Attribution (Marketing attribution)

Worked example

Maya first finds your product through a Hacker News post. A week later she joins your waitlist from a newsletter mention. On launch day she clicks your announcement email and pays.

• First touch credits Hacker News: it’s how she found you at all.
• Last touch credits the email: it’s what she clicked just before paying.
• Multi-touch splits the credit between all three.

Each answers a different question. None is the single truth.

Which one fits when you’re starting out

First touch answers the question that matters most early on: which channels bring people who end up paying. Your first customers tell you where the next ones are, and first touch points there. Last touch tells you what closes the sale (often your own email), which is useful, but it tends to credit your own follow-up rather than the channel that found the person.

Multi-touch models need many customers and many recorded touches to mean anything; they’re a tool for later.

What makes attribution possible

1. Record the source of the first visit: the referring site and UTM parameters, before the person signs up.
2. Carry it to the signup: store it with their email or your user id.
3. Match payments to people: connect each paying customer to that signup, by email or by your own user id.

Skip the first step and it can’t be recovered later: your billing system knows who paid, not how they found you.

How complete is it?

No method traces everyone: some people pay with a different email, block analytics or switch devices. A good report says how many paying customers it could trace, so you know how much of the picture you’re seeing.

In Nordva Launch

Launch credits the first touch. It traces each Stripe customer to a first visit through the visitor id in the customer’s or subscription’s metadata, identify calls after login, or their waitlist signup, and shows how many it could trace and how. Referring sites and countries work on every plan; UTM groupings need Builder.