Marketing Attribution: Where Customers Really Come From (And How to Track It Properly)

You pull last month's report and Google Ads says it drove 60% of revenue. Your content person swears the blog post that took off on Facebook is what's been generating the inbound calls. Somebody's wrong - or more likely, both numbers are technically correct and both are misleading you into the wrong decision.

This is what happens when the only attribution model running in your business is last-click, and last-click is the default in almost every ad platform's dashboard because it's the easiest one to build, not the most accurate one.

The blind spot in last-click attribution

Last-click gives 100% of the credit to whatever channel the customer clicked right before they converted. It doesn't ask what got them interested three weeks ago - only what happened in the final five seconds before checkout.

Last-click attribution doesn't measure what worked - it measures what happened right before the credit card came out.

Here's the sequence that plays out constantly and never gets counted properly: someone sees your product on Instagram, forgets about it, googles your brand name two weeks later out of curiosity, clicks your search ad, and buys. Google Ads logs that as a search-driven sale. It wasn't. The ad just happened to be standing at the finish line.

• Search ads and retargeting get systematically overvalued - they intercept people who already decided, then take credit for the decision.
• Branded search volume going up is usually a symptom of brand awareness working, not proof that search itself is doing the heavy lifting.
• Organic content, SEO, and community posts get undervalued because their effect shows up weeks later, in someone else's attribution window.
• Word of mouth and referrals are nearly invisible unless you deliberately ask - no platform tracks a conversation that happened over dinner.

If you're only looking at platform-reported ROAS, you're not learning how to track attribution properly - you're reading a report that was designed to make its own channel look good, because every ad platform attributes conversions to itself by default.

What a small business can actually afford to do

You don't need a multi-touch attribution platform charging a few thousand dollars a month. Most businesses under a certain scale don't have enough transaction volume to make a statistical model worthwhile anyway - you'd be paying for precision you can't use.

What actually closes 80% of the gap, at close to zero cost: put a unique UTM parameter on every single campaign link so you know exactly which post, which ad, which email drove the click. Then add one manual field at the point of inquiry or purchase - a simple "how did you hear about us?" question, logged into your CRM or even a spreadsheet. Cross-reference the two and you've got a rough but honest picture of the real customer journey, built without buying anything.

We run this exact setup at our own clinic brand, itsherbs.com, across 10 locations and 40+ staff - UTM tagging plus a front-desk question, all feeding into a shared dashboard. It's not elegant, but it's accurate enough to make a budget call every month, which is the only thing attribution needs to do.

How to tell if someone actually knows how to track attribution properly

Here's the test: does the person reporting your numbers ever volunteer which channel is getting more credit than it deserves? Anyone can show you a ROAS chart that only goes up. The ones who actually understand attribution will tell you, unprompted, "your search numbers look strong, but I think a chunk of that is really your content working upstream - I wouldn't cut the content budget to fund more search."

Before you hire an agency or freelancer for this, ask them directly: how do you handle multi-touch customer journeys, do you track assisted conversions and not just last-click, and what percentage of your reported conversions are actually branded or direct traffic wearing a search-ad costume. If the answer is vague or defensive, they haven't done the analysis - they've just built the dashboard.

The goal was never a perfect number

Attribution modeling can turn into an endless academic exercise if you let it. That's not the point. The point is having enough signal, today, to answer one question: does this channel deserve more budget next month, or less. A rough, honest, 80%-accurate answer you can act on this week beats a mathematically perfect model that takes three months to build and arrives after the budget's already spent.

The businesses that get this right aren't running fancier software - they're the ones willing to say out loud which channel is overrated, and move the money anyway.