Attribution & Conversion Counting

Data-Driven Attribution vs Position-Based Attribution

In short: Both are multi-touch models that split credit across a conversion path instead of crowning one touch the winner, but they get the split from very different places. Data-driven attribution (DDA) lets Google's model learn the weighting from your account's converting and non-converting paths; position-based fixes 40 percent on the first touch, 40 percent on the last, and 20 percent on everything between, no matter what your data shows. DDA is the current default in Google Ads and GA4 and needs enough conversion volume to model reliably; position-based was retired from both platforms in 2023 and now lives only in third-party tools. Choose DDA when you trust the platform's evidence and have the volume to support it; choose position-based when you want a transparent, fixed rule that specifically rewards discovery and close.

By the AdFlint research team · Fact-checked against current Google and Meta platform behavior · Last reviewed July 2026

Data-Driven Attribution

Uses a platform's own modeling of converting and non-converting paths to distribute fractional credit based on observed contribution rather than fixed rules.

The platform compares paths that converted with paths that did not, then assigns each touchpoint a fractional share reflecting its measured contribution. It is the default in Google Ads and GA4 and needs enough conversion volume before the model has data to learn from. Two things get misread: the credit is modeled rather than observed, and it remains in-platform, so it cannot see channels outside that platform's data.

Full definition

Position-Based Attribution

Assigns forty percent each to the first and last touchpoints and splits the remaining twenty percent across everything in between.

Also called U-shaped, this model rewards the two moments most advertisers care about, discovery and close, while acknowledging the middle without letting it dominate. It fits multi-step B2B paths where an early content touch and a final branded search both clearly matter. Google Ads and GA4 retired it in 2023. The weakness is that the weights are arbitrary conventions, not estimates derived from how your customers actually behave.

Full definition

Side by side.

The differences that actually change what happens in your account.

	Data-Driven Attribution	Position-Based Attribution
Source of the weighting	Learned from comparing converting and non-converting paths in your account.	Fixed convention, 40 percent first touch, 40 percent last touch, 20 percent split among the rest.
Does the split change per account	Yes, it is trained on and specific to your own data.	No, the 40/40/20 formula is identical for every account that uses it.
Minimum conversion volume needed	Needs sufficient volume for the model to find reliable patterns.	None, the formula runs the same on any path length or volume.
Where it's active today	Default in Google Ads and GA4.	Removed from Google Ads and GA4 in 2023; available in third-party or custom-built models.
How a middle-path touch fares	Credited based on whether it actually correlates with conversion, which could be high or low.	Always capped inside the shared 20 percent bucket, regardless of how important it actually was.
Explainability	Low, you see the resulting split, not the model's reasoning.	High, anyone can recompute the split from path position alone.
What it costs you when wrong	A low-volume account gets a model trained on too little data, which can look confident but rest on thin evidence.	A channel that only ever appears mid-path gets structurally capped at a small share even if it is genuinely influential.

What actually separates them.

DDA's per-touch weight is estimated from data and can vary touch to touch and account to account; position-based's weights are fixed at 40/40/20 regardless of what the data shows.

Position-based always reserves 80 percent of total credit for exactly two touches, first and last; DDA has no such structural guarantee and can spread credit however the data supports, including concentrating it in the middle.

DDA requires enough conversion volume to model reliably; position-based runs identically regardless of volume, even on a single conversion.

A middle touchpoint can receive more credit than an anchor touchpoint under DDA if the data supports it; under position-based a middle touchpoint can never out-earn either anchor, the math forbids it.

Only DDA remains selectable inside Google Ads and GA4 as of 2026; position-based must be recreated in a third-party attribution tool or custom model.

Which one should you use?

Use Data-Driven Attribution when

You are inside Google Ads or GA4 with enough conversion volume for the model to have something to learn from.
You do not want to assume in advance that first and last touch matter most, you want the data to decide.
You are setting Smart Bidding targets and want the underlying conversion credit to reflect measured contribution.
Your funnel does not have a clean, describable discovery-then-close shape that would justify a fixed positional rule.

Use Position-Based Attribution when

You can clearly name the first-touch and last-touch channels in your funnel and believe they deserve the lion's share of credit.
You want a transparent rule you can explain in one sentence, without invoking a model's internal reasoning.
Your conversion volume is too low to trust a modeled approach.
You are auditing a B2B pipeline where an early content touch and a final branded search or demo request are the two moments that matter to the business.

Common questions.

Can I approximate position-based attribution using data-driven attribution's results?

Not directly. DDA's weighting is specific to your data and will not reliably mimic a fixed 40/40/20 split unless your actual conversion patterns happen to concentrate credit that way. If you specifically want the positional logic, you need a tool that still offers position-based as a distinct preset.

Why would data-driven attribution give my first touch more than 40 percent?

Because DDA has no ceiling tied to position. If the model finds that a specific first-touch channel strongly correlates with conversion outcomes across your account, it can assign that channel well above what a fixed positional rule would allow. That is the core tradeoff: DDA can go anywhere the data leads, while position-based never can.

Is position-based attribution still worth understanding if I can't select it in Google's tools?

Yes, mainly because many third-party attribution and marketing analytics tools still offer it, and the underlying logic of rewarding discovery and close specifically is a useful mental model even when you are reading DDA's output. It helps you notice when DDA's modeled result is quietly behaving like a positional split versus something more evenly or unusually distributed.

Or stop choosing between them.

AdFlint picks the setting, writes the ads, and keeps optimizing inside the Google and Meta accounts you already own.

Try AdFlint free See how it works

Related comparisons

All comparisons