Google Ads Attribution Models: Which One Should You Use?

1. What Is a Google Ads Attribution Model?

Quick Answer: A Google Ads attribution model is the rule that decides which ads, clicks and keywords get credit for a conversion when a customer touches several before buying. It does not change what you spend or earn — only how credit is shared across the journey. That, in turn, shapes what your reports say is working and where your budget flows.

Real buying journeys are rarely one click. A Kuala Lumpur shopper might see your Display ad on Monday, search your brand on Wednesday, then click a Search ad on Friday and submit a form. Four touches, one sale. Someone has to decide who gets the credit — and that someone is your attribution model.

The conversion itself can be many things, and the model credits all of them the same way — whether it is a purchase, a lead form submission, or a phone call from a call-focused ad. One point trips people up constantly: attribution is not the same thing as tracking. Tracking is what records that a conversion happened at all; the model only decides how to split the credit for it. If your tracking is broken, no model can save you — which is why clean conversion tracking comes first, always.

2. The 2026 Reality: Six Models Became Two

Quick Answer: Google Ads once offered six attribution models. In 2026 only two remain — Data-Driven Attribution and Last Click. Google began retiring first click, linear, time decay and position-based in 2023 and finished that September, moving every affected conversion action to data-driven. So the real choice today is DDA or Last Click, nothing else.

This is the single most important thing to understand, because it makes the decision far easier than older articles suggest. Starting in mid-2023, Google confirmed it was sunsetting four rule-based models across Ads and Analytics — first click, linear, time decay and position-based — which together accounted for less than 3% of conversions. By September 2023 they were gone, and the conversion actions using them were upgraded to data-driven.

Why did Google cut them? The four retired models were rule-based — they followed a fixed formula a human picked, whether or not that formula matched real behaviour. Data-driven attribution reads your actual account instead of guessing, so Google made it the centre of gravity and let the blunt rules go.

3. Data-Driven Attribution (DDA), Explained

Quick Answer: Data-driven attribution uses Google’s machine learning to share conversion credit across every ad interaction in the path, based on how much each one actually moved the needle for your account. It is now the default for new conversion actions and works across Search, Shopping, Display, YouTube and Demand Gen.

Instead of a fixed rule, DDA compares the paths that converted with the paths that did not, then works out which touches genuinely made a difference. A click that reliably shows up before a sale earns more credit; one that rarely matters earns less. Per Google’s data-driven attribution documentation, this now runs across Search, Shopping, Display, YouTube and Demand Gen in one model.

Two practical points matter for Malaysian SMEs:

• It is the default now. Google made data-driven the default model for new conversion actions, so unless someone changed it, your account is likely already on DDA.
• The old volume gate is gone. DDA once needed roughly 3,000 ad interactions and 300 conversions in 30 days before you could switch it on. That hard minimum has been removed — every conversion action is now eligible — though the model still reads most accurately once you have meaningful, steady conversion volume behind it.

Because DDA looks at the whole path, it finally gives credit to assisting channels a last-click view ignores — the Demand Gen campaigns and Display impressions that warmed the buyer up. It also decides how to weigh clicks from places like Google Search Partners based on whether they actually contribute, not on a flat rule.

4. Last Click: When It Still Makes Sense

Quick Answer: Last Click gives 100% of the credit to the final clicked ad and keyword. It is simple and predictable, and it still fits a few cases — very low conversion volume, genuinely single-touch buying journeys, or when you deliberately want to reward only the closing click. For most multi-touch accounts, though, it over-credits the last step.

Last Click is not wrong — it is just narrow. It answers one question well: which click sealed the deal? For a business where buyers search once and convert on the same click, that is the whole story, and the simplicity is a feature.

The times Last Click still earns its place:

• Very low conversion volume. If you get a handful of conversions a month, DDA has little pattern to learn from, and a simple, stable last-click number is easier to trust.
• Short, single-touch journeys. When people find you and buy in one visit, there is barely a path to distribute credit across.
• You want hard, predictable numbers. Some owners prefer the same fixed logic every time — the same instinct behind managing bids and bid adjustments by hand rather than handing them to a model.
• Very tight geography. A single-suburb service business steering leads through precise location targeting often has such short paths that last-click and data-driven barely differ.

The catch: Last Click systematically over-rewards the closing click — usually brand searches and bottom-funnel terms — while the campaigns that created the demand look worthless. That distortion is exactly what you will see move in the next two sections.

5. Data-Driven vs Last Click, Side by Side

Quick Answer: The practical difference is simple: Last Click rewards the final click only, while DDA spreads credit across the whole path using your account’s own data. DDA usually surfaces assisting campaigns Last Click hides; Last Click gives you a simpler, more predictable number. For most Malaysian SMEs with real multi-touch journeys, DDA reflects reality better.

