Every SFMC engagement that blows up in week four could have been saved in week one. The story is almost always the same: client says "we want to send marketing email," team agrees, team builds Data Extensions and templates and automations, and two weeks later the client mentions three things that change the architecture entirely.

• "Oh, I thought emails would send from each sales rep's address?"
• "My data is in Salesforce CRM, not a CSV file."
• "I also need to segment by purchase behavior from Shopify."
• "Compliance requires GDPR consent tracking."

Each of those is a rebuild, and they all happen in the same conversation because nobody asked upfront. Rework is 3x the cost of doing it right the first time.

Here's the five-group discovery checklist we run before opening Content Builder.

Group 1: Where does the client's data actually live?

• Subscriber data is stored where today - CSV, Salesforce CRM (Sales/Service Cloud), own database (MySQL, PostgreSQL), e-commerce platform (Shopify, Magento), or multiple sources?
• What's the unique identifier per subscriber (CustomerID, MemberID, email)? Does it match across systems?
• Does the data contain PII (name, email, address, phone)? Who has access today?
• Does GDPR or PDPA apply? EU users mean GDPR is mandatory; VN/TH users mean PDPA. Do you need consent tracking?

Group 2: What email types and use cases need to ship?

• Marketing, transactional, or both?
• Monthly volume? Under 250k can live on shared IP; above that needs dedicated IP + warming plan.
• Frequency - daily, weekly, monthly, trigger-based?
• Personalization depth - first-name-only, dynamic content by segment, or full AMPscript logic?
• Multi-language? How many languages, who provides translations?

Group 3: Tech stack and integrations

• Salesforce CRM yes/no? If yes, Marketing Cloud Connect required; which objects need to sync (Contact, Lead, Opportunity, custom)?
• Other systems to integrate - e-commerce (abandoned cart, order confirmation), loyalty system, CDP/DMP, BI tool?
• Data ingestion path - SFTP drop, real-time API, Salesforce Sync, manual import?
• Data export needs - tracking data to warehouse, subscriber list back to CRM, automated reports?

Group 4: Team and operations

• Who operates SFMC after handoff - technical team or marketing team? HTML/CSS literate? SQL literate? This decides template strategy.
• Email approval workflow - Content Builder Approval Workflow needed? How many approvers (Legal, Compliance, Brand)?
• Multiple brands or Business Units? Enterprise account with multiple BUs, data sharing between BUs?
• Reporting audience - marketing team or C-level? Dashboard or exported files? BI tool integration (Tableau, Power BI)?

Group 5: Compliance and legal

• How is the subscriber list collected - explicit opt-in or purchased? Bought lists are a red flag; escalate immediately.
• Unsubscribe mechanism - Preference Center exists? Need a custom one? Does unsubscribe need to sync back to CRM?
• Physical sender address - CAN-SPAM requires it in every commercial email.
• Do any emails need to reach unsubscribed contacts (recalls, safety notices, account security)? Those are Send Classification = Transactional.

The timeline that falls out

Once discovery is clean, the first month looks like:

Week 1: Discovery
  - Run the five-group checklist with the client
  - Map data sources and unique identifiers
  - Identify compliance requirements (GDPR, PDPA, CAN-SPAM)
  - Prioritize use cases
  - Define success metrics

Week 2: Foundation setup
  - Design the data model before creating the first Data Extension
  - Sender Profile, Delivery Profile, Send Classification
  - Verify From Email Address
  - Enhanced SFTP if file import is needed
  - Install MC Connect if needed and test the sync
  - IP warming plan if dedicated IP is in scope

Week 3-4: Build MVP
  - Create DEs per the data model
  - Do the first data import and verify row counts
  - Build the first template
  - Ship use case #1 (usually welcome email)
  - Test thoroughly before go-live

Go-live -> Monitor -> Optimize
  - Watch metrics daily for the first two weeks
  - Watch bounce rate and complaint rate like a hawk
  - Adjust based on real data

Takeaway

Discovery feels unproductive because you're not building anything. It's the most leveraged time on the project. Every hour there saves three hours of rework. When a new SFMC engagement lands on our plate, nothing gets dragged onto a canvas until the five groups are answered, written down, and agreed with the client.

Starting an SFMC engagement? Our Salesforce team runs discovery, data model design, and Marketing Cloud builds end to end. Get in touch ->

See our full platform services for the stack we cover.