Appearance
MVP Canada EA analysis — WS 2 readiness review
| Source | client-resources/Unified Digital Experience - MVP Canada EA Analysis.pptx |
| Type | PPTX, 24 slides — executive readiness review |
| Origin | Canada Segment Enterprise Architecture |
| Dated | August 2026 |
| Verdict | NO GO for the October release |
What it is
The executive-facing companion to the segment assessment. Where that document assesses the ADS against standards, this one assesses the programme against its own delivery record — largely from Jira — and returns a ship verdict. It is the document an executive audience has most likely already seen, which makes it the most important one for us to be consistent with.
The verdict and scorecard
NO GO. "October is a build date, not a ship date."
| Readiness area | Status |
|---|---|
| Scope, architecture and governance | Not ready |
| Operating model | Pending |
| Legal, compliance and consent | Pending |
| BFF caching and resilience | Partial |
| API strategy and performance | Not defined |
| Environments (SIT/UAT/PROD) | At risk |
| Business continuity and DR | Not defined |
The eight executive concerns
- 11 new components. Shell, BFF, Integration API and Redis (4 platform) plus 7 BU micro-frontends. Each multiplies API load, ops burden, security surface and CI/CD gaps — "with no shared platform team funded to run them."
- No agreed North Star. The MVP is being built without a target-state architecture agreed and approved with other segments and the CAN Architecture team. "Every MVP decision risks being reversed."
- Inconsistent engineering, evidenced from Jira ticket coverage.
- Legal and compliance blanks.
- No API volumetry. Zero baseline on GB, CR, BU Mapping or AEM requests per second. Login-time fan-out will amplify legacy backend load with no signed target.
- Resilience posture partial.
- Consent not in MVP.
- Operating model needs review.
The programme discontinuity
The single most useful fact in the deck for roadmap purposes. WS 2 delivery started in Jira project CAUCE (Epic and Milestone grain). Four weeks before the review the WS 2 programme epics were cancelled and CAUEX was created as the replan, at SAFe Capability and Feature grain, with zero cross-project links between the two.
What was kept: BU product tiles, homepage and navigation, a caching placeholder, solution architecture, a design prototype (unassigned), an infra placeholder, plus a new ECID capability where only one of three tickets has an owner.
What was cancelled and not replanned: Readiness, Testing, GTM, Stabilization and Scale Planning — "no CAUEX inheritor." This is listed as the number one release risk.
What is done, and what is missing
Done or in progress: product tile feedback loops, API integrations to UI, API integration into UI, Affinity product tile implementing.
Missing, still in backlog: BFF caching and resilience SIT re-provisioning (the build itself is done); solution architecture sign-off; the caching implementation feature (unassigned, parent capability also in funnel); the NFR catalogue and target SLOs; API strategy.
The gate plan
| Gate | Condition | Target |
|---|---|---|
| G1 | Solution architecture approved, NFRs defined | end Aug |
| G2 | Privacy Office covers MVP, PIA scope | mid Sep |
| G3 | BFF and Redis in SIT | end Sep |
| G4 | Integration and performance test passed | mid Oct |
| G5 | DR drill passed → ship decision | end Oct |
The seven topic deep-dives
Each follows the same shape — what's being done, what's missing, risks and concerns.
| # | Topic | The headline gap |
|---|---|---|
| 1 | Legal, compliance and consent | Privacy Office coverage letter not requested; consent deferred post-MVP1 |
| 2 | BFF approach | Resilience layer built, but SIT re-provisioning in backlog; session and BU aggregates share one Redis |
| 3 | API volumes | No documented TPS, no per-BU rate limits, no cache-hit target, no fallback contract |
| 4 | Circuit breakers | Zero Jira tickets for breaker, retry, timeout or bulkhead; no fallback UX; no owner |
| 5 | Integration and performance | Zero tickets for integration or performance testing; no SLOs; no load environment |
| 6 | Other environments | SIT Redis pending, UAT pending, PROD not ready, no DR-region Redis anywhere |
| 7 | Business continuity and DR | Gold RTO/RPO 2h agreed; zero DR/BC tickets; no runbook; drill unscheduled |
Topic 7 carries one genuinely resolved item: ADR-001 drafted for Redis Enterprise with active geo-replication across Canada Central and Canada East, which satisfies the ≥100 km separation control. It is the only piece of DR design that exists.
The seven questions put to the team
Reproduced because they are the client's own framing of what is unresolved, and several map directly onto SOW deliverables:
- Where is the Privacy Office letter confirming existing PIAs cover MVP1 flows?
- The BFF caching and resilience build is done — why is SIT re-provisioning still in backlog with no owner?
- Where is the signed NFR catalogue? What are we performance-testing against without it?
- Where is the API strategy? What is the blocker?
- Who owns circuit breakers? No ticket, no named engineer.
- Who owns integration and performance testing? Where is the plan and environment?
- When will SIT, UAT and Prod Redis be provisioned?
Why it matters here
It tells us three things the other documents do not.
The delivery record is not reconcilable. Two Jira projects with no links between them means "what is actually built" cannot currently be evidenced. Any roadmap we produce in WS 3 inherits that problem, and reconciling CAUCE and CAUEX into one traceable view is condition 7 of the assessment. It may need to be an explicit early roadmap item.
The cancelled operational half is the real schedule risk. Readiness, Testing, GTM, Stabilization and Scale Planning were cancelled without an inheritor. Those are exactly the workstreams a transformation roadmap has to carry, and they are currently owned by nobody.
"No agreed North Star" is stated by the client, in an executive deck, as concern #2. That is the strongest possible mandate for the WS 2 deliverable, and it means our target state architecture is not a nice-to-have artifact — it is the thing named as missing.
Open questions raised
- Has the October verdict changed since this deck was presented?
- Which project now owns Readiness, Testing, GTM, Stabilization and Scale Planning?
- Is the deck's Gold RTO/RPO of 2h reconciled with the assessment's finding (A-05) that the tier should be Critical Digital Properties at 5 min?
- Does this engagement's target state deliverable become the North Star of record, and who approves it?