Marketing Attribution Models Compared: Which One Fits Your Stage?
Marketing says the campaign worked, sales says they closed it themselves, and finance wants one number by Friday.
Everyone is looking at different data and calling it the truth. Picking an attribution model won't end that argument. It decides whose version of the story becomes the official one.
What an attribution model does
An attribution model is a rule for splitting credit across the touches that led to a deal. That's all it is.
No model captures how a B2B deal closes. What it gives you is one agreed way to judge what's working. Without it, every performance conversation is a negotiation.
Agree on the model before the report goes to the board.
The models side by side
| Model | How credit is split | Good at | Blind to |
|---|---|---|---|
| First touch | 100% to the first touch | What brings new people in | Everything that happened after |
| Last click | 100% to the final touch before conversion | What closes the gap at the end | Everything that created the demand |
| Linear | Equal shares to every touch | Long cycles with lots of nurture | Which touches mattered more |
| W-shaped | 30% each to first touch, lead creation and opportunity creation; 10% across the rest | B2B with a defined sales process | Messy data. It needs clean fields to work. |
| Data-driven (GA4) | Estimated from your conversion data | Checking paid channels | Anything that never reaches GA4 |
Linear: underrated for teams just starting out
Say a contact had eight touches before becoming an opportunity. A LinkedIn ad, two emails, a webinar, a direct visit, a demo request, a follow-up email and a sales call. Linear gives each one 12.5%.
For a six-month cycle, that's useful. Every nurture email and retargeting ad gets some credit, so nobody kills the six things that built the relationship just because a form fill came last.
It's also the easiest to set up and explain. If you're standardizing UTMs and setting up campaign attribution for the first time, start here and add complexity later.
The catch: when everything gets equal credit, it's hard to tell which programs are doing the work and which are along for the ride.
W-shaped: what I recommend most often
For B2B teams with a defined sales process and enough volume to see patterns, it's what I pick most of the time.
It answers the three questions leadership asks. How did they find us? What made them raise a hand? What got them talking to sales? Those moments matter more than the rest, and W-shaped weights them that way.
It also shows which channels bring people in, which ones nurture, and which ones tip deals into sales. A webinar that keeps showing up at opportunity creation is a different asset than one that shows up mid-funnel.
The downside is data. It needs clean fields. If lead source is full of "other" and campaign membership has holes, W-shaped gives you neat percentages built on sand.
Last click: still in a lot of CRM reports
Google Ads has moved most conversion actions to data-driven attribution, but last click is still what plenty of CRM reports and older ad accounts use, and it's why paid search can look like your best channel in those reports. Last click is decent at one question (what closes the gap at the end) and a poor guide to budget, because it ignores everything that created the demand.
Data-driven attribution in GA4
The default in GA4 and in most Google Ads accounts. Instead of a fixed rule, it estimates credit from your conversion data. For B2B it has two limits:
- It only sees what reaches GA4. Sales calls, events and CRM stages are missing unless you send them in.
- Out of the box it scores website conversions, not pipeline or revenue.
Use it to check paid channels. Keep a CRM model for pipeline.
Which attribution model fits your stage
Two things decide it: how clean your data is, and how long your deals take.
- Deals close in under 30 days with a handful of touches? W-shaped is overkill. First touch and last touch side by side, with a linear view for context, covers it.
- Deals take 60+ days and involve several people? W-shaped gives you a story about marketing's contribution you can defend in a pipeline review.
- Data is a mess? Fix that before you pick anything. A fancy model on bad data gives you a precise wrong answer.
Still not sure? The 2-minute quiz picks one from your answers.
Before you pick
- Can you trace your last five closed deals back to a first touch?
- Do your forms capture UTMs into hidden fields?
- Does campaign membership update at every stage, or only when the lead is created?
- What share of pipeline has a lead source of "other" or blank?
- Who will defend the model when someone in finance asks?
The part people skip
Picking the model is the small part. Most of the work is making sure the data going into it is right.
That means UTM naming your whole team follows, hidden fields on every form, campaign membership updated at every stage, and a CRM and marketing automation tool that sync lead source, lifecycle stage and campaign membership.
Write down why you chose the model and what it can't see. Future you will need that paragraph.
Know what it can't see
Every model has blind spots. W-shaped doesn't see the Slack thread, the podcast your buyer heard three times, or the peer who said "just talk to them."
That doesn't make the model wrong. It means the numbers come with a footnote, and somebody has to say the footnote out loud in the pipeline review.
Skip the build
The B2B Attribution System
The locked UTM system, GA4 and CRM setup, the model decision guide and the dashboard. $399.
See the Attribution SystemNot sure your data is ready? Run the free audit first.
Notes from the Field.
Every other week, one broken thing and how I’d fix it.
One email every other week · unsubscribe anytime · Privacy policy