Marketing Attribution, Explained in Plain English

Ask three different tools which ad made a sale happen, and there's a good chance all three take credit for it. That isn't a bug in the software. It's baked into how marketing attribution works, and once the mechanics are visible, the disagreement stops feeling confusing and starts being useful.

What is marketing attribution, really?

Attribution is the accounting system behind a sale. It decides which touchpoint, an ad, an email, a search, a referral, gets credit when someone buys. It isn't one fixed number. It's a set of rules, and different rules produce different "true" answers for the exact same sale.

Many purchases now involve more than one interaction. A customer might see an Instagram ad on Monday, search the brand name on Google on Wednesday, get a reminder email on Friday, and finally buy on Saturday after a friend mentions it over lunch. Four touchpoints, one sale. Attribution is simply the method for deciding how much credit each of those four gets.

That word "method" matters. Attribution isn't a measurement in the way a bathroom scale measures weight. It's a rule a business, or a platform, chooses to apply. Change the rule and the credit moves, even though the sale itself never changed. A small service business with a short sales cycle and a big-ticket B2B company with a six-month sales cycle both need attribution, but they'll reasonably pick different rules, because the shape of their customer's journey is different.

The confusion most business owners run into isn't the concept. It's discovering that Google Ads, Meta Ads, and email marketing software each apply a different rule to the same customer journey, and each one reports the sale as theirs.

The four attribution models most businesses run into

Four models cover almost everything a small business will encounter. Last click gives all credit to the final touchpoint before the sale. First click gives it all to the first. Linear splits it evenly across every touchpoint. Data-driven lets a platform's own conversion data decide the split, weighting whichever steps actually correlate with a sale.

Here's what each one actually does, in the order most people meet them:

1. Last click. All the credit, 100%, goes to whichever ad or link the customer clicked right before buying. Google Ads' own documentation defines it exactly this way: last click "gives all credit for the conversion to the last-clicked ad and corresponding keyword" (Google Ads Help, "About attribution models"). It's the simplest model, and for years it was the industry default, which is part of why so many businesses still assume it's the only option.
2. First click. The mirror image of last click: every credit goes to whatever brought the customer in the door originally, ignoring everything that happened afterward. Useful for judging what's good at starting a relationship, useless for judging what's good at closing one.
3. Linear. Every touchpoint on the path gets an equal slice of the credit. Four touchpoints means each one gets a quarter. It's the fairest-looking model on paper and the least useful in practice, because it treats a passing glance at an ad exactly the same as a considered visit to a pricing page.
4. Data-driven. Instead of applying a fixed rule, this model compares the paths of customers who converted against customers who didn't, and works out which touchpoints actually correlate with a sale (Google Ads Help, "About data-driven attribution"). It's specific to each advertiser's own data, which is both its strength (it reflects your actual customers) and its limit (it needs enough conversion volume to find a real pattern).

None of the four is "correct." Each answers a slightly different question: what closed the sale, what started it, what touched it, or what the data says mattered. The mistake is assuming one number is the truth and the other three are wrong.

Why the platforms all take credit for the same sale

Ad platforms report on their own slice of the customer journey, so each one tends to count a sale it merely touched as a sale it caused. Google's own documentation gives a related but different reason for the shift away from simple rules-based credit: those models weren't flexible enough to reflect how people actually buy.

This is the part that catches small business owners out, and it's worth being direct about it: every ad platform is graded on the sales it can claim, so every ad platform is structurally motivated to claim as many as it plausibly can. That isn't a conspiracy. It's what happens when the same company runs the ad, tracks the click, and reports the result, with nobody independent checking the sum.

Google's own history with attribution is the clearest public evidence of the problem. For years, last click was the assumed default across the industry. Then, on 6 April 2023, Google Ads Help announced that first click, linear, time decay, and position-based models were "going away" across Google Ads and Google Analytics 4, replaced by data-driven attribution as the new default, stating plainly that the old rules-based models "don't provide the flexibility needed to adapt to evolving consumer journeys" (Google Ads Help, "First click, linear, time decay, and position-based attribution models are going away," 6 April 2023). By that point, the same announcement noted, less than 3% of Google Ads web conversions were still being attributed using those older models. If the platform that popularised last-click attribution has spent years moving away from simple rules-based credit, that's a strong signal the simple version was never giving anyone the full picture, on any platform, not just Google's.

Add a second platform into the mix, an email tool, a Meta pixel, a marketplace, and the maths gets worse before it gets better: add up every platform's self-reported conversions for a single month and the total can exceed the number of sales actually made, worth checking against your own numbers rather than assuming. Nobody is lying. Each platform is just reporting honestly on its own version of events, and none of them can see what happened on a competitor's ad account. Where attribution fits inside a full marketing engine is the harder, more honest question, one that starts by accepting no single platform's dashboard is the whole story.

What a small business can actually trust

Without a data team, the goal isn't picking the "correct" attribution model, because none exists. The goal is tracking one honest money metric consistently, and treating every platform's self-reported number as a claim to be checked, not a fact to be repeated.

Here's the practical version, stripped of the software vendor sales pitch.

First, stop expecting the numbers to match. Google, Meta, and email will each report a version of "your" sales that adds up to more than actually happened. That's not a red flag on any single platform, it's what self-reported attribution looks like by design.

Second, anchor to one number outside the platforms. The most reliable check on inflated attribution claims is a simple external measure: actual enquiries, actual bookings, actual sales, counted by hand or in a spreadsheet, not pulled from an ad dashboard. The full breakdown of how to pick and track that one number, without needing software, is worth reading in the one number worth watching every week.

Third, ask any agency reporting to you which model they're using, and why. A vague "the platform said so" answer is a warning sign. A specific answer, "we use last click for this channel because the sales cycle is short, and we cross-check monthly against actual bookings," is what accountability looks like. The full standard for what a report should contain goes further than attribution alone.

Nine years running this agency, across 600+ clients and 50+ still active on retainer, the pattern repeats: in our experience, the businesses that grow fastest aren't the ones with the fanciest attribution dashboard. They're the ones who picked one honest number, stopped trusting platform totals at face value, and kept the discipline up every week.

The short version

Attribution isn't a fact a platform reports. It's a rule a platform applies, and every platform has a reason to apply the rule that makes it look best. Understanding the four common models, last click, first click, linear, and data-driven, is enough to read a report critically instead of nodding along. The one thing worth trusting more than any single dashboard is a number tracked outside all of them.

Tell us about your business. Get a straight read on what to trust in your own reporting.