Building Your Revenue Orchestration Platform
Part 4: Data Architecture
This is Part 4 of a 6-part series. Read Part 3: AI Agents & Automation
The Data Problem
Enterprise revenue teams generate data across a dozen systems:
- CRM contacts and activities
- Email conversations
- Calendar meetings
- LinkedIn connections
- Call transcripts
- Chat messages
- Web engagement
- Marketing automation
Each system has its own data model. Each creates its own version of "the customer."
The result? No single view of the buyer group. No unified understanding of engagement. No way to connect signals across touchpoints.
The Unified Contact Graph
Revenue Orchestration requires a fundamentally different data model: the Unified Contact Graph.
What It Is
A single source of truth for every person your company has ever interacted with:
Person (Ross Sylvester)
├── Identities
│ ├── Email: ross@company.com
│ ├── LinkedIn: /in/rosssylvester
│ └── Phone: +1-xxx-xxx-xxxx
├── Roles (over time)
│ ├── 2024: VP Sales, Company A
│ └── 2026: CRO, Company B
├── Relationships
│ ├── Works at: Company B
│ ├── Reports to: CEO
│ └── Connected to: 47 people in graph
├── Engagement History
│ ├── 23 emails received
│ ├── 5 meetings attended
│ └── 2 calls recorded
└── Signals
├── LinkedIn: Active on AI content
└── News: Mentioned in funding announcement
Why It Matters
With a unified graph:
- Cross-deal intelligence: Know a person's history across all opportunities
- Relationship mapping: See who knows who across accounts
- Signal aggregation: Combine touchpoints for complete picture
- Role tracking: Understand career progression over time
Data Sources
Primary Sources
| Source | Data Type | Sync Frequency |
|---|---|---|
| CRM (Salesforce/HubSpot) | Contacts, accounts, opportunities | Real-time |
| Email (Gmail/Outlook) | Conversations, engagement | Real-time |
| Calendar | Meetings, attendees | Real-time |
| Profiles, connections, activity | Daily | |
| Call recording | Transcripts, sentiment | Post-call |
Enrichment Sources
| Source | Data Type | Purpose |
|---|---|---|
| LinkedIn Sales Nav | Org charts, titles | Role classification |
| ZoomInfo/Apollo | Contact data | Gap filling |
| News/PR | Company events | Signal detection |
| Technographics | Tech stack | ICP matching |
Internal Sources
| Source | Data Type | Purpose |
|---|---|---|
| Marketing automation | Web visits, form fills | Engagement scoring |
| Product analytics | Usage data | Expansion signals |
| Support tickets | Issues, sentiment | Renewal risk |
Identity Resolution
The hardest problem: linking the same person across systems.
The Challenge
- ross@company.com in CRM
- ross.sylvester@personal.com in email
- /in/rosssylvester on LinkedIn
- "Ross" mentioned in a call
Are these the same person?
Resolution Methods
- Email matching - Same email = same person
- LinkedIn URL - Unique identifier when available
- Phone number - Strong match signal
- Name + Company - Probabilistic matching
- Explicit linking - User confirms connection
Confidence Scoring
Every match has a confidence score:
| Score | Meaning | Action |
|---|---|---|
| 95%+ | Certain match | Auto-merge |
| 80-94% | Likely match | Flag for review |
| <80% | Possible match | Keep separate |
Data Quality
Bad data kills revenue orchestration. Quality must be systematic.
Automated Cleanup
- Duplicate detection - Merge or flag duplicates
- Stale data - Alert on contacts without recent activity
- Missing fields - Surface gaps in critical data
- Role changes - Detect job changes via LinkedIn
Human-in-the-Loop
- Merge approval - Confirm uncertain identity matches
- Role correction - Fix misclassified stakeholder roles
- Relationship editing - Adjust reporting lines
CRM Sync
Revenue Orchestration must be bidirectional with CRM:
From CRM
- Contact records
- Account hierarchy
- Opportunity data
- Activity history
To CRM
- Enriched contact data
- Engagement scores
- Buyer group mapping
- Deal risk signals
Sync Principles
- CRM remains source of truth for pipeline and forecasting
- Orchestration enriches but doesn't override
- Selective sync - Only push high-confidence data
- Audit trail - Every change traceable
Implementation Checklist
- Map data sources - Inventory every system with people data
- Define identity keys - Which fields link records across systems?
- Build resolution pipeline - Automated + human-in-loop matching
- Establish quality metrics - Completeness, accuracy, freshness
- Configure CRM sync - Bidirectional with clear ownership
- Design enrichment strategy - Which gaps to fill automatically?
What's Next
Data is the foundation. Workflow turns data into action.
In Part 5, we'll cover Workflow & Execution---from insight to action, how orchestration closes the loop.
Next week: Part 5 - Workflow & Execution