Attribution & Multi-Channel Tracking Basics

The attribution problem

Picture a realistic customer journey: someone sees a social media post about your business, does not click. A week later they see a search ad and click through, browse, and leave without converting. A few days after that, they remember your name and search for it directly on Google, click the organic result, and finally fill out the contact form. Three touchpoints, one conversion — and a genuine question with no single correct answer: which channel deserves credit?

This is the attribution problem, and it exists because analytics tools have to pick some rule to assign credit, even though the real answer is usually "all of them contributed something." Different rules produce meaningfully different pictures of which channels are "working," which is why understanding the rule your reports are using matters as much as the numbers themselves — two people looking at the same underlying data through two different attribution models can walk away with opposite conclusions about which channel to invest in.

This is not a purely theoretical concern for a small business. Even a modest marketing setup — a bit of paid search, some organic content, a social presence — is enough to generate multi-touch journeys constantly. The larger the gap between when someone first hears about you and when they actually convert, the more touchpoints are likely involved, and the more attribution choice matters to how you interpret the results.

First-click, last-click and data-driven attribution, in plain language

• Last-click attribution gives 100% of the credit to whichever channel the customer interacted with immediately before converting. It is simple and was the long-time default in most tools, but it systematically undervalues awareness-building channels — a social post or display ad that started the journey gets zero credit if a search click closed it.
• First-click attribution does the opposite — full credit to whatever channel first introduced the customer to your business. This flatters awareness channels but ignores whatever ultimately closed the sale, which can be equally misleading in the other direction.
• Data-driven attribution (GA4's default model) distributes credit across all the touchpoints in a customer's journey, weighted by how much each one actually appears to have influenced the outcome, based on patterns across your own conversion data. It is more realistic than either single-touch model, though it is also less intuitive to explain and requires enough conversion volume for the underlying patterns to be meaningful — a very low-traffic site may see this model behave more conservatively simply because there is less data for it to learn from.

There is no need to become an expert in the mathematics behind data-driven attribution to use it well. The practical takeaway is simpler: know which model your reports are using before drawing conclusions from them, and treat any single-touch model — first-click or last-click — as a simplification that flatters one part of the journey at the expense of the rest.

Why single-channel thinking misleads

The practical danger of picking the wrong attribution lens is not academic — it changes real budget decisions. A business that judges every channel purely on last-click conversions will tend to overfund search (which naturally captures a lot of "closing" clicks, since people often search for a brand right before converting) and underfund social or display, which more often start journeys than finish them. Cut the "underperforming" awareness channel based on last-click data alone, and search conversions can quietly decline months later — because nothing was introducing new customers into the top of the funnel anymore, and last-click reporting was never going to show that connection until it had already happened.

This does not mean every channel is secretly working regardless of what the numbers show — some genuinely are not. It means the honest answer requires looking at more than one attribution view, and being skeptical of any single number that claims to settle the question of "which channel matters" on its own. A reasonable practical habit is checking a campaign's performance under both last-click and data-driven models before making a significant budget decision about it — if the two models tell a similar story, you can be more confident either way; if they diverge sharply, that divergence itself is useful information about how much of the channel's value is happening earlier in the journey than the last click.

UTM parameters — tagging campaigns properly

Attribution can only credit a channel correctly if GA4 knows where a visit actually came from — and for anything beyond organic search and direct traffic, that requires tagging your links with UTM parameters: small pieces of text appended to a URL that tell GA4 the source, medium, and campaign a click belongs to. For example: ?utm_source=facebook&utm_medium=social&utm_campaign=spring_sale.

Without UTM tags, a link shared on social media, in an email newsletter, or in a WhatsApp message often gets bucketed into GA4's generic "referral" or even "direct" traffic — because there is no campaign information attached to it, and GA4 cannot always infer the true source. This quietly understates the channels doing real work and makes attribution reports far less useful than they should be. A simple, consistent naming convention — always lowercase, always the same medium names across campaigns (social, email, paid) — matters more than the tool used to generate the tags; Google's free Campaign URL Builder is enough for most small businesses.

Consistency matters more than completeness here. A business that tags ninety percent of its campaigns carefully but leaves the medium name spelled differently ("social" versus "Social" versus "social-media") across different links will find those campaigns fragmented into separate rows in every report, understating each one individually even though the underlying traffic was real. Agreeing on a naming convention once, in writing, before tagging anything is worth the ten minutes it takes.

One more source of attribution confusion worth knowing about: self-referrals. If a checkout or a linked tool sends a visitor briefly to a different domain and back without cross-domain measurement configured (Chapter 2), GA4 can register the return trip as a new session with your own site listed as the "referrer" — which shows up in reports as a channel called something like "self-referral," inflating a channel that is not really a channel at all. Seeing your own domain listed as a traffic source is usually a sign that cross-domain tracking needs attention, not a real acquisition channel worth analysing.