A multichannel attribution model is a set of rules for splitting conversion credit across every marketing channel a customer touched, not just the last one they clicked. Marketers usually mean one of six models: last-click, first-click, linear, time-decay, position-based, or data-driven. GA4 only lets you select two of these today, which is where most confusion starts.
A multichannel attribution model is a rule, or a set of rules, for deciding how much credit each marketing channel gets for a single conversion. If a customer finds you through organic search, comes back through a retargeting ad two weeks later, then clicks an email the day they buy, the model decides whether email gets all the credit, whether search does, or whether the credit gets shared. Different models answer that question differently, and the answer changes which channels look like they are working.
This matters because almost nobody converts on a single interaction anymore. If your reporting only counts the last click, every channel that opened the door earlier in the journey looks invisible, even when it was doing the heavy lifting. A multichannel model exists to correct that blind spot.
Every attribution conversation eventually comes back to the same six models. Some are still selectable inside GA4, some only exist now in third-party tools and manual analysis, but understanding all six is what lets you read any attribution report correctly, wherever it comes from.
| Model | How credit is split | Best used when | Still in GA4? |
|---|---|---|---|
| Last-click | 100% to the final touchpoint before conversion | Short sales cycles, low data volume, simple reporting | Yes |
| First-click | 100% to the touchpoint that started the journey | Judging which channels build awareness, not close sales | No, retired 2023 |
| Linear | Equal credit split across every touchpoint | A quick, unbiased first look across many channels | No, retired 2023 |
| Time-decay | More credit to touchpoints closer to conversion | Sales cycles with a clear, gradual build-up | No, retired 2023 |
| Position-based (U-shaped) | 40% first touch, 40% last touch, 20% split between | Valuing both discovery and closing channels equally | No, retired 2023 |
| Data-driven (DDA) | Machine learning assigns credit from your own conversion data | Enough conversion volume for the model to find real patterns | Yes, the default |
Last-click gives all the credit to whatever channel the customer interacted with right before converting. It is the oldest, simplest model, and it is still one of only two options GA4 lets you select today. Its strength is that it is easy to explain to anyone. Its weakness is obvious the moment a customer's journey has more than one step: it completely ignores every channel that came before the final click, even the one that introduced the brand in the first place.
First-click flips the logic and gives 100% of the credit to whatever touchpoint started the journey. It is useful for judging which channels are good at generating initial awareness, since it answers "what got this person's attention" rather than "what closed the deal." Google removed it from GA4 and Google Ads in 2023, so today you would need a third-party platform, a CRM report, or a manual spreadsheet model to see it.
Linear attribution splits credit equally across every touchpoint in the path. A customer who touched five channels gives each one 20% of the credit. It is the fairest-looking model on paper and a reasonable starting point when you have no strong reason to weight any stage of the journey more than another, but it treats a passing glance at an ad the same as a deliberate return visit, which is rarely accurate.
Time-decay gives more credit to touchpoints that happened closer to the conversion, on the logic that recent interactions are more likely to have influenced the final decision. It sits between last-click and linear: less extreme than crediting only the last touch, but still weighted toward the end of the journey rather than splitting evenly.
Position-based attribution, often called U-shaped, gives 40% of the credit to the first touchpoint, 40% to the last, and spreads the remaining 20% across everything in between. It is a compromise: it rewards both the channel that created the opportunity and the channel that closed it, while still acknowledging the middle of the funnel mattered a little.
Data-driven attribution uses machine learning instead of a fixed rule. Google Analytics Help explains that DDA evaluates both converting and non-converting paths in your own account, using signals such as time from the key event, device type, and the number and order of ad interactions, to calculate how much each touchpoint actually contributed. It is the only model here that adapts to your specific business rather than applying the same formula to everyone, which is why Google made it the GA4 default.
Google retired first-click, linear, time-decay, and position-based attribution from both GA4 and Google Ads between May and September 2023. Search Engine Journal reported Google's stated reason: those four rule-based models combined accounted for fewer than 3% of conversion actions, a rate Google called too low to keep supporting alongside data-driven attribution. If you are reading an older blog post, a course, or a certification guide that still lists six selectable GA4 attribution models, that content is describing a version of GA4 that no longer exists.
