How to fix attribution gaps between marketing qualified leads and closed won deals
Attribution gaps between marketing qualified leads (MQLs) and closed won deals happen when lead source data breaks somewhere in the handoff—usually missing IDs, overwritten fields, or no shared definition of a qualified lead. Fix it by enforcing consistent record linking (lead-to-contact-to-opportunity), locking source fields, and reconciling MQL and closed won counts on a single timeline.
Why attribution breaks between MQLs and closed won
Most teams get this wrong because they treat attribution as a reporting problem when it's really a data plumbing problem. The gap shows up when you can't trace a closed won deal back to the marketing touch that created the MQL. A few root causes account for nearly all of it:
- Lead-to-contact conversion wipes source data. In Salesforce, converting a Lead into a Contact often blanks
LeadSourceon the new Opportunity unless you map it explicitly. - No persistent identifier. Without a stable ID carried from form fill to opportunity, you're matching on email—which changes, gets duplicated, or has typos.
- MQL and closed won live on different timelines. An MQL from Q1 closing in Q3 looks like an unattributed deal if your report windows don't account for sales cycle length.
- Manual SDR edits. Reps overwrite source fields during cleanup, breaking the chain.
If your SDR team is booking meetings that never convert, the same underlying data hygiene issues often inflate MQL counts while real attribution erodes.
Step 1: Define MQL and closed won with one shared standard
Attribution can't reconcile two stages that nobody agrees on. Write down exactly what triggers an MQL—a score threshold, a demo request, a specific form. Tie it to a qualification framework so marketing and sales score leads the same way. If you use the BANT framework to qualify B2B leads, make sure the MQL criteria reference the same budget, authority, need, and timing signals the AE checks later.
Document both definitions in your CRM as field-level criteria, not a wiki page nobody reads.
Step 2: Carry a persistent ID through every stage
The single highest-leverage fix is a stable identifier that survives the lead-to-opportunity conversion.
- Generate or capture a unique ID at first touch (UTM-based campaign ID, form submission ID, or a hidden CRM field).
- Map that field through lead conversion so it lands on the Contact and the Opportunity.
- Validate the mapping in a sandbox before pushing live.
In Salesforce, configure this under Setup > Object Manager > Lead > Fields & Relationships > Map Lead Fields. The official guidance on lead conversion field mapping walks through which standard and custom fields transfer.
Don't rely on email matching alone
Email-based matching fails on free-text entry errors, role changes, and shared aliases like info@. Use the persistent ID as the primary key and email only as a fallback.
Step 3: Pick an attribution model and apply it consistently
There's no perfect model—just pick one and apply it everywhere.
| Model | What it credits | Best for |
|---|---|---|
| First touch | The campaign that created the MQL | Top-of-funnel demand gen |
| Last touch | The final touch before opportunity | Bottom-funnel conversion |
| Multi-touch (linear) | Every touch equally | Long, multi-stakeholder cycles |
| W-shaped | First, MQL, and opportunity creation | Most B2B pipelines |
For account-heavy motions, multi-touch fits better because deals involve several contacts. This matters more if you run account based marketing for B2B sales, where a single closed won deal may trace back to five or six MQLs across one account.
Step 4: Reconcile on a single timeline
MQLs and closed won deals rarely occur in the same reporting period. To avoid phantom gaps:
- Build a cohort report: group deals by the period the MQL was created, not the period it closed.
- Account for your median sales cycle—if it's 90 days, a Q1 MQL closing in Q2 is expected, not lost.
- Track MQL-to-closed-won conversion rate per cohort to spot real leakage versus timing lag.
A cohort view turns "we have 200 unattributed deals" into "these deals came from January MQLs with a 110-day cycle."
Step 5: Lock fields and audit the handoff
Protect the data once it's flowing.
- Make source fields read-only for SDRs and AEs after conversion.
- Set up a weekly validation report flagging Opportunities with blank source or campaign IDs.
- Add a required handoff checklist so reps confirm the MQL record before creating an opportunity.
Clean handoffs depend on clean inputs upstream. If your funnel is full of low-quality leads—say, cold emails landing in spam producing weak engagement—you'll see attribution noise no model can fix.
Common attribution errors and quick fixes
- Blank
LeadSourceon opportunities → add explicit field mapping in lead conversion settings. - Duplicate contacts splitting credit → run dedupe rules keyed on the persistent ID.
- Self-reported source vs. tracked source mismatch → store both; trust tracked for reporting, self-reported for context.
- Offline conversions missing → import event and call data with the original campaign ID attached.
Key takeaways
- Attribution gaps are usually broken data plumbing, not a reporting tool problem.
- Agree on one MQL and one closed won definition before measuring anything.
- Carry a persistent ID from first touch through the opportunity—don't match on email.
- Use cohort reporting to separate real leakage from sales cycle timing.
- Lock source fields and audit handoffs weekly to keep the chain intact.
Get these five fixes in place and your MQL-to-closed-won attribution will reconcile cleanly, giving marketing and sales a shared, defensible view of what actually drives revenue.