Last-click attribution gives 100% of the credit for a sale to the thing that happened five minutes before the conversion. That’s not measurement. That’s alibi construction. It’s the equivalent of walking into a courtroom, pointing at the closing lawyer, and saying “this person built the entire case” — while the detective, the witnesses, the six months of investigation, and every piece of evidence that actually made the case possible get nothing. Zero credit. As if they weren’t there.
First-click isn’t much better. It credits the thing that first got someone’s attention — sometimes six months before they were remotely ready to buy — and calls that the cause. Which is a bit like crediting the movie trailer for the Oscar, rather than the actual movie.
Both attribution models are telling you a story. Neither is telling you the full one. Here’s what each actually measures, where each breaks down, and what to use instead when you need to make a real budget decision.
What Last-Click Attribution Actually Does
Last-click attribution assigns 100% of the conversion credit to the final touchpoint before a user converts. Clicked a Google ad and then bought something? The ad gets the sale. Found the site via organic search, read three articles, came back via direct two weeks later, clicked a retargeting ad, and then converted? The retargeting ad gets the sale. Everything else gets nothing.
It’s the default model for most ad platforms because it makes the platform’s own ads look fantastic. Google Ads defaults to last-click within its own reporting. Meta does too. The platforms that benefit most from last-click attribution built last-click into their dashboards. This is not a coincidence.
Last-click is popular for a second reason: it’s simple. One touchpoint, one conversion, one clear line of causation. For leadership teams who want a clean story about where leads come from, it delivers one. The story just happens to be wrong most of the time.
What First-Click Attribution Actually Does
First-click attribution flips the logic entirely. 100% of the credit goes to the very first touchpoint — the thing that introduced a user to the brand or product in the first place. Find a company via a LinkedIn post? LinkedIn gets the conversion credit, even if the actual purchase happened seven weeks and twelve touchpoints later.
The appeal here is obvious. First-click makes content, organic search, and brand awareness look like the drivers of revenue — which in long sales cycles, they often are. It’s a corrective to the last-click bias. But it overcorrects. It makes the last three things a buyer did look completely irrelevant, which is rarely true either.
Why Both Are Wrong in Most Real Funnels
Real buyer journeys don’t look like a single click followed by a purchase. They look like this: someone discovers a brand via organic search, reads an article, leaves. Sees a LinkedIn post two weeks later, reads two more articles. Gets retargeted with an ad, clicks through to a pricing page. Signs up for a newsletter. Gets a nurture email. Schedules a demo. Converts.
First-click gives all the credit to the organic search hit. Last-click gives all the credit to whatever triggered the demo booking. Neither model captures the actual purchase journey. And if you’re using either model to allocate budget, you’re defunding the channels that were actually doing useful work — you just can’t see which ones because your attribution model isn’t designed to show you.
This is the core tension documented in What Is Marketing Attribution and Why It Matters: different attribution models answer different questions, and the question “which channel gets the budget?” requires a model that accounts for the full journey — not just the beginning or the end.
The Multi-Touch Models Worth Understanding
Between first-click and last-click sit a range of multi-touch models that distribute credit across the full journey:
- Linear attribution distributes credit equally across every touchpoint. If there were six interactions, each gets 16.7%. Fairer than single-touch models, but it assumes every interaction mattered equally — which is almost never true. A homepage visit and a pricing page visit are not equivalent signals.
- Time decay attribution gives more credit to touchpoints closer to the conversion. This makes intuitive sense for short sales cycles. It makes less sense for a six-month enterprise deal where the initial discovery event was as important as anything that followed.
- Position-based (U-shaped) attribution splits 80% of the credit between the first and last touchpoints, distributing the remaining 20% across the middle. This is a reasonable compromise. It acknowledges both the introduction and the close while not completely ignoring the middle. It’s also arbitrary, because there’s no data-driven reason why 40/20/40 is the right split for any specific funnel.
- Data-driven attribution uses machine learning to distribute credit based on observed patterns in your actual conversion data. According to GA4’s documentation, this is the default attribution model in GA4 — and it’s the right default for most businesses, with one important caveat covered below.
Data-Driven Attribution: The Actual Right Answer (With Caveats)
Data-driven attribution is better than any fixed-rule model because it doesn’t assume anything about which touchpoints matter. It observes which combinations of touchpoints actually correlate with conversions across your real data, and distributes credit accordingly. Google’s own research on attribution consistently shows that data-driven outperforms single-touch models in predicting which campaigns will produce results.
