1.3 Deconstructing GCP Case Studies & Scenario Questions
Key Takeaways
- Each PCA exam covers 2 of the 4 published case studies — EHR Healthcare, Cymbal Retail, Altostrat Media, and KnightMotives Automotive — comprising roughly 20-30% of exam questions.
- Every case study consists of five structural elements: Company Overview, Solution Concept, Technical Requirements, Business Requirements, and Executive C-Suite Directives.
- The optimal exam technique is 'Question-First Deconstruction': read the question stem and answer choices before referencing the case study split-screen to pinpoint the exact governing constraint.
- Distractor answer choices often provide valid technical configurations that fail specific business constraints, such as regulatory compliance (HIPAA/PCI-DSS), budget limits, or operational simplicity.
- Identifying legacy anti-patterns (e.g., custom cron scripts, self-hosted MySQL on Compute Engine, public IP egress for private services) allows rapid elimination of incorrect options.
Deconstructing GCP Case Studies & Scenario Questions
The presence of in-depth, enterprise-scale Case Studies is the defining characteristic of the Google Cloud Professional Cloud Architect exam. Case study questions do not test isolated product facts; they test your ability to synthesize business requirements, technical requirements, regulatory constraints, and executive statements into an integrated cloud architecture.
On the live exam, case study questions represent approximately 10 to 15 questions (roughly 20% to 30% of the exam). Understanding the structural patterns of the case studies and mastering a systematic decomposition technique will save you valuable time and prevent costly misinterpretations.
The Four Current Official Case Studies
Google Cloud periodically refreshes its case studies. The current exam guide publishes exactly four case studies, and your exam form will draw 2 of the 4, so you must prepare all of them. Legacy prep materials built around Mountkirk Games, TerramEarth, Helicopter Racing League, or Dress4Win describe retired case studies that no longer appear on the exam; use them only as generic pattern practice.
+-------------------------------------------------------------------------+
| PCA CURRENT CASE STUDY MATRIX (v6.1) |
+---------------------+---------------------+-----------------------------+
| Case Study | Industry Vertical | Primary Architecture Focus |
+---------------------+---------------------+-----------------------------+
| EHR Healthcare | Healthcare SaaS | Hybrid migration, K8s |
| | | consistency, 99.9% avail. |
| Cymbal Retail | Online Retail | Gen-AI catalog enrichment, |
| | | conversational commerce |
| Altostrat Media | Media / Content | Gen-AI engagement, content |
| | Publishing | moderation, summarization |
| KnightMotives | Automotive / | AI in-vehicle UX, hybrid |
| Automotive | Connected Vehicles | legacy modernization, EU |
| | | data privacy, data sales |
+---------------------+---------------------+-----------------------------+
1. EHR Healthcare (Regulated SaaS & Hybrid Modernization)
- Profile: A SaaS provider of electronic health record software to multinational medical offices, hospitals, and insurance providers. Applications run in colocation facilities — one data center lease is expiring — and Google Cloud has been chosen to replace the colos.
- Core Business Goals: On-board new insurance providers quickly, provide minimum 99.9% availability for all customer-facing systems, gain centralized visibility into performance and usage, produce insights and predictions on healthcare trends, reduce customer latency, maintain regulatory compliance, and decrease infrastructure administration cost.
- Core Technical Challenges: Legacy file- and API-based insurance integrations stay on-premises for the foreseeable future; databases are a mix of MySQL, MS SQL Server, Redis, and MongoDB; users live in Microsoft Active Directory; monitoring is fragmented open-source tooling with email alerts that staff ignore.
- Standard GCP Architecture Patterns: Dedicated or Partner Cloud Interconnect (with HA Cloud VPN for backup or low-volume links) for secure hybrid connectivity; GKE plus GKE Enterprise / Anthos to manage the multiple container environments consistently across cloud and on-premises; Pub/Sub + Dataflow pipelines to ingest new provider data feeds; BigQuery for healthcare-trend analytics; consolidated Cloud Monitoring / Cloud Logging with real alert policies; Cloud Build + Cloud Deploy for the requested continuous deployment cadence.
2. Cymbal Retail (Generative AI Digital Commerce)
- Profile: A fast-growing online retailer with a large product assortment across several retail sub-verticals and a catalog that is painful to manage manually.
- Core Business Goals: Three modernization tracks: (1) Catalog and Content Enrichment — use generative AI to derive product attributes, descriptions, and images from supplier-provided information; (2) Conversational Commerce with Product Discovery — AI virtual agents on the website and mobile app that understand natural language and surface the most relevant products; (3) Technical Stack Modernization — cloud infrastructure, secure data handling, third-party integrations, and proactive monitoring. Explicit cost goals include reducing call-center staffing and data-center hosting spend.
- Core Technical Challenges: Hybrid on-premises/cloud estate; MySQL, MS SQL Server, Redis, and MongoDB databases; Kubernetes clusters; legacy SFTP file transfers and ETL batch jobs; a custom web app querying relational databases; an IVR phone tree with human agents re-keying orders; fragmented Grafana/Nagios/Elastic monitoring.
