What each model actually credits
First touch attribution gives 100% of the credit for a conversion to the very first tracked interaction a person had with your brand. Last touch attribution gives 100% of the credit to the final tracked interaction that happened immediately before they converted. Everything in between, every touch the person made on the way from first to last, gets nothing in either model. That's the whole mechanism, and it's also why the two models can point at completely different channels for the exact same sale.
Take a real sequence. Someone taps a link from an Instagram Reel, browses for two minutes, and leaves without buying. Three days later they search your brand name on Google and buy from the first result. First touch attribution credits Instagram, because that was the first recorded interaction. Last touch attribution credits Google, because that branded search was the last thing that happened before checkout. Neither model is wrong. They are answering different questions about the same customer.
The same disagreement shows up with email. A subscriber joins your list from a giveaway post, gets nurtured for a month, then clicks a promo email and buys. First touch credits the giveaway post that got them on the list in the first place. Last touch credits the email. Both are true descriptions of what happened, and a business that only ever looks at last touch will keep concluding that email is its best channel, while quietly starving the content that fills the list email depends on.
First touch vs last touch, side by side
| First touch attribution | Last touch attribution | |
|---|---|---|
| Credits | The first touchpoint that brought the person into your funnel | The final touchpoint immediately before conversion |
| Best for | Judging which channel introduces new people, awareness and content spend | Judging which channel closes people, retargeting and bottom-of-funnel spend |
| Typical winner | Organic social, content, influencer posts, SEO | Branded search, email, retargeting ads |
| Weakness | Ignores whatever actually closed the sale | Ignores whatever actually created the demand |
| Matches ad platform reporting | Rarely, platforms report their own last touch by default | Yes, this is what most ad platforms show unless you change the setting |
Why platforms double count the same sale
This is where it gets messy once you run more than one paid channel. Meta reports a sale as attributed to a Meta ad whenever that ad was the last Meta touchpoint the person had inside Meta's own attribution window, regardless of whether they also clicked a Google ad afterward. Google does the same thing with its own last click, inside its own window, blind to whatever happened on Meta. Neither platform can see the other's touchpoints. So the same single sale gets full credit from Meta's dashboard and full credit from Google's dashboard, at the same time, for the same purchase.
That's why adding up the attributed revenue or attributed conversions shown across every ad platform's own reporting reliably produces a number higher than actual total revenue, sometimes by a wide margin. It isn't fraud and it isn't a bug. It's what happens when every observer runs last touch attribution over its own partial view of the same person, with no visibility into what the other platforms saw.
Why some tools no longer show you last touch by default
Google Analytics 4 defaults to a data-driven attribution model rather than plain last click, which means the channel report you see there is already running an algorithmic model behind the scenes, not the simple last touchpoint. That's a reasonable default for a large account with enough conversion volume to train the model, but for a small account it can produce a channel breakdown that looks nothing like the raw click order you'd expect, purely because the model is guessing at credit with too little data to guess well. When a data-driven report disagrees sharply with what you'd expect from looking at the actual clicks, switching the comparison to a plain last click view, or checking a link tracking tool's raw click history, is a useful sanity check before trusting the fancier number.
Which one should you actually use
For most small operators running content plus one or two paid channels, last touch is the better operational default, because it answers the question that decides where budget goes next: what to keep paying for. But check first touch on a regular cadence too, because a channel that never shows up in last touch but keeps showing up in first touch is still doing real work. It's introducing people who convert somewhere else later. Cut that channel because it never gets last touch credit, and the channels that do get credit will eventually run dry, because nothing is feeding them new people anymore.
When first touch and last touch already agree
If you only run one real acquisition channel, first touch and last touch will almost always name the same channel, because there's nothing else in the journey to disagree with it. The debate only starts once a second channel enters the picture, whether that's paid alongside organic, or email alongside social. Worth checking before you build out any comparison: how often is the first touch and the last touch literally the same click? If that number is high, you don't have an attribution problem yet, you have a short, simple journey, and either model will tell you roughly the same thing. Save the deeper comparison work for the channels where the two models genuinely split.
Common mistakes when comparing the two models
- Comparing first touch numbers from one date range against last touch numbers from a different one, which makes the disagreement look bigger or smaller than it actually is.
- Sending traffic through untagged links for part of a campaign, so half the journey has no UTM data and neither model can see it, rather than a genuine model disagreement.
- Treating a spike in direct or branded traffic as organic brand strength, when it's often just last touch attribution defaulting to direct because the referring app or in-app browser stripped the referrer.
- Switching which model a dashboard uses partway through a campaign and then comparing before and after as if performance changed, when only the measurement did.
How to actually see both, instead of picking one
The practical fix is to stop treating this as a single choice. A link tracking tool that records both a first-touch and a last-touch cookie on every click lets you look at the same click history two ways without switching tools. Raydar does this by default on every tracked link, alongside the UTM parameters and referrer that identify the specific post or campaign. Tag links consistently first, with a tool like the UTM builder, so the comparison between first touch and last touch is actually comparing the same campaigns rather than a mess of inconsistent naming.
Once you can see both numbers side by side, the gap between them tells you something real: a wide gap means your channels are specializing, some for introduction and some for closing, which is normal and healthy. A gap that never closes no matter how long you wait might mean the attribution window you're using is too short for how long your customers actually take to decide.
When to move past both
Once you're running enough paid channels that first touch and last touch genuinely disagree in a way that changes what you'd do with a marketing budget, and you have the conversion volume to trust a model built on more than a handful of data points, multi touch attribution starts to earn its complexity. Before that point it mostly adds a spreadsheet without adding a decision. Get comfortable with campaign tracking and the first touch and last touch comparison first. That combination already answers which attribution model to use for the overwhelming majority of businesses reading this.