Beyond Last-Click: Practical Multi-Touch Attribution

Multi-touch attribution without a data team takes three things: a UTM convention on every outbound link, a server-side click log you trust, and one model you commit to for a quarter (first-touch, last-non-direct or position-based 40/20/40). Last-click gives all the credit to the final touch and starves the channels that created the demand.

Last-click attribution is the friend who shows up at the end of a long project and takes credit for everything. It's not malicious — it's just how the default analytics view works. And it consistently steers budget toward the wrong channels.

You don't need a data team to fix this. You need three things: a UTM discipline, a click log you trust, and a model. LinkPilot's link click analytics is one source for that click log; the model is what this post is about.

Why does last-click attribution break teams?

Because it hands all of the credit to the final touch, which is usually a branded search or a direct visit, and starves the channels that created the demand. Imagine a customer journey:

1. Sees a YouTube ad → forgets about it.
2. Reads a blog post that ranks for "link tracking" → bookmarks the site.
3. Three days later, types the brand name into Google → converts.

Last-click attributes the sale to branded search. The YouTube ad and the blog post — the actual demand drivers — get zero credit. A few quarters of that and the budget consolidates into bottom-of-funnel channels that literally cannot grow because nothing is feeding them.

A practical multi-touch framework

You need three layers:

1. Capture every touch

Wrap every outbound link in a tracked short link with UTMs. The shortener logs the click server-side; the UTMs let downstream tools (GA4, your CRM) attribute the session, using the parameters Google documents.

2. Stitch sessions to identities

When the user converts (signup, purchase), record the cookie ID + the UTMs of every prior session against the new account. This is the single piece of glue that makes attribution work.

Most CRMs and analytics tools support this; you just have to turn it on.

3. Pick a model and commit

Three models, in order of complexity:

• First-touch. Credit the first known channel. Best for awareness teams.
• Last-non-direct. Like last-click but ignores direct visits. The pragmatic default for most teams.
• Position-based (40/20/40). First and last each get 40%, middle touches share 20%. Best for blended teams.

Pick one. Use it for at least a quarter. Compare the ranking of channels to your last-click ranking. (GA4's own attribution reports offer a data-driven model and last-click models, per Google's attribution guide; the spreadsheet method below is how you run first-touch or position-based without it.) The differences are the channels you've been under- or over-investing in.

How do you run a model without an attribution tool?

With a spreadsheet and the journey above. Take the three touches: YouTube ad, blog post, branded search. One conversion, worth one unit of credit.

• First-touch: YouTube ad 1.0; blog 0; branded search 0.
• Last-non-direct: branded search is a search, not a direct visit, so it keeps the credit: 1.0. If the final touch had been a typed-in URL, the credit would fall back to the blog post.
• Position-based (40/20/40): YouTube ad 0.4; blog post 0.2; branded search 0.4.

Now do that for every conversion in the quarter and sum by channel. The spreadsheet needs three columns per conversion: the ordered list of touches (source and medium from the UTMs on each tracked link), the conversion date, and the value if you have one. The click log provides the touches; the CRM provides the conversion; the model is a formula.

What the data must contain for this to work: a stable identifier that ties sessions together (a cookie ID or a login), UTMs on every outbound link so the source is never "direct" by accident, and server-side click logs so ad blockers do not erase the first touch. If any of the three is missing, start there; the model choice matters less than the gaps.

What does this look like in practice?

A typical spend reallocation after switching from last-click to position- based:

| Channel | Last-click share | Position-based share | | -------------- | ---------------- | -------------------- | | Branded search | 38% | 22% | | Paid social | 12% | 19% | | Content/SEO | 8% | 24% | | Email | 22% | 18% | | Direct | 20% | 17% |

Notice: branded search and direct shrink. Content and paid social grow. That's the demand-generation work finally showing up in the numbers.

When should you upgrade to a dedicated attribution tool?

When paid spend reaches six figures a month, the CRM holds months of clean UTM data, and real decisions are blocked on better attribution. You'll know it's time when:

• You're spending six figures a month on paid media.
• Your CRM has 5+ months of clean UTM data.
• You can name three decisions that are blocked on better attribution.

Before that, the upgrade has negative ROI. Discipline beats software at this stage.

One more test before buying anything: run the spreadsheet model for a quarter and write down the three decisions it changed. If you cannot name them, a tool will produce prettier charts of the same indecision.

References

• Google Analytics Help, Get started with attribution.
• Google Analytics Help, Collect campaign data with custom URLs.