Here is how the two live models compare on the dimensions that actually affect your decisions.

One more thing the model quietly controls: the conversion window it reads across. If your sales bunch around promotions, pair your attribution view with proper seasonality adjustments, and read the crowded periods alongside auction insights so you know whether a shift came from your model or the competition.

6. What Changes When You Switch to DDA

Quick Answer: When accounts move from Last Click to DDA, total conversions stay the same but the credit moves. Brand and bottom-funnel campaigns lose some; assisting campaigns like Display, Demand Gen and non-brand Search gain. That shift can change which campaigns look profitable — and therefore where your budget should go.

The number that does not change is total conversions. Nobody bought more or less; the credit for the same sales simply gets reassigned. Across ZenWeb-managed accounts, the pattern is consistent: last-click was quietly overpaying the closing click, and DDA hands some of that credit back up the funnel.

Read that as a warning about last-click habits. Under Last Click, it is easy to conclude that Demand Gen and upper-funnel campaigns “do not convert” and cut them — right as they were quietly feeding the brand searches you were praising. Before you act on any shift like this, prove it on your own account with a controlled Google Ads experiment rather than a gut read.

7. Attribution, Your Bidding, and Why GA4 Disagrees

Quick Answer: Your attribution model is not just a reporting choice. Smart Bidding optimises toward whatever it counts as a conversion, so the model quietly steers your budget. It is also a leading reason Google Ads and GA4 report different numbers — the two systems can attribute the same sale in different ways.

This is where the model earns its keep. If you use Smart Bidding instead of manual CPC, the algorithm chases the conversions your attribution model credits. Feed it last-click credit and it optimises toward closing clicks; feed it data-driven credit and it learns to value the assists too. In other words, the model and the bidding are joined at the hip — a point worth remembering when you read how Smart Bidding actually works.

Attribution is also the usual suspect when your platforms disagree. If Google Ads and GA4 show different conversion numbers, differing attribution and windows are a common cause — the same sale, credited two ways. And most ZenWeb-managed accounts have quietly settled onto one model.

The takeaway from that mix is not “everyone picked DDA” — it is that DDA is the default and, for most accounts, the right one to leave in place. The handful on Last Click chose it for a reason: very low volume or a deliberately simple read.

8. So, Which Attribution Model Should You Use?

Quick Answer: Use Data-Driven Attribution unless you have a specific reason not to — it is the default, it reads the whole journey, and it has no volume gate anymore. Choose Last Click only for very low-volume or genuinely single-touch accounts. Above all, fix your conversion tracking first, because no model can fix bad data.

Run your account through this short checklist. The more boxes DDA ticks, the more clearly it is your model:

• Tracking is clean. Deduplicated, counting real leads and sales — not junk. If not, fix this before touching the model; start with the basics of what conversion tracking does.
• Journeys are multi-touch. Buyers see and click more than once before converting? DDA reflects that; Last Click hides it.
• You have steady volume. A regular flow of conversions gives DDA the pattern it needs to read your account well.
• The account is tidy. A sound account structure makes any attribution read cleaner — messy campaigns give messy credit.
• You will not chase every prompt. Do not let a 100% optimisation score or a one-click suggestion talk you into a change your data does not support.

For most Malaysian SMEs, that adds up to one answer: stay on DDA, keep it stable, and put your energy into the tracking beneath it. This is the exact read our team makes on every Google Ads account we manage — not “which model is fashionable,” but “which one this account’s data can actually support.”

9. Frequently Asked Questions

How many attribution models does Google Ads have in 2026?

Two: Data-Driven Attribution and Last Click. Google retired the other four — first click, linear, time decay and position-based — during 2023 and moved every conversion action that used them to data-driven attribution. Any guide still listing six models is out of date.

What is the default attribution model in Google Ads?

Data-driven attribution. Google made DDA the default for new conversion actions, so unless someone deliberately changed the setting, your account is almost certainly already using it. You can still switch a conversion action to Last Click if you have a specific reason.

Is data-driven attribution better than last click?

For most accounts, yes. DDA spreads credit across the whole path using your own data, so it surfaces the assisting campaigns Last Click ignores. Last Click is only the better pick for very low-volume or genuinely single-touch accounts, where there is little path to model.

Do I need a minimum number of conversions for data-driven attribution?

Not anymore. DDA once required roughly 3,000 ad interactions and 300 conversions in 30 days, but Google removed that hard minimum — every conversion action is now eligible. The model still reads most accurately once you have steady, meaningful conversion volume behind it.

Why do Google Ads and GA4 show different conversion numbers?

Different attribution settings and conversion windows are a common cause. Google Ads and GA4 can credit the same sale in different ways, so their totals rarely match exactly. Attribution differences, tracking setup and timing windows together explain most of the gap you see.