Marketing Attribution Explained — Which Channels Actually Drive Sales

Marketing attribution is how you figure out which of your marketing actually produced a sale. Not which ad got the most clicks or which post got the most likes. Which channel put a paying customer in front of you. Most small businesses skip it, then wonder why the budget conversation always turns into a guess. It's the backbone of the marketing analytics work we do, and you can get most of the value without fancy software.

This post covers what attribution is, how the common models hand out credit, why every ad platform seems to take credit for the same sale, and a simple setup any small business can run this month. If you've been trying to bring your cost per lead down, this is the step that tells you which leads were worth paying for in the first place.

What Marketing Attribution Actually Means

Customers rarely buy on the first touch. A typical path looks like this: someone sees your Instagram ad on Monday, searches your name on Google on Thursday, reads two reviews, and calls you the next week from your Google Business Profile. Three or four touchpoints, one customer.

Attribution is the set of rules that decides how the credit for that customer gets split between those touchpoints. Did Instagram earn it because it started the journey? Did Google earn it because it was the last click before the call? Or do they share it?

There's no perfect answer. There's only a consistent one. The goal isn't to know the exact truth about every customer. It's to have a reliable enough picture that you can move money from what isn't working to what is.

The Common Attribution Models, in Plain English

Last click. All the credit goes to the last thing someone clicked before converting. Simple, easy to read, and it's the default way most people think. Its blind spot: it ignores everything that got the customer interested. Brand searches and retargeting look like heroes, and the ad that first introduced you looks useless.

First click. All the credit goes to the first touch. It flatters awareness campaigns and ignores what closed the deal. Useful as a second opinion, not as your main view.

Linear, time decay and position-based. These split credit across touchpoints: evenly, weighted toward the most recent, or weighted toward the first and last. They sound fair, but the splits are arbitrary rules someone picked.

Data-driven. The platform looks at the paths that converted and the ones that didn't, and assigns credit based on what actually seems to make a difference. In Google's words, it "distributes credit for the key event based on data for each key event."

Worth knowing: Google Analytics 4 retired the first click, linear, time decay and position-based models in November 2023. GA4 now offers data-driven attribution plus two last-click options (Google Analytics Help). If you read an older guide telling you to switch GA4 to linear, that setting no longer exists. For most small businesses running paid ads alongside SEO, the default data-driven model is fine. What matters far more is whether the data going into it is clean.

Why Every Platform Claims the Same Sale

Here's the part that trips up almost every owner we talk to. You add up the conversions Meta reports, the conversions Google Ads reports and the conversions your email platform reports, and the total is bigger than the number of customers you actually got.

Nobody is lying, exactly. Each platform only sees its own touchpoints, and each one counts a conversion if its ad was anywhere in the path within its attribution window. Meta sees that the customer viewed an ad and counts it. Google sees the search click and counts it. Your email tool sees the newsletter open and counts it. Same customer, counted three times.

That doesn't make platform numbers useless. They're good for comparing campaigns inside the same platform. They're bad for deciding how to split money between platforms. Retargeting ads are the classic example: they show up right before people buy, so they look incredible in-platform, even when many of those buyers were coming back anyway.

The fix is a single source of truth that sits outside the ad platforms: your own record of leads and sales, with the source of each one attached.

A Simple Attribution Setup for a Small Business

You don't need an enterprise tool. You need four things working together.

1. Tag every link you control. Add UTM parameters to the links in your ads, emails, social bios and QR codes. Google recommends always using three: utm_source (where it came from, like "newsletter"), utm_medium (the type of channel, like "email" or "cpc") and utm_campaign (the promo or offer) (Google's URL builder guide). Pick one naming style and stick to it. "Facebook," "facebook" and "fb" show up as three different sources.

2. Track the conversions that matter. Form submissions, booked calls, phone calls from the site, purchases. Set them up as key events in GA4 and in each ad platform. If a click-to-call button isn't tracked, every phone lead shows up as "nothing happened."

3. Save the source on every lead. Your CRM should store where each lead came from, pulled from the UTM tags or the form, so it follows the person all the way to a sale. If you're still picking a system, our guide to the best CRM for small business covers which ones do this well.

4. Ask people. Add "How did you hear about us?" to your forms and to your intake call. It's low-tech and imperfect, and it catches what tracking misses: word of mouth, a podcast mention, the truck they saw parked outside a job.

How to Use the Numbers Once You Have Them

Once a month, pull one simple report: leads and customers by source, next to what you spent on each. Then ask three questions.

What does a customer cost by channel? Not a click, not a lead. A customer. A channel with cheap leads that never close is more expensive than it looks. This is the same reasoning behind our Google Ads vs SEO breakdown: compare channels on cost per customer, not cost per click.

Where do the self-reported answers and the tracking disagree? If tracking says "direct" but customers keep saying "Instagram," your social is doing more work than the dashboard shows.

What happens if we pause it? The most honest attribution test there is. Turn off a campaign in one area or for a few weeks and watch total leads, not platform-reported leads. If nothing drops, it wasn't driving much. If leads fall off, you've found something worth protecting.

Common Attribution Mistakes

Adding up platform conversions. They overlap. Use your CRM or sales records as the total.

Judging awareness on last click. Video and social ads that introduce people to you will almost always look weak on last click. Look at whether total leads rise when they run.

Optimizing for leads instead of customers. If your ad platforms only ever hear about form fills, they'll find you more form fills, including the junk. Feed sales back where you can. It's one of the biggest levers in lead generation.

Changing the rules every month. Pick a model and a reporting format and keep them. Trends only mean something if the measuring stick stays the same.

Where to Start

This week, do two things: add "How did you hear about us?" to your main form, and tag the links in your next email and your social bios with UTMs. Next month, you'll have your first real view of where customers come from, and you'll be making budget calls on evidence instead of a hunch.

If you'd rather have the tracking, CRM sources and monthly report set up for you, that's what we do. The work we've done shows how we approach it.

Frequently Asked Questions