First touch vs last touch attribution for content-led SaaS

A reader finds your post about Stripe tax settings through Google in March. They read half of it and close the tab. In April they click a link in someone's newsletter, look at your pricing page and leave again. In May they type your domain into the browser, start a trial, and pay two weeks later.

First touch vs last touch attribution is the argument over which of those visits gets the credit. First touch says the Google search and the blog post. Last touch says direct traffic, or the newsletter if your tool skips direct visits. Same customer, same revenue, two different stories about what worked.

Neither model is wrong. They answer different questions, and the useful skill is knowing which question you're asking before you open the report.

What does first touch attribution credit?

First touch gives all the credit for a customer to the first visit you recorded: the landing page, the referrer, and any UTM parameters on that first URL. Everything after it is ignored for attribution.

The question it answers is "what introduced this person to us?"

Two details matter in practice. First, "first" means the first visit you saw, which is not always the first visit that happened. If someone cleared their storage, refused consent, or read your post on a phone and signed up on a laptop, the first touch you have is a later one, or none at all. Second, a first touch never changes once it's stored. A customer who arrived from a blog post in March is still credited to that post in December, which makes past months stable and easy to compare.

What does last touch attribution credit?

Last touch gives all the credit to the final visit before the conversion, usually the signup or the first payment. A common variant, last non-direct touch, skips direct visits and credits the most recent visit that had a source.

The question it answers is "what got them over the line?"

Last touch is the default way most people think about marketing without meaning to. You post on X, signups arrive that afternoon, and you credit the post. That's last touch reasoning, and for short journeys it's usually right.

Why first touch suits content-led SaaS

Software for founders and small teams is rarely bought on the first visit. People read, compare, put it off, come back when the problem gets worse. The gap between discovering you and paying you can be weeks or months.

Under last touch, that gap makes content look useless. The final visit before payment tends to be:

• a direct visit, because they remembered your name,
• a branded search for your product,
• a click in one of your own trial onboarding emails.

All three are effects of earlier marketing. Nobody searches for a brand they've never heard of. If you judge your writing by last touch, the posts that created demand get no credit, and the channels that harvested it get all of it.

First touch puts the credit back where the introduction happened. For a business whose main growth lever is publishing things people find, that's the number that tells you what to write next. It's also cheap to compute: one stored touch per visitor, one row per customer, no need to keep a visit history. We go through that setup in how to see which blog posts bring paying customers.

The cost is real, though. First touch over-credits discovery and ignores whatever did the closing. A comparison page or a pricing page that convinced people who were already aware of you will look weak under first touch, even if it's doing important work.

When is last touch the better model?

Last touch is the better lens when cause and effect are close together in time.

Paid campaigns. You're paying per click and need to know whether the click converts. A retargeting ad has no other job than being the last touch, so judging it by first touch makes no sense.

Launches. A Product Hunt day, a newsletter sponsorship, a pricing change announced to your list. The window is short, and you want to know what the push produced that week.

Short purchase cycles. Cheap consumer apps bought on impulse often convert on the first visit. First and last touch are the same visit, and the debate goes away.

Onboarding and lifecycle email. If you're testing which trial email gets people to pay, last touch is the right frame, because the email's job is to close.

A quick test: look at the time between first visit and first payment for your recent customers. If most of them converted within a day, the models will mostly agree and the choice barely matters. If the typical gap is weeks, the choice changes your conclusions, and you should be deliberate about it.

Are multi-touch attribution models worth it for a small SaaS?

Multi-touch models spread the credit across several visits. Linear splits it evenly. Time decay gives more to recent visits. Position-based, sometimes called U-shaped, usually gives 40% to the first touch, 40% to the last and divides the remaining 20% among the middle. Data-driven models fit the weights from your own conversion history.

At small volume these are mostly false precision. Take a hypothetical month with 15 new customers. Splitting each €29 subscription into 40/20/40 shares produces tidy decimals on every channel, but the weights are a choice you made, not something the data told you. With 15 conversions there's nothing for a data-driven model to learn from either.

There's a second problem. Multi-touch can only spread credit over the touches it saw. The podcast someone heard on a run, the friend who recommended you in a private Slack, the post they read on a work laptop: none of those are in the path. A model that divides credit precisely among the visible touches can look more complete than it is.

And it costs more to run. You need to keep every visit per visitor over a long window, which means a persistent identifier, more storage, and a heavier consent question.

For most bootstrapped products, two simple models you understand beat one elaborate model you have to trust.

How to read first touch and last touch together

If your tools let you see both, compare them per channel. The pattern says more than either number alone.

Before any of this, check how much of your revenue you could attribute at all. If only a fifth of new MRR has a known first touch, a ranking by channel is mostly noise, whichever model produced it. The attributed share belongs next to every attribution report, and it's defined on our definitions page.

Clean campaign tags help both models equally. If the same channel shows up under four different spellings, no model can add it up, so it's worth settling on a UTM naming convention early.

How Joinsight handles it

Joinsight uses first touch only, on purpose. The snippet records the first visit, and a customer's MRR at first payment is credited to that visit's page and channel, with the channel decided by UTM first, then Google Ads click IDs, then the referrer domain, otherwise direct. For paid campaigns where you want last touch, read the ad platform's own conversion reports next to it.