Prompt
I'm building a B2B attribution dashboard that connects pipeline and revenue—what's the best approach?
Latest observation
The best approach is to build a hybrid attribution system that combines:
- Deterministic identity resolution
- A clear event model for all touchpoints
- A rules-based attribution layer for operational reporting
- A revenue/pipeline stitching layer that ties marketing activity to CRM objects
- A pipeline quality framework, not just source tracking
That gives you something that’s both accurate enough for sales/marketing decisions and flexible enough to avoid the common “last-touch only” trap.
Recommended architecture
1) Capture the right data
Track all meaningful buyer interactions:
- Web visits
- Form fills
- Email clicks
- Ad clicks/impressions
- Webinar attendance
- Content downloads
- Demo requests
- Sales calls / meetings
- CRM stage changes
- Opportunity creation, close won, expansion
Each event should have:
anonymous_idknown_user_id/lead_idaccount_id- timestamp
- channel/source/campaign
- UTM params
- landing page
- referrer
- event type
- revenue object linkage if known
2) Resolve identity early
You need a way to connect anonymous browsing to known contacts/accounts:
- First-party cookie / device ID
- Email capture on forms
- CRM contact matching
- Reverse IP / account enrichment for ABM use cases
This is critical because attribution breaks if pre-conversion activity remains anonymous.
3) Build a canonical event table
Create a single normalized table for all touchpoints.
This becomes the source of truth for reporting.
Example dimensions:
- person
- account
- opportunity
- campaign
- channel
- timestamp
- touchpoint type
- role in buying committee
- lifecycle stage at time of touch
4) Use multiple attribution models
Don’t rely on one model.
At minimum support:
- First touch – acquisition efficiency
- Last touch – conversion efficiency
- Multi-touch – influence across journey
- Time-decay – more weight to recent touches
- U-shaped / W-shaped – important for B2B with long cycles
- Custom weighted – tailored to your sales motion
For B2B, the most useful operational setup is usually:
- Pipeline creation attribution
- Pipeline influence attribution
- Revenue attribution
- Sourced vs influenced split
5) Tie touches to CRM lifecycle milestones
You should attribute separately to:
- Lead created
- MQL
- SQL
- Opportunity created
- Stage progression
- Closed won
- Expansion / renewal
This matters because “what created pipeline” is different from “what closed revenue.”
Best-practice attribution logic
A. Source of truth for pipeline creation
Assign the opportunity’s “source” based on:
- first known meaningful touch before opp creation
- or the marketing campaign that generated the lead/contact
- or the first acquisition source at account level
Use this to answer:
- Which channels create pipeline?
- Which campaigns generate qualified opportunities?
B. Influence model for all touches before close
Allocate credit to all relevant touches inside the lookback window.
Rules to define:
- Lookback window: e.g. 90/180/365 days
- Include only pre-opportunity touches, or all pre-close touches
- Exclude internal and duplicate touches
- Weight by stage proximity and role
Use this to answer:
- Which programs influence deals?
- What content/campaigns assist revenue?
C. Revenue attribution at contact/account level
For B2B, a single deal often has multiple contacts.
So you need to attribute at both levels:
- Contact-level for person journey
- Account-level for buying committee and enterprise deals
Then roll up to opportunity.
Dashboard metrics that matter
Pipeline metrics
- Marketing-sourced pipeline
- Marketing-influenced pipeline
- Pipeline creation rate by channel
- Opportunity conversion rate by source
- Average deal size by source
- Stage progression velocity
Revenue metrics
- Marketing-sourced revenue
- Marketing-influenced revenue
- CAC by channel
- ROI by campaign
- Payback period
- Win rate by source
- Expansion influenced by marketing
Quality metrics
- Attribution coverage %
- Match rate for anonymous → known
- % opportunities with at least one marketing touch
- Duplicate event rate
- CRM sync lag
- % revenue unattributed
Important design decisions
1) Define “sourced” vs “influenced”
These are not the same thing.
- Sourced: the interaction that initiated the opportunity or lead
- Influenced: any marketing touch that participated in the journey
If you don’t separate them, the dashboard becomes misleading.
2) Choose the right level of granularity
You can report at:
- campaign
- channel
- source / medium
- asset
- ad set
- keyword
- account segment
Start with channel and campaign first.
Too much granularity too early creates noisy reporting.
3) Be explicit about windowing
Your attribution results depend heavily on:
- 30 vs 90 vs 180-day windows
- whether you include post-opportunity touches
- whether you include closed-lost deals
Document this in the product so users trust the numbers.
4) Handle long sales cycles
B2B deals often span multiple months and multiple stakeholders, so:
- use account-level attribution
- capture all members of the buying committee
- support multiple touches per deal
- don’t force a single “winner” channel
Suggested implementation sequence
Phase 1: Foundation
- Unify event collection
- Build identity resolution
- Normalize CRM and marketing data
- Create canonical campaign and touch tables
Phase 2: Basic reporting
- First-touch and last-touch
- Sourced pipeline and revenue
- Channel performance
- Campaign performance
Phase 3: Multi-touch attribution
- Time-decay / U-shaped / W-shaped
- Opportunity-level touch weighting
- Account-level rollups
Phase 4: Advanced analytics
- Cohort analysis
- Channel efficiency by stage
- Predictive attribution
- Incrementality testing
- Contribution to expansion/retention
Common pitfalls to avoid
- Over-trusting last-touch attribution
- Mixing lead, contact, and account definitions
- Counting internal CRM updates as marketing touches
- Ignoring anonymous pre-conversion behavior
- Not deduping campaigns and contacts
- Treating all touches equally
- Failing to document attribution rules
If you want the “best” practical setup
For most B2B teams, I’d recommend:
- Primary dashboard: sourced pipeline + sourced revenue
- Secondary dashboard: influenced pipeline + influenced revenue
- Attribution model: W-shaped or time-decay
- Unit of analysis: opportunity + account
- Identity resolution: contact + account
- Lookback window: 180 days as a starting point
- Source breakdown: channel, campaign, and lifecycle stage
If you want, I can also help you design:
- a data model/schema for this dashboard,
- a SQL/dbt implementation, or
- the exact attribution rules for B2B SaaS.