Any conversion action still configured on one of the four retired models was automatically switched to data-driven attribution, unless the account owner manually chose last-click instead.
This is a real gap in a lot of attribution content still online. Plenty of guides describe the six classic models as if they are all live options inside GA4's Attribution settings menu today. They are not. Two are: data-driven and last-click, plus a Google Ads-specific variant of last-click that prioritizes Google Ads clicks in the path.
Data-driven attribution builds a model from your own account's history rather than applying someone else's fixed formula. According to Google Analytics Help, it looks at every path a user took, both the ones that ended in a conversion and the ones that did not, then uses that comparison to estimate which touchpoints actually increased the odds of converting. A click on a channel that shows up constantly in both converting and non-converting paths gets less credit than one that shows up mostly on paths that convert.
The default lookback window is 90 days for most key events, meaning GA4 only considers touchpoints from the 90 days before a conversion when assigning credit. You can shorten that window in Attribution settings if your sales cycle is genuinely faster, which keeps the model from crediting a visit that happened months before someone had any real intent to buy.
Two practical limits are worth knowing before you lean on it. First, data-driven attribution needs enough conversion volume in the lookback window to find a reliable pattern. Google Analytics Help notes that when there is not enough account-specific data, the model can fall back on broader aggregated data instead of your own, which makes the output less precise for smaller sites. Second, DDA only sees what GA4 can track: a phone call your sales team logs in a CRM, or a trade show conversation that leads to a sale three months later, is invisible to it unless you feed that data back in.
To check or change your setting, go to Admin, then Data display, then Attribution settings. From there you can review the reporting attribution model, which channels are eligible for credit, and the lookback window, and GA4's Model comparison report will show you how switching models would have changed your numbers historically. Google Ads has its own version of this same setting, buried inside each conversion action rather than one property-wide toggle, and it is worth a look too since our Google Ads conversion tracking guide covers what most accounts get wrong there.
There is no universally correct model, only the right one for your situation. Use this as a starting checklist rather than a rulebook.
| Your situation | Recommended model | Why |
|---|---|---|
| New site, low conversion volume | Last-click | Data-driven needs data it does not have yet; last-click is simple and transparent |
| Established site, steady conversions | Data-driven (GA4 default) | Enough volume for the model to find genuine patterns in your own data |
| Long B2B sales cycle, sales-assisted | Data-driven plus a CRM-linked or dedicated multi-touch tool | GA4 alone cannot see offline touchpoints like calls or demos |
| Judging top-of-funnel channels specifically | First-click (via a third-party tool) | Answers which channel starts journeys, a question data-driven blends away |
| Quick, unbiased gut check across channels | Linear (via a spreadsheet or third-party tool) | No assumptions about which stage of the journey matters most |
A practical way to sanity-check any model before committing budget to it: pull your last-click numbers and your GA4 Model comparison report side by side. If a channel's contribution swings wildly between models, that is usually the channel worth a second look, either because it is under-credited by last-click or over-credited by an assumption that does not fit how your customers actually buy.
GA4 is not the only place attribution modeling happens, and for some businesses it should not be the last word. Dedicated multi-touch attribution platforms, CRM-linked reporting in tools like HubSpot or Salesforce, and even careful manual spreadsheet models still use the four rule-based models GA4 retired, because those tools are not bound by Google's own product roadmap. If your sales team logs calls, demos, or in-person meetings that never touch your website, none of that shows up in GA4's model at all, no matter which one you pick.
This is also where the classic models earn their keep again. A B2B company with a six-month sales cycle spanning a webinar, three emails, a sales call, and a final proposal page benefits from seeing that journey through more than one lens: data-driven for the digital touchpoints GA4 can see, and a position-based or linear view layered on top through a CRM for the parts it cannot. If you want to test how each classic model would treat a specific customer journey before committing to one, Rankite's free attribution model calculator lets you enter real touchpoints and see first-touch, last-touch, linear, time-decay, and U-shaped attribution side by side in seconds.
