Marketing attribution for small businesses, explained simply

Ask ten founders what "marketing attribution" means and most will give a vague answer involving dashboards and jargon. The actual idea is simple: a customer rarely buys from the very first ad they see. They might see an Instagram post, search your brand name a week later, click a Google ad, and finally buy after an email reminder. Marketing attribution is just the method for deciding which of those touchpoints gets credit for the sale — and that choice changes which channels look like they're working.

What "attribution" actually means, in plain English

Attribution answers one question: when a sale happens after several marketing touches, how much credit does each one get? It's not a measurement of reality so much as a rule you apply to messy reality, because there's no perfect way to know exactly how much each touchpoint mattered to the customer's decision. Different rules — different attribution models — will hand credit to different channels for the exact same sequence of events.

Why attribution matters for a small business specifically

Budget decisions get made based on which channels "look like" they're working. If your attribution model quietly overcredits one channel and undercredits another, you can end up cutting a channel that was actually contributing, and doubling down on one that was just getting lucky with the last click. For a small business with a tight budget, that misallocation costs more relatively than it does for a company that can afford to run every channel regardless.

The common attribution models, explained without jargon

Last-click attribution

100% of the credit goes to whatever the customer clicked right before buying. Simple to understand, but it systematically favours channels that close a sale — like branded search or retargeting — over channels that introduced the customer earlier, like a blog post or an influencer mention.

First-click attribution

100% of the credit goes to the very first touchpoint that brought the customer into your world. Useful for understanding what generates awareness, but it ignores everything that happened between discovery and the actual purchase.

Linear / multi-touch attribution

Credit is split evenly across every touchpoint in the journey. Fairer in principle than picking one winner, but it treats a passing glance at an ad the same as a deliberate return visit, which isn't quite right either.

Data-driven attribution

This is what GA4 uses by default. Instead of a fixed rule, it looks at patterns across your actual conversion data to estimate how much each touchpoint really contributed, based on what tends to precede a sale versus what doesn't. It's the most realistic of the common models, provided the underlying data feeding it is accurate.

Attribution doesn't tell you the truth. It tells you a story, consistently, according to a rule. The question is whether that rule is honest about how your customers actually buy.

A simple example: one customer, three touchpoints

A customer sees an Instagram ad on Monday, searches your brand name and visits your site on Wednesday without buying, then clicks a Google remarketing ad on Friday and purchases. Under last-click, Google remarketing gets 100% of the credit. Under first-click, Instagram gets 100%. Under data-driven attribution, credit is split across all three based on how similar journeys tend to convert — probably giving Instagram meaningful credit for starting things, even though it never got the final click.

Which model should a small business actually use?

For most small businesses, GA4's default data-driven attribution is a sensible starting point — it's built in, requires no extra setup, and avoids the most obvious distortions of pure last-click or first-click. The bigger lever isn't picking a fancier model; it's making sure every channel's events are tracked correctly in the first place. A sophisticated attribution model applied to incomplete data is still wrong — just wrong in a more convincing way.

The prerequisite nobody mentions: tracking has to work first

Attribution can only distribute credit among the touchpoints it actually sees. If your Meta Pixel is missing events, or your GA4 purchase tracking undercounts, no attribution model — however sophisticated — can correct for data that was never captured. This is why every attribution conversation eventually leads back to the same starting point as a basic ad spend audit: confirm what's actually being measured before debating how credit should be split.