Guide

Last touch or first touch: which visit should get the credit

Somebody found you through a podcast in March and bought after a comparison page in May. One of those gets the sale. Choosing which is a decision about your own marketing, not a technical setting.

Updated

What each model actually says

First touch credits the visit that brought somebody to your site for the first time. It answers "what makes people discover us" and over-credits anything at the top of the funnel: a viral post, an aggregator, a mention in a newsletter with a hundred thousand readers.

Last touch credits the most recent visit before the payment. It answers "what makes people decide" and over-credits anything close to the purchase: your own pricing page, a comparison query, a retargeting ad, and above all direct traffic from people typing your domain because they already knew about you.

Both are the same simplification in opposite directions. The journey had several steps and the model keeps one.

The direct-traffic trap, and the fix

Pure last touch has a failure mode that makes it useless for small sites: most people who buy do it on a return visit they started by typing your domain or clicking a bookmark. That visit has no referrer, so it is direct, so direct wins every report and you learn nothing.

The fix is to credit the last visit that was not direct. Somebody who arrived from Reddit, came back directly three weeks later and paid is credited to Reddit. Direct still wins when a visitor has never arrived from anywhere else, which is the honest answer rather than a blank.

Last non-direct touch is the default here. It keeps the question "what made them decide" while refusing to answer it with "they typed your name".

How to choose

If you are deciding where to spend time to be discovered by people who have never heard of you, first touch is closer to your question. If you are deciding what to improve so the people who already know you buy, last touch is.

A practical approach for a small product: run last non-direct, and when a channel looks disappointing, check whether it is bringing people who buy later through something else. You do not need a multi-touch model to see that, you need the visitor's own history, which is one click from any payment.

Switching is per website and takes effect on the next read. Nothing is rewritten: the models differ in which visit is credited at read time, so the same month can be looked at both ways without losing anything.

Why not multi-touch

Multi-touch splits the sale across every visit, by position or by a model fitted to your own data. On a site with thousands of payments a month it can be worth it. On a site with forty it produces confident-looking fractions from nothing, and the fractions move when a single customer buys.

The version that is honest at small scale is the visitor timeline: every visit, every page, every goal and the payment at the end, in order. It does not reduce the journey to a number, and for deciding what to do next it is usually the more useful artefact anyway.

What that looks like

The dashboard this is describing

Live demo, real queries, generated traffic. No sign-up.

NNimbus ‹Last 30 days ›Daily LiveUpdated just now
Visitors
5,221
540% ↑
Revenue
$6,987
710% ↑
Conversion rate
1.02%
18% ↑
Revenue/visitor
$1.34
27% ↑
Bounce rate
61%
0% ↓
Session time
1m 15s
3% ↑
Online
12
View on map →
ChannelReferrerCampaignKeyword
Visitors · Revenue
  • Organic Social2,412$3,120
  • Direct1,072$1,240
  • Referral771$960
  • Organic Search720$580
  • AI246$310
CountryRegionCity
Visitors · Revenue
  • 🇺🇸United States885$2,410
  • 🇮🇳India887$140
  • 🇬🇧United Kingdom427$980
  • 🇩🇪Germany427$760
  • 🇧🇷Brazil451$210

Questions

Does changing the model rewrite old data?

No. Both models read the same visits and the same payments; the choice decides which visit is credited when the numbers are assembled. Change it, reload, and the same months are reported the other way.

What happens when a visitor has only one visit?

The two models agree, which is most first purchases on a small site. They diverge on the people who took their time, and those are usually the customers worth understanding.

How are renewals credited?

To the same visit as the first payment, and marked as renewals so they can be counted separately. A subscription that renews for a year should not keep re-crediting the channel every month as though it brought a new customer, and it should not vanish from that channel's lifetime value either.

Stop guessing which traffic pays.

Add the script, connect the payments you already take, and every channel on the dashboard carries the revenue it produced.

14 days free · no commitment · cancel anytime