Marketing Attribution Models Explained: Why GA4, Meta and Google Ads Disagree

Here are marketing attribution models explained in one line: an attribution model is the rule that decides which ads, emails and visits get credit for a sale. Last click, first click, linear, time decay, position-based, data-driven and custom models each split that credit differently. That is a big reason GA4, Meta and Google Ads rarely agree on your results.

What is a marketing attribution model?

A marketing attribution model assigns credit for a conversion to the touchpoints that came before it. Google's GA4 attribution settings page describes it as "a rule, a set of rules, or a data-driven algorithm" that determines how credit is assigned.

The model never changes how many sales you made. It changes which channel looks responsible for them, and so where next month's budget goes.

Marketing attribution models explained: the seven you will meet

Seven models cover almost every report. All except last click and first click are multi-touch attribution models:

• Last click: 100% of the credit goes to the final touch before the conversion. Simple and stable, but blind to whatever created the demand.
• First click: 100% goes to the touch that introduced the buyer. Useful for judging prospecting, blind to what closed the sale.
• Linear: credit is split evenly across every touch on the path.
• Time decay: touches closer to the conversion get more credit, usually on a half-life where credit halves for every set number of days back.
• Position-based (40/20/40): 40% to the first touch, 40% to the last, and the remaining 20% shared across the touches in between.
• Data-driven attribution: a machine learning model compares the paths of people who converted with those who did not. Google says its GA4 model uses "a counterfactual approach".
• Custom: you set a weight for each touch or position, for example extra credit for a demo request in a B2B journey.

How does each attribution model split one sale?

Each model hands the same sale to different channels. Here is one AED 1,200 order run through every model (read it as £1,200 in the UK). The journey and figures are illustrative.

1. Day 1: the buyer clicks an Instagram ad and browses.
2. Day 8: they click a non-brand Google search ad.
3. Day 15: they click a link in your email newsletter.
4. Day 22: they search your brand name, click a Google brand search ad and buy for AED 1,200.

The four touches happen 21, 14, 7 and 0 days before the purchase. This is how the AED 1,200 is credited:

• Last click: brand search AED 1,200, everything else AED 0.
• First click: Instagram AED 1,200, everything else AED 0.
• Linear: AED 300 to each touch.
• Time decay (7-day half-life, for illustration): Instagram AED 80, non-brand search AED 160, email AED 320, brand search AED 640.
• Position-based: Instagram AED 480, non-brand search AED 120, email AED 120, brand search AED 480.
• Custom (example weights 30/30/10/30): Instagram AED 360, non-brand search AED 360, email AED 120, brand search AED 360.
• Data-driven: no fixed answer. The split depends on how often paths with and without each touch convert in your own data.

Instagram swings from AED 0 under last click to AED 1,200 under first click. Judge prospecting on last click and you will probably cut it.

Why do GA4, Meta and Google Ads report different numbers?

Each platform measures something different: different touches, windows, reporting dates and, sometimes, estimated conversions. None of them de-duplicates against the others.

Each platform credits its own ads

Suppose the buyer above also scrolled past an Instagram video ad on the morning of the purchase without clicking. Meta's Marketing API reference says its default reporting window means 7-day click plus 1-day view, so Meta reports the AED 1,200 sale. Google Ads reports it too, through the brand search click.

Add the two dashboards and you see AED 2,400 of revenue for one AED 1,200 order.

Attribution windows do not match

A lookback window is how far back a touch can still earn credit. The defaults differ:

• GA4: 90 days for most key events and 30 days for acquisition events such as first_visit, according to Google's attribution settings help.
• Google Ads: a 30-day default click-through window, adjustable from 1 to 90 days, plus a 1-day view-through and 3-day engaged-view default, per Google's conversion windows article.
• Meta: 7-day click and 1-day view by default.

The day 1 Instagram click sits inside GA4's 90-day window but outside Meta's 7-day click window, so GA4 can credit it and Meta cannot.

Views count in ad platforms, not in most analytics

Ad platforms count conversions after an ad was seen but not clicked. GA4 never sees an impression on Instagram, TikTok or Snapchat, so the more video-heavy your mix, the wider the gap.

Some conversions are modeled

When Google cannot observe a conversion, for example because cookies were declined under consent mode or the journey switched devices, it estimates it. Google's modeled conversions article says that in the Conversions column, "Google reports both modeled and observed conversions." Your order system contains no estimates.

The same sale lands on different dates

Google Ads reports a conversion on the date of the ad interaction, while GA4 uses the date it happened. In Google's Analytics Help example, a key event on May 10 after an ad click on May 5 shows on May 10 in Analytics and May 5 in Google Ads.

Which attribution models can you use in GA4, Google Ads and Meta?

As of late 2025, GA4 and Google Ads offer data-driven and last-click models only. Meta reports through click and view windows rather than the classic models.

