Pillar · The model your dashboard defaults to
Last-Click Attribution
Is Lying to You
Last-click attribution gives 100% of the credit for every sale to the final touch before purchase — and hides every channel that actually created the customer. It isn't noisy. It's biased in a fixed direction, which is worse, because it always misleads your budget the same way.
Definition
Last-click attribution (also called last-touch attribution) is a rule that assigns all the credit for a conversion to the last marketing touchpoint a buyer interacted with before purchasing. Every earlier touch — the ad that introduced you, the email that nurtured the sale, the review that convinced them — gets exactly zero.
It is still the default lens in most dashboards because it is easy to compute, not because it is true. A buyer's journey has a beginning, a middle, and an end; last-click keeps only the end and then presents that fragment as the whole story. Everything below is the anatomy of that lie — and what to use instead.
The mechanism: it over-credits closers and starves creators
Think about which touches sit at the end of a buying journey. Branded search — the buyer already knows your name. Direct traffic — they typed your URL. Retargeting — they had already visited. Email — they were already on your list. These are closing touches. They harvest demand that something else created.
Last-click hands each of those closers the whole sale, every time. So the discovery channels — the TikTok video, the prospecting campaign, the article that introduced the problem — report terrible numbers precisely because they do their job at the start of the journey. Turn one off, and weeks later the "high-performing" closer channels quietly shrink too, because the demand they were harvesting stopped arriving.
That is the practical cost: budget flows toward the channels that collect credit and away from the channels that cause revenue. Not because anyone made a mistake — because the measurement rule itself points the money the wrong way.
Three places you can watch it lie in your own GA4
Branded search and direct “win” journeys they never started
Open your conversion paths and look at how many end in branded search or direct. Those buyers already knew you — some earlier touch made that happen. Last-click credits the reunion, not the introduction.
Email looks unbeatable because it fires last
Email lands on people who already gave you their address — the warmest audience you have. It closes sales it did not create, and last-click reads that closing position as causation.
Your platforms claim more sales than you shipped
Each ad platform attributes the same orders to itself under its own last-touch-flavored rules. Add their claimed conversions and compare with your actual order count — the total routinely exceeds what you sold. They cannot all be right. Under last-click logic, they all are.
The €99 second opinion
Last-click is lying to you. Find out what it's costing you — €99, once.
Watch the model work on a sample store first, no signup. Then upload your GA4 export and get each channel's incremental contribution — the sales it actually caused — with confidence intervals, in 5–10 minutes. No pixel, no subscription, no setup call.
The alternatives to last-click attribution, compared
Most "alternatives" swap one credit-splitting rule for another. That changes who gets flattered; it does not change the fact that a rule is describing the journey instead of measuring what caused the sale. Here is the honest map:
| Model | What it rewards | Where it misleads |
|---|---|---|
| Last-click / last-touch | The closing touch | Inflates branded search, direct, retargeting, email; starves discovery channels |
| First-touch / first-click | The discovery touch | The same lie, mirrored — ignores everything that nurtured and closed the sale |
| Linear / position-based | Every touch, by formula | The split is arbitrary — being present in a journey is not the same as causing the purchase |
| Platform "data-driven" | What the platform's model values | Scored by the referee who is also a player; sees only its own touches; hides its uncertainty |
| Causal measurement | Incremental sales a channel caused | Needs real statistical care — which is why every estimate should ship with a confidence interval |
The fix is a different question, not a different split
Every rule-based model answers "who touched this sale?" The question your budget actually needs is "what would have happened without this channel?" That is a causal question, and it is answerable from data you already have: your own sales history contains natural variation — spend going up and down, channels pausing, seasonality — that a causal model uses to estimate each channel's incremental contribution, the way economists evaluate policy.
That is what causal attribution does: revenue caused, not revenue claimed, per channel, with a confidence interval on every estimate so you know how sure the model is before you move money. The difference between that read and your last-click report is, concretely, what the lie has been costing you.
