Google Ads Attribution Models Explained: Which One Should You Use?

Google Ads attribution models decide which touchpoint gets credit when a user clicks multiple ads before converting. The model you choose shapes which campaigns look successful, which look like they are wasting money, and where your Smart Bidding algorithm steers the budget.

Most accounts we audit use the wrong attribution model — almost always last-click — and end up underfunding upper-funnel campaigns that are actually doing the heavy lifting. This guide walks through every Google Ads attribution model, when each one fits, and how to switch without disrupting performance.

The framework is simple. The decisions are not. Done right, switching attribution can shift 10 to 30 percent of your conversion credit between campaigns and unlock budgets that were stuck.

What attribution actually does

Attribution decides how to split conversion credit across multiple ad clicks. If a user clicks your Display ad on Monday, your YouTube ad on Wednesday, and your Search ad on Friday before converting, attribution decides which click gets credit.

Last-click gives 100 percent of the credit to the Friday Search click. First-click gives 100 percent to the Monday Display click. Linear gives equal credit to all three. Data-driven uses machine learning to distribute credit based on what actually moved the needle.

The model you choose tells Google's algorithm where to spend. If last-click hides the value of upper-funnel touches, the algorithm starves those campaigns of budget — and your pipeline shrinks even though everything looks fine in last-click reporting.

The six Google Ads attribution models

There are six attribution models available in Google Ads. Two are now deprecated for new accounts but still appear in older setups.

Last-click attribution

Gives 100 percent of the conversion credit to the final ad click before conversion.

Best for: very short funnels (under 1 day), strong commercial-intent campaigns, brand defence, or accounts where Smart Bidding has not had time to work with anything more sophisticated.

The bias: undervalues every touch except the last. Upper-funnel campaigns look unprofitable. Long-cycle B2B underweights awareness work.

Google's stance: last-click is being phased out as the default. Data-driven is now the recommendation for most accounts.

First-click attribution (deprecated)

Gives 100 percent of credit to the first ad click in the conversion path.

Best for: rare cases where you want to optimise hard for awareness and demand creation.

The bias: undervalues closing touches. Last-click campaigns look unprofitable. Branded search gets no credit even though it closed the deal.

Google removed this model for new accounts in 2023. If your account still uses it, switch.

Linear attribution (deprecated)

Splits credit equally across every ad click in the path.

Best for: very long, even-touch funnels where every interaction matters equally.

The bias: treats a random Display impression as equal to a high-intent Search click. Almost never reflects reality.

Like first-click, linear is no longer available for new accounts. Older accounts running it should switch.

Time-decay attribution (deprecated)

Gives more credit to clicks closer to the conversion event. A click an hour before conversion gets more than one a week before.

Best for: funnels where recency is the dominant factor — short commercial-intent journeys with multiple touches.

The bias: still mostly favours last-touch, just with slightly more credit to earlier touches than pure last-click.

Now deprecated for new accounts.

Position-based attribution (deprecated)

Gives 40 percent of credit to first click, 40 percent to last click, and 20 percent split among middle clicks.

Best for: accounts where awareness and closing are equally important, with middle touches as nurture.

The bias: arbitrary 40/40/20 split rarely matches real influence. Loved by some marketers because it tells a balanced story, but it is not data-driven.

Also deprecated for new accounts.

Data-driven attribution

Uses Google's machine learning to distribute credit based on the actual contribution of each click. Different models trained per account.

Best for: most accounts with 300+ conversions per month. This is the default for new accounts and Google's recommended model.

The strength: it reflects what actually drives conversions in your specific account, not a rule-of-thumb assumption.

The limitation: requires conversion volume. Below 300 monthly conversions, Google falls back to last-click anyway.

Why last-click is wrong for most B2B and considered-purchase accounts

Last-click undervalues every touch that is not the closer. For most B2B and considered-purchase accounts, this creates a systematic bias.

A typical B2B buyer journey looks like this — search "what is X", read a blog post, get retargeted, search again "X vs Y", read a comparison, get retargeted, finally search the brand name and convert.

Under last-click, the branded search gets all the credit. The earlier non-brand campaigns that built awareness look like they did nothing. So the algorithm starves them, awareness drops, branded search volume falls, and the whole pipeline shrinks.

We see this pattern on most B2B accounts. The fix is switching to data-driven attribution, which spreads credit across the full journey.

When data-driven is the right call

Data-driven attribution is the right model for almost every account that meets two conditions.

• At least 300 conversions per month across the relevant time window
• A funnel with multiple ad touches before conversion

That covers most e-commerce above €5K monthly ad spend and most B2B above €10K monthly ad spend.

If your account does not meet these thresholds, data-driven will silently fall back to last-click — Google needs the volume to train the model. In that case, focus on building conversion volume first.

When last-click is still acceptable

Last-click works in specific cases.

• Pure brand defence campaigns where the click is the entire journey
• Very short funnels (under 1 day from first click to conversion)
• Accounts with under 100 conversions per month where data-driven cannot train
• Single-campaign accounts where attribution mostly does not matter

In all other cases, data-driven is the upgrade.

How attribution interacts with Smart Bidding

Smart Bidding uses your attribution model to decide which clicks to value most. This is where attribution choice has its biggest budget impact.