Attribution model in GA4

GA4 offers data-driven, paid and organic last click, and Google paid channels last click, according to Google's GA4 attribution guide. It states that "The first click, linear, time decay, and position-based attribution models are no longer available as of November 2023." All GA4 models exclude direct visits from credit unless the whole path is direct.

Attribution model in Google Ads

Google Ads offers data-driven and last click. In its April 6, 2023 announcement, Google said "less than 3% of Google Ads web conversions are attributed using first click, linear, time decay, or position-based models". It planned to remove those options from June 2023 and switch remaining conversion actions to data-driven from September 2023.

Attribution model in Meta Ads

Meta credits a conversion to its ads when it falls inside the attribution setting, 7-day click and 1-day view by default. A shorter setting such as 1-day click lowers reported results without changing what happened.

How do you choose an attribution model for your business?

Choose by the decision you are making and the length of your sales cycle. Most teams need two views: each platform's own model for automated bidding, and one cross-channel model on de-duplicated conversions for budget decisions.

DTC and ecommerce brands

Journeys are usually short. Data-driven or last click inside each ad platform keeps bidding stable. For budget splits, a position-based view on de-duplicated orders stops brand search from absorbing credit that prospecting earned.

B2B and long sales cycles

Deals take months and involve several people. GA4's longest lookback is 90 days, so a six-month deal can lose its first touches entirely. Use position-based or a custom model that weights first touch, demo request and opportunity, and bring CRM stages in so the model sees pipeline, not just form fills.

Our B2B SaaS marketing page covers long-cycle attribution and includes a free CAC payback calculator.

A five-step way to decide

1. Pick one source of truth for conversions: your order system or CRM, not an ad platform.
2. De-duplicate, so each order counts once however many platforms claim it.
3. Choose one cross-channel model and write down why.
4. Leave platform bidding on the platform's model and judge budgets on your cross-channel view.
5. Compare models quarterly. Where two models disagree sharply on a channel, test it with a holdout.

What should marketers in the UAE, Saudi Arabia, Qatar and the UK watch for?

The models are global, but local buying habits decide which gaps hurt most.

• Cash on delivery (UAE, Saudi Arabia, Qatar): a COD order fires a purchase event at checkout even if the customer later refuses delivery. Reconcile attributed sales with delivered orders before you trust any ROAS.
• WhatsApp and phone sales: orders closed in a chat or on a call never reach the ad platforms unless you upload them as offline conversions, so every model undervalues the ads that started them.
• Snapchat and TikTok heavy plans: check how many of their reported conversions in Riyadh, Dubai or Doha come from views before you move budget.
• Peak weeks: around White Friday and Eid, buyers compare offers within days and last click piles credit onto brand search and retargeting. Our White Friday playbook for the GCC covers that peak.
• Consent in the UK: some visitors decline cookies, so GA4 sees fewer journeys than happened. With consent mode, Google models part of the gap, one more reason Google Ads can show more conversions than GA4.

How Grower helps

Grower's attribution software sits in the Streamline module. Its attribution engine de-duplicates conversions from every connected channel and app into one unified conversion count, so a sale that Meta and Google Ads both claim is counted once.

You then credit each touchpoint with the model you choose: last click, first click, linear, time decay, position-based, or a custom model with your own weights. Streamline integrates with Google Ads, Meta Ads, TikTok Ads, Snapchat Ads, LinkedIn Ads, Google Analytics, HubSpot, AppsFlyer and Adjust, so ecommerce, B2B and app journeys sit in one view.

Frequently asked questions

What is the best marketing attribution model?

No single model is best. Last click suits quick decisions on short journeys, and position-based gives a fairer view of prospecting and closing channels. Data-driven works well inside an ad platform with enough conversions to learn from. For most teams, the best setup is one cross-channel model on de-duplicated conversions, plus each platform's own model for bidding.

Why does Meta report more conversions than GA4?

Meta counts conversions within seven days of a click and within one day of a view by default. GA4 never sees Meta ad impressions and only credits sessions it can track. Meta also claims every conversion it can link to its own ads, even when another channel closed the sale. Together, these usually push Meta's count above GA4's.

Can I still use position-based attribution in Google Ads?

No. Google Ads now offers data-driven and last click attribution only. Google retired first click, linear, time decay and position-based models in 2023 and moved conversion actions still using them to data-driven. To see position-based results, use a cross-channel attribution tool that applies the model to your own de-duplicated conversions.

How long should my attribution window be?

Match it to your buying cycle. Check the typical time from first visit to purchase in your own data, then pick windows that cover most journeys. A fast-moving ecommerce brand may find 7 or 30 days covers most sales. For B2B deals that take months, GA4's 90-day maximum may be too short, so bring CRM data into your attribution.

Want your channels on one de-duplicated conversion count, under the model you choose? Book a Grower demo and we will walk through attribution on your stack.