Last-click attribution: frequently asked questions
What is last-click attribution?
+
Last-click attribution assigns 100% of the credit for a sale to the final marketing touchpoint before purchase. If a buyer discovers you on TikTok, reads two emails, then buys after a branded Google search, last-click records that sale as caused by Google search — and the channels that actually created the customer get nothing.
Why is last-click attribution inaccurate?
+
Because the last touch is usually the closing step of a journey some other channel started. Branded search and direct traffic sit at the end of journeys almost by definition, so last-click systematically inflates them while starving the discovery channels. The error is structural, not random: it always flatters the same kinds of channels.
What is the best alternative to last-click attribution?
+
Rule-based alternatives (first-touch, linear, position-based) redistribute credit but still describe the journey rather than measure causation. The stronger alternative is causal measurement: estimating what each channel incrementally caused by comparing against what would have happened anyway. That's the question a budget decision actually needs answered.
Is data-driven attribution better than last-click?
+
Usually, but with a catch: platform data-driven attribution is trained and reported by the platform whose ads it evaluates, only sees the touches that platform can observe, and rarely shows its uncertainty. It fixes some of last-click's mechanics while keeping the referee-and-player conflict of interest.
How do I find out what last-click is costing my store?
+
Run a causal read on your own history: upload 40–90 days of your GA4 export to Causality Engine and you get per-channel incremental contribution with confidence intervals in 5–10 minutes, for €99, once. The gap between that and your last-click report is what the lie is costing you.
Is first-touch attribution more accurate than last-click?
+
No — it's the same rule pointed at the other end of the journey. First-touch gives everything to the discovery touch and nothing to what nurtured or closed the sale. Any model that assigns credit by position describes the journey; none of them measure what caused the purchase.
The €99 second opinion
Last-click is lying to you. Find out what it's costing you — €99, once.
Watch the model work on a sample store first, no signup. Then upload your GA4 export and get each channel's incremental contribution — the sales it actually caused — with confidence intervals, in 5–10 minutes. No pixel, no subscription, no setup call.
Go deeper
The last-click file, by question
The best alternative to last-click
If you replace it with one thing, replace it with this.
Last-click attribution problems
The failure modes, catalogued one by one.
First-touch vs last-touch
Why both ends of the journey tell the same lie.
Last-click vs data-driven
What the platforms' own fix gets right — and what it can't.
Multi-touch vs last-click
Why the two reports will never match.
Email's inflated numbers
The channel last-click flatters most.
Google Ads beyond last-click
The platform view, taken apart for Shopify.
Causal attribution
The measurement that answers the budget question.
Real reports
What a causal read looks like on real data.
Anonymised reports from stores that ran the €99 analysis on their own GA4 exports — per-channel incremental contribution with confidence intervals.
Browse the report library →For Ecommerce brands on GA4 · Defensible in a budget meeting
Five things to know before you upload.
If you run an Ecommerce site with Google Analytics 4 installed, you are a fit. Any ecommerce platform (Shopify, WooCommerce, BigCommerce, custom) works as long as GA4 is the analytics layer.
Proprietary causal-inference model
Not an LLM. Not last-click in a trench coat. Our model is the same statistical machinery used to evaluate medicine and policy, applied to your Shopify and GA4 data.
Confidence intervals on every estimate
Honest uncertainty, not a single confident-looking number. A causal claim without an interval is a guess in a suit.
Methodology open on request
Every assumption documented: prior, functional form, covariate set, robustness checks. The methodology document ships to any customer who asks. The goal is a number you can defend in a budget meeting.
EU data residency. First-party only.
Your Shopify and GA4 exports are processed inside the EU and never sold. No pixel, no SDK, no third-party tracking. GDPR-compliant by construction.
No engineering ticket
Standard exports from Shopify and GA4 go in. Two minutes of setup, no developer needed, no 90-day onboarding, no platform migration.
Want the methodology document? Email hi@causalityengine.ai. Reply within one business day. Or jump to pricing or the interactive demo.