Which Attribution Model Should You Actually Be Using?

The attribution model your analytics tool defaults to is quietly shaping every budget decision you make — and most teams never change it. Understanding how each of the six main models works, which suits your product type, and why Matomo and GA4 will report different numbers for the same conversions is one of the most practically useful things a digital analyst or product owner can know.

The attribution model you’re using determines which marketing channels get credit for a conversion — and most teams are using whichever default their analytics tool ships with, without knowing it. That default then quietly shapes every budget decision made from the data.

This happens in every company that runs digital marketing across more than one channel. The marketing director believes certain campaigns are driving revenue. The analyst looks at the same data and sees a different story. Neither is wrong — they’re applying different attribution rules to the same conversion events.

Understanding which model is telling you what — and why it differs from what another tool is reporting — is one of the more practically useful things a digital analyst or product owner can know.

What is digital attribution?

Digital attribution is the process of assigning conversion credit to the marketing touchpoints that preceded it. In practice: when someone books, buys, or fills out a form, which ad, email, organic search, or social post gets the credit — and how much.

It sounds like an analytics problem. It is actually a business decision. The digital attribution model you choose determines where budget goes next quarter. Teams using last-click attribution typically overspend on bottom-of-funnel channels. Teams using data-driven attribution typically see a different split entirely.

There is no objectively correct attribution model. There is only the model that best reflects how your customers make decisions — and being honest about which one you are using.

The six digital attribution models

Last-click

Every conversion is credited to the final touchpoint before it happened. Simple to implement, easy to explain to stakeholders, and deeply misleading if your acquisition funnel has more than one step. This is still the default in many analytics tools. If you are using last-click without knowing you are using it, you are probably undervaluing brand awareness, email, and organic search.

First-click

The opposite problem. The first touchpoint gets all the credit, regardless of what drove the final decision. Useful for understanding where customers first hear about you — useless for measuring what converts them.

Linear

Credit is distributed equally across every touchpoint in the journey. It avoids the extremes of first- and last-click, but it treats a casual scroll past a banner ad as equivalent to someone reading a 1,500-word product comparison page. It rarely matches reality, but it is a defensible neutral position when you don’t have enough data to do better.

Time decay

More credit goes to recent touchpoints, less to earlier ones. The assumption is that what happened closest to conversion was most influential. Reasonable for short sales cycles — less useful for travel, B2B, or any product with a long consideration phase, where an email sent six weeks ago may have been the thing that kept the brand in the running.

For an online travel agency with a 3–6 week booking window, time-decay attribution consistently undervalues the early discovery phase and overstates the role of last-minute retargeting. I’ve seen this cause teams to cut spend on channels that were actually doing the heavy lifting.

Position-based (U-shaped)

40% of credit goes to the first touchpoint, 40% to the last, and the remaining 20% is spread across everything in between. This model acknowledges that both first contact and final conversion matter. It is a reasonable default for businesses where brand discovery and final push are both meaningful — which is most businesses.

Data-driven

The only model that uses your own conversion data to assign credit weights. Google’s version uses machine learning to analyse which combinations of touchpoints actually correlate with conversion. It requires minimum data volume to work — Google requires at least 400 conversions per month — but when you have the data, it’s the most defensible choice for budget decisions.

Matomo’s data-driven attribution is available as a premium plugin and uses a different methodology. For sites with sufficient volume, the outputs rarely match Google’s — which tells you something about how much tool assumptions shape the numbers.

How to choose the right attribution model

The answer depends on three things: your sales cycle, your channel mix, and what decision the attribution data is actually going to drive. Google’s guidance on overhauling attribution is a useful starting point for the strategic framing.

Short sales cycle, simple funnel — last-click is defensible, but test it against linear before trusting the numbers.

Long sales cycle (travel, B2B, high-consideration purchases) — time-decay will mislead you. Start with position-based and move to data-driven once you have the volume.

If the purpose is budget allocation — you need data-driven or at minimum linear. Last-click will consistently over-reward retargeting and paid search.

If the purpose is stakeholder reporting — pick the model that tells the truthful story for your product, and be explicit about which model you are using.

One thing I always do before presenting attribution data: state the model. “According to last-click attribution, paid search drove 40% of conversions” is a very different claim from “paid search drove 40% of conversions.” The first is honest. The second causes arguments.

Attribution in Matomo vs GA4: what actually changes

Both tools offer multiple attribution models, but the defaults differ — and most teams never check.

GA4 defaults to data-driven attribution for conversions where data is sufficient, and last-click otherwise. Matomo defaults to last-click across all reports unless you configure otherwise or purchase the Multi Attribution plugin.

The result: if you are comparing Matomo and GA4 reports side-by-side without accounting for this, you will get different numbers. Not because one is wrong — because they are making different assumptions about credit distribution.

In GDPR-sensitive markets where teams use both tools (Matomo for cookieless tracking, GA4 for benchmarking), this is worth documenting explicitly. The attribution model should appear in any report header. If it doesn’t, the number is incomplete.

The real risk with digital attribution modelling is not choosing the wrong model — it is not knowing which model you are using. Most analytics tools have a default. Most teams never change it. That default then silently shapes every budget decision made from the data.

The real risk with digital attribution is not choosing the wrong model — it is not knowing which model you are using.

Knowing the mechanics of each model does not make attribution easy. But it does make the inevitable “why do your numbers not match mine?” conversation much shorter.

TL;DR

• The attribution model your analytics tool defaults to silently shapes every budget decision — most teams never change it or know which model they’re using
• Last-click over-rewards retargeting and paid search; time-decay misleads for long-consideration products like travel or B2B; data-driven is most accurate but requires volume
• For long sales cycles: start with position-based, move to data-driven when you have the data; for budget allocation, never use last-click alone
• GA4 defaults to data-driven; Matomo defaults to last-click — always state which model you’re using before presenting attribution numbers

---

Delphine Ragazzi is a Product Owner with 20 years of experience across digital analytics, CRO, and product delivery. She writes about product decisions, data, and AI at douli.com.