The caveat: data-driven attribution requires a minimum volume of conversions — roughly 400 per month across your property — to produce statistically meaningful results. Below that threshold, GA4 falls back toward last-click behavior without clearly announcing it has done so. If you’re running a business with 30 conversions a month, data-driven attribution is doing less work than you think it is.
And the model is only as reliable as the tracking underneath it. A data-driven attribution model built on top of broken conversion tracking produces sophisticated-looking charts that are confidently wrong. This is the exact problem covered in How to Fix Conversion Tracking in GA4 Step by Step — the fix has to happen before the attribution model means anything.
When First-Click or Last-Click Is Actually Useful
Here’s the honest answer: both models have legitimate use cases. They’re just not attribution models — they’re diagnostic tools.
First-click is useful for answering “what’s driving discovery?” — which channels are introducing new audiences to the brand. If you’re trying to decide whether to invest in a new content channel or a new platform, first-click data tells you where people are finding you for the first time. That’s a legitimate signal.
Last-click is useful for answering “what closes deals?” — which channels and messages appear right before a conversion. If you’re optimizing the bottom of the funnel and trying to understand what triggers the final action, last-click data is relevant. Just don’t use it to evaluate top-of-funnel channels, because by definition it can’t see them.
The mistake isn’t using either model. It’s using either model as the single source of truth for budget allocation across the entire funnel. That’s the decision that HubSpot’s State of Marketing research consistently identifies as one of the top sources of misallocated marketing spend — and the one that Marketing Week has documented as a recurring theme in attribution debates.
What Attribution Guides Conveniently Leave Out
- View-through attribution is where budgets quietly inflate. Most ad platforms offer view-through attribution — crediting a conversion to an ad the user saw but never clicked. Meta’s default includes a 1-day view-through window. That means any conversion within 24 hours of someone scrolling past your ad gets credited to the campaign, even if they never clicked it. Understand what window your platforms are using before reading performance numbers.
- Cross-device journeys make every model less accurate. A user who discovers you on mobile, researches on desktop, and converts on mobile again is being tracked as three separate anonymous sessions by most attribution systems — unless they’re logged into a platform that can connect the dots. The resulting attribution gap is structural and real.
- The attribution window is a setting, not a fact. GA4’s lookback window is adjustable in Admin → Attribution Settings. Most accounts have never touched it since setup. A 30-day lookback on a 90-day sales cycle misattributes a meaningful percentage of conversions. Match the window to your actual buying cycle.
- Changing attribution models doesn’t change reality — it changes what you can see. Switching from last-click to data-driven doesn’t move budget to the right channels automatically. It shows you that the budget was in the wrong place. You still have to move it.
- Attribution accuracy depends entirely on what happens before attribution. If events are misfiring, if parameters are missing, if server-side and client-side are double-counting without deduplication — the attribution model is processing garbage. The measurement layer has to be right first. Every time.
The Visibility vs Conversion Question Is an Attribution Question in Disguise
The debate about whether to invest in brand visibility or conversion optimization is, underneath, an attribution debate. Teams that run on last-click attribution systematically undervalue visibility channels because those channels never appear as the last touch. Teams that understand the full attribution picture make different decisions — they see that the organic article published eight months ago is still the first touchpoint for 30% of the deals that closed this quarter.
The framework for making that decision correctly is covered in How Growth Engineers Prioritize Visibility vs Conversion Goals. The short version: get the attribution honest first, then make the call.
The Right Model for Your Business
If you have more than 400 conversions per month and your tracking is clean: use data-driven attribution in GA4, set the lookback window to match your sales cycle, and stop looking at platform-native last-click numbers as if they’re telling you the same story.
If you have fewer than 400 conversions: use position-based attribution as a reasonable approximation, understand that you’re applying a rule of thumb rather than statistical inference, and make sure the tracking is solid before worrying about which model to run.
In both cases: understand that attribution models are lenses, not cameras. They help you see the data differently. They don’t capture everything that’s there.
This is the measurement foundation behind the attribution model work OJC Labs sets up for clients — clean tracking first, then the right model for the actual sales cycle, then decisions that reflect what’s really happening rather than what the last-click dashboard wants you to believe.
If your attribution setup doesn’t match your buying cycle, see how we’ve fixed this in practice.