Under last-click Smart Bidding, the algorithm bids hardest on the closing clicks — branded search, high-intent commercial queries. Upper-funnel campaigns get less budget.

Under data-driven Smart Bidding, the algorithm bids based on each click's actual contribution. Upper-funnel campaigns get more budget if they really do contribute. Branded search gets less if it would have converted anyway.

We covered the full Smart Bidding decision tree in our Smart Bidding strategies guide. Attribution is the layer underneath that decides what those bidding strategies optimise toward.

How to switch attribution models without crashing performance

Switching attribution disrupts Smart Bidding learning. The algorithm has to re-learn which clicks matter. Expect a 2 to 4 week wobble.

Step 1 — Document the baseline

Capture the past 30 days under your current attribution model. Save: conversions, CPA, ROAS, campaign-level conversion split, top converting campaigns.

Step 2 — Switch during a stable window

Avoid switching during launches, sales periods, or budget changes. Pick a normal two-week window.

Step 3 — Expect campaign-level conversion shifts

Under data-driven, branded search conversions typically drop 20 to 40 percent (because the credit is redistributed) and non-brand commercial campaigns rise 10 to 30 percent.

This is not lost conversions — it is redistributed credit. Total conversions hold or improve.

Step 4 — Hold all targets steady for 2 weeks

Do not tighten Target CPA or Target ROAS during the transition. The algorithm needs time to re-stabilise.

Step 5 — Adjust budgets after the dust settles

After 4 weeks under the new attribution, you will see which campaigns deserved more budget all along. Reallocate based on the new data.

We documented the full conversion tracking foundation in our Google Ads conversion tracking setup guide. Attribution sits on top of tracking — both need to be right.

How to verify your account is on the right model

Open Google Ads. Tools → Measurement → Conversions. Click on each primary conversion action. Look at the "Attribution model" field.

If it says last-click, evaluate whether you meet the 300-conversion threshold for data-driven. If yes, switch. If no, focus on volume first.

If it says first-click, linear, time-decay, or position-based — those are deprecated. Switch to data-driven (or last-click if volume is too low).

A real example — Marseille e-commerce account

We took over a Marseille cosmetics e-commerce account using last-click attribution. The account was scaling profitably but stuck at €40K monthly revenue.

After switching to data-driven attribution, the Performance Max campaigns suddenly showed 35 percent more conversion value than under last-click. The algorithm reallocated budget toward upper-funnel asset groups that were previously starved.

Three months later, monthly revenue was at €68K. Same product, same offer — the attribution switch unlocked the budget flow that was already working. The full story is in our Marseille cosmetics case study.

Attribution beyond Google Ads — multi-channel reality

Google Ads attribution only covers Google Ads clicks. It does not see Meta, LinkedIn, organic search, email, or direct traffic.

For accounts with multiple paid channels, Google Ads attribution always overstates Google's contribution because it cannot deduct credit for non-Google touches. The fix is multi-touch attribution at the GA4 or warehouse level.

This matters most for B2B accounts running Google Ads alongside Meta or LinkedIn. We covered the comparison in our Google Ads vs Meta Ads for B2B guide. Multi-channel attribution is the only way to see the true picture.

Common attribution mistakes

These are the patterns we see most often.

Sticking with last-click out of habit. Most accounts above 300 conversions per month should be on data-driven. Habit costs money.

Switching attribution mid-launch. Always switch during a stable window. Mid-launch creates noise that lasts months.

Comparing campaign performance across attribution models. A campaign at €30 CPA under last-click and €45 CPA under data-driven did not get worse — the credit was redistributed.

Letting Smart Bidding optimise on an outdated model. If you upgraded conversions but forgot to update Smart Bidding targets, the algorithm is using new data with old targets.

Ignoring GA4 attribution settings. GA4 has its own attribution model independent of Google Ads. Keep them aligned for consistency.

Frequently asked questions

What is the best attribution model for Google Ads?

Data-driven for most accounts above 300 conversions per month. Last-click for accounts below that volume or pure brand defence.

Will switching attribution change my Google Ads costs?

Indirectly yes. The bidding algorithm redistributes budget based on the new attribution, which can lift cost per click on some campaigns and lower it on others. Total spend usually holds.

Is data-driven attribution available for all accounts?

It requires 300 conversions per month and 3,000 ad interactions in the past 30 days. Below that, Google falls back to last-click.

How is data-driven attribution different from GA4 attribution?

Both use machine learning, but they are trained on different data sets. Google Ads data-driven uses only Google Ads clicks. GA4 data-driven uses all GA4 touchpoints.

Should I use the same attribution model across all my conversion actions?

Usually yes, for consistency. The exception is if you have one conversion event with a very different funnel — for example, a phone call conversion in a single-touch funnel might stay last-click while a form-fill conversion in a multi-touch funnel uses data-driven.

Does attribution model affect Quality Score?

No. Quality Score is independent of attribution. We covered Quality Score mechanics in our Quality Score playbook.

Get an attribution audit

If your Google Ads account looks stuck and you suspect attribution is hiding value, we audit accounts free of charge. We pull your conversion paths, model what each campaign would look like under data-driven, and deliver a switch recommendation in 48 hours.

Book a free 30-minute audit. We screen-share, look at your conversion paths, and walk you through what to change.

Or explore our Google Ads service for the full system we run on accounts spending €5K to €50K monthly.