Lead attribution assigns credit for an enquiry to the recorded marketing interactions that came before it. A visitor might find you through a Google ad, return through search, and submit a form after clicking an email. The attribution model determines which of those interactions gets credit for the lead.
TL;DR
The same lead can have different attributed sources. First-touch, last-touch, and linear models apply different rules to the same recorded journey.
Choose the question before the model. Discovery, the visit before an enquiry, and shared credit are different things to measure.
Keep the lead connected to its sales outcome. Source data helps you compare channels once you know which enquiries became customers.
What is lead attribution?
A source record tells you where a recorded visit came from. Attribution adds a rule for deciding which marketing touchpoints receive credit for a conversion. Here, the conversion is a website enquiry: someone submits your contact or quote form.
A touchpoint is an interaction included in the analysis. To keep this guide simple, each touchpoint in our example is a website visit from an identifiable marketing source.
That distinction matters when a person visits more than once. The source of their final visit may be different from the source that first brought them to your site. Both can be useful, but they answer different questions.
Source tracking captures the evidence. An attribution model decides how to distribute credit across it. Keeping a journey is useful even before you build a report that assigns percentages to each touch.
One lead, three different attribution results
Imagine someone is looking for a business to redesign their website. In this hypothetical example, all four visits are recorded, belong to the same person, and fall within the period included in the report.
| Visit | Recorded channel | What happens |
|---|---|---|
| 1 | Google Ads | They click an ad and read your services page. |
| 2 | Organic search | They return through an unpaid search result and read a case study. |
| 3 | Referral | They follow a link from a partner's website to your pricing page. |
| 4 | They click your newsletter, return to the site, and request a quote. |
Google Ads → Organic search → Referral → Email → Form submitted
The form submission creates one lead. It marks the end of the journey we're analysing here; it isn't a fifth marketing source.
Now apply three different models to those same four visits:
| Model | Google Ads | Organic search | Referral | |
|---|---|---|---|---|
| First-touch | 100% | 0% | 0% | 0% |
| Last-touch | 0% | 0% | 0% | 100% |
| Linear | 25% | 25% | 25% | 25% |
The journey hasn't changed. The rule for assigning credit has.
First-touch: credit the first recorded interaction
First-touch gives all the credit to the earliest eligible touch. In our example, Google Ads gets 100%. It answers a discovery question: which recorded channel introduced this lead?
It leaves out the later visits, so it doesn't explain what happened immediately before the enquiry.
Last-touch: credit the interaction before the enquiry
Last-touch gives all the credit to the final eligible touch before conversion. Email gets 100% in our example. It tells you which recorded channel preceded the quote request.
It leaves out how the visitor first found you and the visits in between.
Linear: share credit equally
Linear divides credit equally across the eligible touches. Each of our four visits receives 25%. This is a multi-touch approach: more than one interaction gets a share.
Equal credit is a rule, not evidence that each visit had equal influence. These three model definitions follow the principles described in Adobe's attribution documentation.
You still have one lead. In the linear report, its credit is split into four shares; you haven't generated four enquiries.
Which attribution model should you use?
Start with the decision you're trying to make. A model that helps explain discovery may be less useful for understanding the visit that led directly to an enquiry.
| Your question | Useful starting view | What to remember |
|---|---|---|
| Which recorded channel introduced this lead? | First-touch | Later interactions receive no credit. |
| Which recorded channel came just before the enquiry? | Last-touch | Earlier discovery receives no credit. |
| How would the result look with equal credit across the recorded touches? | Linear | Equal shares don't prove equal influence. |
For a small team, comparing first-touch and last-touch can reveal something useful without adding a complicated weighting system. In our example, Google Ads introduces the visitor while email brings the visit that ends in a quote request. Looking at only one of those views hides the other role.
Other approaches use different weights. Time-decay gives more credit to recent interactions, while position-based models favour selected points in the journey. More detailed rules also mean more assumptions to explain.
Before choosing one, check whether you have the underlying journey data. A weighting formula can't recover a visit that was never recorded. Also check which models your reporting tool supports; this guide's examples aren't settings available in every analytics product.
TrackLeadSource
See the story behind each lead.
Keep source, campaign, and journey context together when a visitor becomes a lead.
7-day free trial. No card required
Why attribution reports disagree
Two reports can describe the same enquiry differently because they use different rules or different evidence. Check the details before assuming one is broken.
They include different periods or interactions
An attribution window defines how far back the report looks for eligible touches. A visit outside that window won't receive credit. Scope matters too: a report covering one visit can differ from a report covering several visits by the same person. Adobe explains both scope and lookback windows.
For our example, excluding the first Google Ads visit would change the first-touch result. Organic search would become the earliest remaining touch.
They count different conversions
A form enquiry, a qualified opportunity, and a completed sale happen at different points. A report ending at the form should not be compared directly with one ending at the sale.
Write down what the report counts. “Email received credit for the enquiry” is more precise than “Email generated the customer” when the customer hasn't bought anything yet.
They handle direct visits differently
Some reporting rules exclude direct visits when another eligible source is available. A report that includes the final direct visit can therefore disagree with one that credits the previous identifiable channel. Google's attribution guide explains how its reporting models handle direct traffic.
Our example has no direct return visit, so this rule doesn't affect its table. If you add one, check the rule before interpreting the result.
They haven't captured the same journey
Someone may use another device, clear browser storage, or hear about you through a private recommendation. Your records may only show part of that path.
A “How did you hear about us?” answer can add context. Keep it alongside the tracked information rather than replacing it. A person can first hear about you from a friend and later arrive through a Google ad; those answers can both be useful.
A lead isn't a customer yet
Return to the website redesign enquiry. Email received the last-touch credit, but the quote is still open. You don't yet know whether this enquiry will become a customer.
Later, sales marks the deal as won in the CRM. You can now connect that outcome to the recorded journey. The sale is an additional fact; it doesn't change which visits came before the original form submission.
Keep the distinction clear in your reporting:
- Lead attribution: how credit is assigned for the enquiry.
- Sales outcome: whether that enquiry became a customer and what the sale was worth.
- Revenue attribution: how credit is assigned for a revenue conversion, with its own defined rules.
As more leads move through sales, compare qualified enquiries and customers by channel using a consistent definition. Give leads from the same period enough time to complete your sales process. A channel with fewer enquiries may still bring more suitable projects.
Avoid cutting a channel just because it receives little last-touch credit. Check whether it appears earlier in the journeys of customers who buy. That is a reason to investigate its role, not proof that it caused those sales.
Start with the source data you can capture
Useful attribution starts with keeping the recorded source and journey attached to the enquiry. Once that information is saved, you can interpret it with a clear model and follow what happens to the lead.
TrackLeadSource captures source, medium, campaign, source journey, and visited-page information to pass with website form submissions. You keep the sales outcome in your CRM or spreadsheet. Recording that context and assigning model-based credit are separate parts of the work.
If your forms currently save only a name, email, and message, start with our guide to tracking lead source in website forms. It covers how to keep the source with the submitted enquiry.
Common questions
Sometimes. A referring website or other captured source information may provide context for an untagged visit. Missing UTMs don't automatically mean a visit was direct. Keep unknown origins explicit when the evidence isn't available.
For the three models in our example, that one eligible touch receives all the credit. The result doesn't tell you whether the visitor had other interactions that weren't recorded.
You need reliable source records and a clear reporting question first. A simple first-touch or last-touch view can be useful. Add more detailed modelling when you have the journey data and a decision that the extra detail will help you make.