- Standard GCP Architecture Patterns: Vertex AI (including Imagen-class image models) for image generation and enhancement, Vision AI / Natural Language AI / Document AI for attribute extraction, Vertex AI Search for product discovery, Conversational Agents to replace the IVR, a human-in-the-loop review UI so associates approve/reject generated content, BigQuery for unified analytics, and consolidated Cloud Monitoring replacing the open-source patchwork.
3. Altostrat Media (Media Publishing & Generative AI Engagement)
- Profile: A media company with a vast library of podcasts, interviews, news broadcasts, and documentaries. Already substantially on Google Cloud: GKE, Cloud Storage, BigQuery, and Cloud Run functions (event-driven transcoding and metadata extraction), plus some legacy on-premises ingestion/archival systems slated for modernization.
- Core Business Goals: Use generative AI for personalized recommendations, natural-language audience interaction, 24/7 self-service support, automated content summaries, and metadata extraction; detect and filter inappropriate content; drive revenue through dynamic pricing and targeted marketing — while optimizing storage cost and workflow reliability.
- Core Technical Challenges: Modernize CI/CD for containerized deployments on a centralized platform; secure, high-performance hybrid connectivity for content ingestion; scalable Kubernetes both on-premises and in cloud; rising media storage costs; AI systems that are auditable and explainable; securing generative AI endpoints.
- Standard GCP Architecture Patterns: Conversational Agents for natural-language self-service, Natural Language AI / Video AI / Vision AI for summarization and metadata, Vertex AI custom models for harmful-content detection with Vertex Explainable AI, Model Armor and Sensitive Data Protection for generative AI safety, Cloud Build + Cloud Deploy for centralized CI/CD, Anthos/GKE Enterprise for on-prem + cloud Kubernetes, and Cloud Storage Autoclass/lifecycle rules for cost optimization.
4. KnightMotives Automotive (Connected & Autonomous Vehicles)
- Profile: A global car manufacturer (BEVs, hybrids, and ICE vehicles) that must modernize the in-vehicle experience across all models within five years. Its unreliable build-to-order online ordering system is straining dealer relationships, and dealers have no budget for new equipment.
- Core Business Goals: Personalized, consistent AI-powered driver experience; transparent build-to-order commerce; monetize corporate data to fund AI investment (current AI infrastructure is obsolete and data is siloed); security is paramount after past data breaches; comply with EU data protection regulations for autonomous platforms; invest in autonomous driving in favorable regulatory regions; upskill employees.
- Core Technical Challenges: Outdated mainframe supply chain and ERP systems; fragmented codebases across vehicle lines; rural network connectivity gaps for vehicles; hybrid on-premises/cloud estate requiring gradual, not big-bang, modernization.
- Standard GCP Architecture Patterns: Vertex AI with Cloud TPUs for autonomous-vehicle model training and simulation, partner MQTT broker feeding Pub/Sub (Cloud IoT Core was retired, so no managed IoT registry exists), Dataflow + Cloud Bigtable for vehicle telemetry time series, BigQuery as the governed data-monetization warehouse, Apigee for dealer and partner APIs, Network Connectivity Center with Cloud Interconnect / SD-WAN partner appliances for plant connectivity, GKE Enterprise (Anthos) for hybrid modernization, and Security Command Center + Sensitive Data Protection + Assured Workloads / resource-location Organization Policies for breach recovery and EU compliance.
Structural Anatomy of a Case Study
Every official Google Cloud case study follows an identical, standardized five-part structure:
- Company Overview: Establishes the company's background, current industry standing, and high-level mission.
- Solution Concept: Outlines the envisioned future-state cloud architecture and desired organizational transformations.
- Existing Technical Environment: Describes the legacy infrastructure, hardware constraints, database engines, network connections, and deployment tooling currently in production.
- Business Requirements: Lists explicit non-functional and commercial drivers, such as reducing licensing costs, speeding up global market expansion, and complying with industry standards.
- Technical Requirements: Details hard operational specifications, such as multi-region failover, latency ceilings, data residency rules, encryption mandates, and scaling targets.
- Executive Statements: Quotations from the CEO, CTO, CFO, or CISO that highlight high-priority organizational imperatives.
The 5-Step Systematic Deconstruction Strategy
When tackling case study questions under strict exam time limits, never read the entire case study document from start to finish on your first question. Instead, apply this proven 5-step workflow:
┌─────────────────────────────────────────────────────────────────────────┐
│ 5-Step Deconstruction Workflow │
└─────────────────────────────────────────────────────────────────────────┘
│
[Step 1: Read Question Stem & 4 Choices First (Isolate the problem)]
│
[Step 2: Identify Governing Constraint Category (Cost, Sec, Latency)]
│
[Step 3: Scan Split-Screen Text for Relevant Requirement Bullet]
│
[Step 4: Eliminate Distractors Violating Stated Constraints]
│
[Step 5: Select the Google-Recommended Best Practice Pattern]
Step 1: Read the Question Stem and Answer Options First
Before glancing at the case study text on the left panel, read the specific question prompt on the right panel. The question will usually target a single subsystem (e.g., "How should Cymbal Retail automate its product catalog enrichment?" or "How should EHR Healthcare connect its on-premises systems to Google Cloud securely?").
