SAMM Assessment Methodology (process)¶
| Status | Template. Adapt it and have it approved in your organization. |
| Owner | AppSec lead. Operated by Security Governance. |
| Calculation | SAMM/ASVS mapping section A.3, config/maturity-methodology.yaml, targets config/samm-targets.yaml |
This document covers how assessments are run. The calculation is defined in the SAMM/ASVS mapping document and the configuration above.
1. Assessment types¶
| Type | Scope | Frequency | Assessors | Evidence |
|---|---|---|---|---|
| Continuous automated | Every application | Continuous | Platform | C1–C3 control states |
| Application assessment | One application (non-automatable activities only) | Annually: Tier 1 (M2), Tier 2 (M1 + AppSec sample), Tier 3 (sampled 10%) | Champion + AppSec engineer | M1/M2 |
| Organisational assessment | Governance practices (G-SM, G-PC, G-EG) and programme-level parts of others | Annually | Governance analyst + AppSec lead. External reviewer optional. | M2 |
| Baseline | First applications + organisation | At adoption | AppSec | M2 |
2. Application assessment procedure¶
- Plan (platform creates the assessment 30 days before due): scope = the application, its repositories and the target profile for its tier.
- Pre-fill: the platform pre-populates every SAMM question that has mapped controls with the automated answer, read-only, with the control states and evidence behind it.
- Interview (60–90 min): champion + tech lead + AppSec assessor answer only the remaining questions, using the official SAMM question text and answer options. Each answer needs a rationale. Evidence references are expected for answers ≥ 0.5.
- Cap: where a manual answer exceeds the automated answer by more than the tolerance, it is capped and flagged ("evidence does not support this answer").
- Submit and approve: approver ≠ lead assessor. M2 needs an approver outside the team. Approval creates
assessment.samm_interviewevidence (valid 365 days). - Improvement plan: for each practice below target, the team records 1–3 improvement actions (Jira-linked). The platform suggests actions from the gaps in mapped controls.
3. Organisational assessment procedure¶
Document review (charter, standards, KPI packs, training records, champion programme) → interviews with Security Governance, AppSec, two business unit leads and three champions → answers for Governance practices plus programme-level parts of other practices → Board review → published as an organisation-scope assessment.
4. Quality controls¶
| Control | Detail |
|---|---|
| Official text only | Questions and options are imported verbatim. Never paraphrased in the UI. |
| Calibration | Quarterly: AppSec manually checks 5% of automated SATISFIED states (EVD-04 ≥ 97%) |
| Assessor consistency | Annual calibration workshop. Two assessors double-score three apps and compare. |
| Methodology change | Recorded in the snapshot. Trend charts mark discontinuities. |
| Use restrictions | Scores are reported as distributions with provenance. No league tables. Never used for individual performance (charter constraint). |
5. Outputs¶
Application maturity report (practice/stream scores, target comparison, provenance, improvement actions) · business unit roll-up (distribution, % meeting target) · organisation report (Governance practices plus application distributions, trend).