderate or High, ISO 27001, SOC 2 Type II, HIPAA, PCI DSS. Required for regulated estates.
  • 3-year total cost. Licensing + professional services + integration + tuning over 3 years. Year-one licensing is often the smallest line.
  • Vendor stability. Financial position, product roadmap clarity, M&A risk, leadership turnover. Public companies disclose more; private vendors require careful diligence.
  • Reference customer fit. Three references of comparable size, sector, and existing-stack composition. Phone calls, not case studies.
Frequently asked

Vendor evaluation questions

Why does this page not list specific vendor pricing?
Three reasons. First, vendor-disclosed list pricing is rarely the price organisations actually pay. Enterprise contracts negotiate 20-40% discounts off list, multi-year terms, and bundle adjustments that make published pricing misleading. Second, regional pricing varies meaningfully (US vs EU vs APAC) and any single number misrepresents the others. Third, this site is independent and explicitly avoids fabricating numbers we cannot stand behind. The vendor-neutral framework is more useful: it tells you which vendor categories fit which scenarios, and which dimensions to negotiate on. For specific pricing, request published rate cards or partner quotes.
How do we evaluate a zero trust platform vendor?
Eight dimensions matter. (1) Pillar coverage - which of the five pillars does the platform genuinely cover, and which is bolt-on? (2) Deployment complexity - how long is the average rollout for an organisation your size? (3) Existing-stack fit - does the platform integrate cleanly with your IdP, SIEM, ITSM, HR system? (4) Lock-in risk - is policy configuration extractable in standard formats? (5) FedRAMP / compliance authorisation if regulated. (6) Total cost (licensing + services + integration + tuning over 3 years), not just year-one licensing. (7) Vendor financial stability and product roadmap clarity. (8) Reference customer fit - request references from organisations of comparable size, sector, and existing-stack composition.
Is best-of-breed or single-vendor better?
Single-vendor (typically Microsoft-suite) is cheaper in licensing, faster to deploy, and easier to operate for organisations already committed to that vendor's productivity suite. Best-of-breed is more expensive in licensing and integration but gives you better-in-class capability per pillar and lower vendor lock-in. The practical answer for most mid-market: single-vendor for identity and device pillars (where bundle economics dominate), best-of-breed for network (where pure ZTNA platforms can be substantially cheaper than bundled SSE) and workload (where CNAPP specialists outperform generic security suites). Pure best-of-breed multi-vendor is best for organisations with mature security teams who can absorb the integration overhead.
How important is FedRAMP authorisation?
Mandatory if you are a federal contractor or process federal data. Important if you intend to bid on federal contracts within 2-3 years (the procurement cycle requires authorised tools to be already in place). Largely irrelevant for non-government estates. FedRAMP authorisations restrict the vendor pool to roughly the top 8-12 zero trust platforms in each category. FedRAMP Moderate is the minimum tier; FedRAMP High is required for higher-sensitivity federal data. Authorisation is a moving target, validate current status with each vendor.
Should we issue a formal RFP?
Yes for purchases above $250K annual licensing or any platform that will run for 3+ years. RFPs force vendors to commit in writing to the dimensions that matter (deployment time, integration depth, FedRAMP status, support response SLAs) and produce a comparable evaluation across vendors. The RFP should explicitly require: pillar-by-pillar coverage attestation, three reference customers of comparable size, deployment timeline guarantee, contract-level lock-in protections (data extract on exit), and total-cost-of-ownership pricing including services and integration. Skip RFPs for purchases under $50K annual or for tooling that will be replaced within 2 years.
What is the worst evaluation mistake?
Pillar conflation. Vendors that primarily address one pillar often claim coverage of all five through bolt-on modules. The bolt-ons are typically immature compared to dedicated platforms in those categories. The worst evaluation mistakes confuse breadth with depth. A network-centric SSE platform may technically include identity through an OEM SSO, but the identity capability will not match a dedicated identity platform. A Microsoft suite includes ZTNA via Entra Private Access, but the ZTNA capability is newer than dedicated ZTNA vendors. Evaluate each pillar separately and pick the right tool for each, even if it means multi-vendor.