Correctly attributing revenue is also the first step before you can trust any channel's reported growth. We saw this directly with Zluri, where cleaning up how conversions were being counted and credited was part of a project that grew their organic traffic by 45%. Without fixing the attribution picture first, it would have been hard to prove which changes actually moved the needle.
Every one of these mistakes has the same root cause: treating attribution as a settings toggle you configure once, instead of a lens you actively check against reality.
What is a multichannel attribution model in simple terms? It is a set of rules for splitting credit for a conversion across the different channels a customer touched before buying, instead of giving all the credit to one click. Instead of asking which channel closed the sale, it asks which channels actually contributed.
Which attribution model does GA4 use by default? Data-driven attribution. Google Analytics Help states GA4 uses machine learning to distribute credit across the touchpoints in a customer's path based on your own account's conversion data, and Google recommends keeping it as the default reporting model.
Can I still use first-click or linear attribution in GA4? Not as a selectable model. Google retired first-click, linear, time-decay, and position-based attribution from GA4 and Google Ads between May and September 2023, citing adoption under 3% of conversions. Only data-driven and last-click remain in the Attribution settings menu, though the concepts still matter for manual analysis and third-party tools.
Is multi-touch attribution the same as multichannel attribution? They describe the same idea from two angles. Multichannel attribution focuses on which channels get credit, while multi-touch attribution focuses on how many touchpoints in the journey get credit. In practice both terms point to any model that splits credit across more than one interaction, as opposed to single-touch models like last-click.
What is the best attribution model for a small business? Last-click, at least to start. Data-driven attribution needs enough conversion volume flowing through GA4 to find real patterns, and a low-traffic site will not have that. Last-click is simple, transparent, and good enough until volume grows, at which point GA4's own data-driven default becomes the better choice.
Does data-driven attribution work for a small website with few conversions? Not well. Google Analytics Help notes that when there is not enough conversion history in the lookback window, GA4 leans on broader aggregated data instead of patterns specific to your account, which makes the output less reliable. Sites with low conversion volume typically get cleaner, more explainable results from last-click.
How do I change the attribution model in GA4? Go to Admin, then under Data display select Attribution settings. Choose your reporting attribution model, set which channels are eligible for credit, and set the lookback window, which defaults to 90 days for most key events. Changing the model is retroactive and will recalculate your historical reports.
Why do so many touchpoints happen before a customer converts? Buyers research across more channels than they used to. NP Digital's consumer survey found the average number of touchpoints before a purchase rose from 8.5 in 2021 to 11.1 in 2025, which is exactly why single-touch models like last-click undercount the channels doing real work earlier in the journey.
Should I use a multi-touch attribution tool outside of GA4? Consider it once your sales cycle stretches across weeks or months, spans offline touchpoints like sales calls, or needs to survive a customer switching devices. GA4's data-driven model only sees what it can track inside its own cookie and login boundaries, so long B2B cycles often need a dedicated multi-touch platform or a CRM-linked model on top.
What is position-based attribution and can I still use it anywhere? Position-based, also called U-shaped, gives 40% credit to the first touch, 40% to the last touch, and splits the remaining 20% across everything in between. GA4 no longer offers it natively, but it is still available in many third-party attribution platforms, in spreadsheet models, and in Rankite's own free attribution model calculator.
Open GA4's Model comparison report today and see how far your last-click numbers drift from data-driven attribution. If the gap is small, you are probably fine on the default. If it is large, that gap is telling you which channels your current reporting is quietly punishing. For a longer sales cycle or an offline-heavy business, pair that GA4 view with our free attribution model calculator to see how the classic models would treat the same journey. And if attribution is only one symptom of a bigger measurement gap, a full Google Analytics audit is usually the faster fix than tuning one setting at a time.
Fill out the form and we'll get back to you within one business day. Prefer email? Write to us directly at contact@rankite.com.
