Sign in

Blog · How-to guides

Lead attribution: connect sources to customer outcomes

Lead attribution connects a recorded source to a lead and its later customer outcome. Work through duplicate removal, original-source preservation, and account mapping with a reconciled CRM example.

The short answerLead attribution assigns a recorded source to a lead and connects that lead to later qualification and customer outcomes. Start with stable lead identifiers, preserve the original source, and document how several leads map to one account. Reconcile source totals to unique leads and unique customers separately so duplicate records cannot inflate channel performance.

Lead attribution connects a lead's recorded source to its later qualification and customer outcome. To compare channels fairly, first resolve duplicate lead records, preserve the original source, and decide how multiple people at the same company receive customer credit.

The useful output is a reconciled source table. It should show which leads were accepted, which qualified, which accounts became customers, and what remains unknown.

Choose the unit before assigning credit

A form submission, a lead, and a customer account are different units. Repeated submissions can belong to the same lead. Several leads can belong to one account. A source report needs explicit rules at each level.

For lead acquisition, use the earliest eligible recorded source attached to the canonical lead ID. Keep the latest interaction in a separate field. An original-source correction needs a reason and a retained history; it should not happen simply because someone attended another event.

Scope matters even within analytics exports. Google's channel documentation distinguishes first-user, session, and event scopes. These fields answer different questions, so a session channel should not silently replace the source preserved on a CRM lead.

Customer credit requires another choice: which lead supplies the account's source? The example below credits the earliest accepted lead attached to a new account. This is a reporting convention, chosen before reviewing results. Other conventions can be valid if they are clearly labeled and reconcile.

For allocating an outcome across several recorded interactions, use the separate marketing attribution comparison. Here the task is to establish identity and source ownership before any allocation.

Bring the registers that establish the relationship

Use an intake export with submission ID, canonical lead ID, capture timestamp, original source, campaign, and account ID. Include an acceptance flag and exclusion reason. Keep the raw source text beside a standardized channel so a changed naming rule can be reviewed.

The qualification export needs the canonical lead ID and first qualification date. The customer register needs the account ID and first customer date. Define what makes an account a customer, such as an accepted initial contract, and apply the same definition across channels.

Do not infer that two people belong to the same account because their names look similar. Use the CRM's approved account mapping and keep unresolved relationships visible. Marketing operations owns source capture and duplicate rules; sales operations owns account mapping and qualification history. The team responsible for the customer register confirms customer status.

Work one cohort from intake to customer

Consider fictional US company Cascade Industrial Services. All the leads below arrived on January 5, 2026, in the sequence shown. Qualification and customer outcomes are observed through day 120 after capture. The observation window is complete for every row.

The intake file contains a repeated submission from L01. It records a later webinar interaction, but it does not create a new lead or replace L01's search origin.

Intake sequence Canonical lead Account Recorded source Qualified by day 120 Treatment
1 L01 A01 Search Yes Keep
2 L02 A01 Webinar Yes Keep
3 L03 A02 Partner Yes Keep
4 L04 A03 Search No Keep
5 L05 A04 Webinar Yes Keep
6 L06 A05 Unknown Yes Keep
7 L07 A02 Search No Keep
8 L01 A01 Webinar Yes Exclude repeated lead

The first reconciliation is eight intake rows minus one repeated lead equals seven accepted unique leads. Five of those seven qualified: 5 / 7 = 71.4%, rounded to one decimal place. The repeated row contributes neither an extra lead nor another qualification.

Now apply the account credit rule and join customer outcomes by account ID:

Account Earliest accepted lead Credited source New customer by day 120
A01 L01 Search Yes
A02 L03 Partner Yes
A03 L04 Search No
A04 L05 Webinar No
A05 L06 Unknown Yes

The seven leads map to five accounts, and three became customers. Account conversion is 3 / 5 = 60.0%. It is an account measure; dividing these three customers by seven people would change the unit halfway through the calculation.

Read the source table without double counting

The two registers produce this combined view. Lead columns use each lead's original source. Account columns use the earliest accepted lead's source for that account.

Source Unique leads Qualified leads Credited accounts New customer accounts
Search 3 1 2 1
Webinar 2 2 1 0
Partner 1 1 1 1
Unknown 1 1 1 1
Total 7 5 5 3

Every column ties back to its register. Search qualification is 1 / 3 = 33.3%; webinar qualification is 2 / 2 = 100.0%. Yet webinar receives no new-account credit under the chosen rule. Its L02 belongs to A01, whose earlier L01 gives search the account credit.

That result does not prove the webinar was ineffective. It describes an acquisition-origin convention applied to recorded relationships. Google's attribution overview defines attribution as assigning credit along an observed path; credit alone does not establish what would have happened without an interaction.

Use the result to inspect handoffs, missing source capture, and channel composition. The revenue operations metrics guide covers those handoffs, while stage conversion defines stage-to-stage measurement. Before moving budget, compare customer segments and buying conditions as explained in win rate by segment, source, and deal size.

Where it goes wrong

  • Updating origin with every response. A later campaign overwrites the acquisition history. Retain original source and subsequent interactions separately.
  • Counting every contact as a customer. Buying groups inflate channel totals when the customer register is joined to people without account deduplication.
  • Dropping unknowns. Known-source totals stop reconciling to the business total, and apparent channel shares change without any business change.
  • Mixing observation windows. Older leads have had longer to become customers. Compare complete windows or label cohorts still developing.
  • Treating credit as incremental demand. A source assignment describes the recorded relationship. Additional evidence is needed to claim a causal budget effect.

Covirage is not affiliated with Google.

Reconcile lead attribution on your own file in Covirage

Bring your intake, qualification, and customer exports with the lead-to-account mapping and stated credit rules. Covirage maps the columns, and its tools compute and check the source totals; the external AI model explains the results with every figure traced to its rows. Create a free Covirage account and bring a sample of your own export.

See sales insights for the supporting file-based workflow.

Questions people ask

Is lead attribution the same as revenue attribution?

Lead attribution identifies the source credited with a lead or subsequent customer relationship. Revenue attribution allocates a monetary outcome across eligible interactions. A reliable lead-to-account mapping can support either analysis, but it does not determine how revenue should be divided.

How should unknown lead sources be handled?

Keep unknown sources as an explicit category and retain the underlying records for review. Do not distribute them across known channels just to make a report look complete. Missing source information is a coverage issue that the marketing operations owner should investigate.

Can several leads receive credit for the same customer?

Several leads can belong to a customer's buying group, but a unique-customer report must count that customer once. Set a documented account credit rule or show shared credit separately. Keep person-level qualification counts separate from the account-level customer total.

Should a later campaign replace the original lead source?

Preserve the original recorded source when measuring acquisition origin. Store later campaign interactions separately so they can answer engagement questions. If a correction is necessary, retain the prior value and reason so historical channel results remain explainable.