Step 2: Identify the Governing Constraint Category
Determine the primary evaluation dimension:
- Is it a Security/Compliance question? (Look for HIPAA, PCI-DSS, CMEK, VPC Service Controls, IAM least privilege).
- Is it a Reliability/DR question? (Look for RTO, RPO, multi-region failover, 99.99% availability).
- Is it a Cost Optimization question? (Look for Spot VMs, CUDs, serverless auto-scaling to zero, lifecycle storage tiering).
- Is it a Performance/Latency question? (Look for Cloud CDN, Anycast ALB, Bigtable low-latency writes).
Step 3: Scan the Case Study Split-Screen for the Exact Binding Requirement
Navigate directly to the Business Requirements, Technical Requirements, or Executive Statements section on the left panel to locate the specific bullets governing that subsystem. For example, if the CISO statement requires "complete control over encryption keys and key rotation schedules," you immediately know that Google-default encryption is insufficient and Cloud KMS CMEK is required.
Step 4: Eliminate Distractor Choices Violating Stated Constraints
Evaluate each option against the binding constraints identified in Step 3. Eliminate options that solve the technical problem but violate operational or business constraints (e.g., proposing self-managed MongoDB on VMs when the business requirement demands minimal administrative overhead).
Step 5: Select the Cloud-Native Recommended Pattern
Among the remaining viable options, select the one that aligns with Google Cloud Well-Architected best practices (declarative IaC, managed serverless/containers, native IAM integration, and high availability).
Dissecting C-Suite Executive Statements
Executive statements in Google Cloud case studies are not decorative flavor text; they contain vital architectural constraints that frequently distinguish the correct answer from tempting distractors:
| Executive Title | Core Priority Focus | Architectural Translation on the Exam |
|---|---|---|
| Chief Executive Officer (CEO) | Speed to market, global expansion, brand trust | Choose global services that scale instantly without multi-month capacity planning (Cloud Spanner, Global Cloud Load Balancing, Firebase). |
| Chief Technology Officer (CTO) | Cloud-native patterns, developer agility, open standards | Select open-source compatible managed services (GKE, Cloud Run, BigQuery, standard SQL/PostgreSQL) to avoid vendor lock-in and eliminate manual operational patching. |
| Chief Financial Officer (CFO) | Predictable OPEX, TCO reduction, zero idle waste | Choose autoscaling serverless architectures, Committed Use Discounts (CUDs) for baseline loads, and Spot VMs for batch pipelines. |
| Chief Information Security Officer (CISO) | Zero trust, regulatory compliance, data loss prevention | Mandate Customer-Managed Encryption Keys (CMEK), VPC Service Controls, BeyondCorp Identity-Aware Proxy (IAP), and IAM least privilege. |
Spotting Common Distractor Patterns & Legacy Anti-Patterns
Exam writers intentionally craft believable distractors. Learn to identify and eliminate these common anti-patterns on sight:
| Distractor Anti-Pattern | Why It Is Incorrect on the PCA Exam | Correct Architectural Alternative |
|---|---|---|
| "Custom Cron Scripts on a Bastion VM" | Fragile, single point of failure, high operational maintenance toil. | Cloud Scheduler triggering Cloud Tasks, Pub/Sub, or Cloud Run jobs. |
| "Self-Hosted MySQL/Postgres on GCE" | Requires manual backups, patching, OS maintenance, and custom failover scripts. | Cloud SQL (regional HA) or AlloyDB / Cloud Spanner. |
| "Assigning Basic Owner/Editor IAM Roles" | Violates the foundational security principle of least privilege. | Predefined roles or custom IAM roles granted to Google Groups. |
| "Exposing Backend VMs via Public IP Addresses" | Violates network security isolation; expands the attack perimeter. | Internal Load Balancers, Private Google Access, and IAP TCP Forwarding. |
| "Manual Shell Script Deployment via SSH" | Non-repeatable, prone to drift, lacks auditability and version control. | Declarative Terraform (IaC) with automated Cloud Build CI/CD pipelines. |
Refer to the EHR Healthcare case study. EHR Healthcare must sign new insurance providers quickly, while its legacy file- and API-based insurance integrations remain on-premises for the next several years. New providers will push patient encounter feeds that must be ingested and normalized for analytics. Which architecture best satisfies these requirements?
What is the primary operational advantage of applying the 'Question-First Deconstruction' technique when answering case study questions on the Professional Cloud Architect exam?
Refer to the Cymbal Retail case study. Cymbal wants to automate catalog and content enrichment: generating accurate product attributes from supplier-provided titles, descriptions, and images, and producing image variations such as color changes and background swaps. Associates must approve generated content before it reaches the live catalog. Which combination best fits the stated requirements?
In a case study scenario, the Chief Information Security Officer (CISO) explicitly states: 'Our enterprise must retain sole cryptographic custody over all encryption keys and manage key rotation lifecycles without Google personnel having access to raw key material.' How should the cloud architect translate this executive directive into the technical solution?