First-touch, last-touch and multi-touch attribution each decide in a different way which campaign gets the credit for a lead. The model you use changes which campaigns look good, so choose it on purpose instead of inheriting whatever each ad platform uses by default.
What each attribution model credits
| Model | Who gets the credit | Good at showing |
|---|---|---|
| First touch | The first campaign that brought the person in | Which campaigns find new people |
| Last touch | The last campaign before the lead was sent | Which campaigns close the decision |
| Linear | Every recorded touch, equally | The whole path, at the cost of simplicity |
| Position-based | Most to the first and last touches, with the rest shared | Discovery and decision together |
| Data-driven | Shares estimated from patterns in the platform’s own data | Bidding and budgets inside one platform |
Google retired the first click, linear, time decay and position-based models from Google Ads and Google Analytics 4 in 2023. Data-driven and last click remain, so first-touch reporting today is something you capture yourself rather than a setting you pick.
Why the ad platforms’ totals overlap
Each platform sees only its own touches and credits itself by its own rules. A person who saw a social ad, later searched for the company and then filled in the form can appear as a conversion in both ad accounts. Add the platforms together and the total can exceed the leads the client actually received.
That is fine for steering bids inside each platform. It becomes a problem in a client report, where the totals are supposed to add up.
The case for one recorded source per lead
For many local lead generation clients, the path from ad to enquiry is short, and the useful question is simple: which campaign brought this person in, and did they buy? Recording one original source and campaign on each lead when it arrives answers that in a way the client can check.
- Every lead counts once, so campaign totals add up to the lead total.
- The client can open any campaign and see the actual people behind it.
- The source never changes after the fact, so last month’s report still matches when you open it this month.
Where one source per lead falls short
Single-source reporting undervalues work that supports a decision without starting it, such as retargeting, reviews and searches for the company name. It also strains when journeys run for months; a solar buyer may see several campaigns before asking for a quote. The solar industry page shows how deals that close months later still count toward the month the lead arrived.
If those longer journeys matter to the client, report the original source as the headline and use each platform’s own model as a supporting view. If they are the whole story, a tool that tracks every visit and click fits better, and our lead attribution tools comparison explains when.
How to choose a model for each client
| If the client… | Report |
|---|---|
| Gets most leads from one or two channels and decides quickly | The original source per lead, and nothing more |
| Runs prospecting and retargeting to the same audience | The original source per lead, plus platform views labelled as estimates |
| Has sales cycles measured in months | The original source per lead, credited to the month the lead arrived |
| Needs credit split across many touches | A visit-level tracking tool alongside the CRM |
Keep unknown sources visible
Whatever model you choose, some leads will arrive with no source. Giving them their own row is more honest than guessing, and the size of that row shows how much source capture needs fixing; passing UTM parameters into the CRM closes most of the gap.
Attrival’s client report credits each lead to the first source and campaign captured, and can switch to each lead’s last touch, the visit just before the form. It doesn’t spread credit across touches or fill in a missing source, so an unknown campaign keeps its own row. Campaign attribution shows how that reads in a client report.
