Attribution models in marketing: which one should a small business use?

Attribution models in marketing explained simply: last click, first click, linear, data-driven and more, why Google and Meta disagree, and which to use.

You spend money on Google Ads, Meta ads and maybe SEO. At the end of the month, Google says it brought you 20 customers, Meta says 15, and your bank account says you got 22. Every platform is taking credit for the same people. An attribution model is the rule that decides who gets the credit, and the rule you pick changes which marketing looks like it works.

This guide explains the main attribution models in plain English, which ones you can still use in Google Ads, GA4 and Meta, why the numbers never match, and a simple setup that tells a small business where its customers really come from.

Quick answerWhich attribution model to use
  1. Know the modelsLast click, first click, linear, time decay, position-based and data-driven split the credit differently.
  2. Use what the tools offerGoogle Ads and GA4 now offer data-driven and last click. Meta uses attribution windows instead.
  3. Don't trust one platformEach ad platform counts the sales it touched, so the totals add up to more than you sold.
  4. Track the source on every leadSave where each lead came from in your CRM, ask "how did you hear about us?", and compare with real sales every month.

What is marketing attribution?

Marketing attribution is working out which marketing led to a lead or a sale. An attribution model is the rule that decides how to share the credit when a customer saw more than one thing before buying.

That happens all the time. A typical customer journey for a service business looks something like this:

Example: one customer, four touchpoints
Day 1Instagram ad. Sees your ad while scrolling and visits your website.
Day 6Google search. Searches for the service, finds your article and reads it.
Day 12Email. Opens your newsletter and clicks through to your prices.
Day 14Google ad. Searches your business name, clicks your ad and books a call.
Four touchpoints, one customer. Each attribution model gives the credit to a different one.

So who brought in this customer? The Instagram ad that introduced you? The Google ad they clicked last? Every model answers that question differently.

The main attribution models, explained

Here is how each model would split the credit for the customer above:

ModelHow it splits the creditWho gets the credit in the example
Last click100% to the last thing clicked before convertingThe Google ad
First click100% to the first thing that brought them inThe Instagram ad
LinearEqual shares to every touchpoint25% each to the Instagram ad, the article, the email and the Google ad
Time decayMore credit the closer a touchpoint is to the saleMost to the Google ad, least to the Instagram ad
Position-based (U-shaped)Most credit to the first and last touch, the rest shared in the middleMostly the Instagram ad and the Google ad
Data-drivenUses your account’s data to estimate how much each touchpoint raised the chance of convertingDepends on your data

Last click

The simplest model, and for years the default almost everywhere. It is easy to understand, but it favours whatever happens right before the sale: branded searches, retargeting ads and email. In the example, the Google ad gets all the credit even though the customer was already searching for your name.

First click

The opposite. It shows which channels find new people, which makes it useful for judging awareness ads and content. It ignores everything that helped close the sale.

Linear

Every touchpoint gets an equal share. It is fair in a simple way, but it treats a quick glance at an email the same as the ad that introduced you.

Time decay

Touchpoints closer to the sale get more credit. It suits longer buying decisions where the last few steps really do matter more, but it still undervalues the first touch.

Position-based

Also called U-shaped. The first and last touch get most of the credit and the middle steps share the rest. It is a reasonable middle ground for businesses that care both about finding new people and closing them.

Data-driven

Instead of a fixed rule, the platform compares the paths of people who converted with those who didn’t and estimates how much each touchpoint added. It is the most sophisticated option, but you can’t see exactly how it decides, and it only knows about the touchpoints the platform can track.

Which models you can actually use today

Most of the models above are now only theory in the big ad tools.

Google Ads. Google removed first click, linear, time decay and position-based models in 2023. According to Google’s announcement, conversion actions still using them were switched to data-driven attribution in September 2023. Today you choose between data-driven (the default) and last click.

Google Analytics 4. GA4 lost the same four models. In its attribution settings you can choose data-driven, paid and organic last click, or Google paid channels last click. The last-click options ignore direct visits, so someone typing your address in after clicking an ad still credits the ad.

Meta. Meta doesn’t offer these models at all. Instead it uses attribution settings: how long after someone clicked or saw an ad a conversion still counts. As Jon Loomer explains, the default is a 7-day click, 1-day engage-through and 1-day view-through window. Since March 2026, click-through only counts clicks on the ad’s link. Likes, comments and other interactions now fall under engage-through.

That last point matters. Meta can count a sale from someone who saw your ad and never clicked it. Google Ads and GA4 would not give Meta credit for that same sale. This is one reason the Google Ads vs Facebook Ads question is hard to answer from the platforms’ own reports.

Why Google, Meta and GA4 never agree

If you have ever added up your platform reports and got more sales than you actually had, this is why:

ReasonWhat happens
Each platform only sees itselfGoogle Ads doesn’t know the customer saw your Instagram ad. Meta doesn’t know they clicked a Google ad. Each one counts the sale.
Views vs clicksMeta can count people who only saw an ad. Google Ads and GA4 mostly count clicks.
Different time windowsA sale 20 days after a click counts in one tool and falls outside the window in another.
Different modelsData-driven in one place, last click in another, so the same journey is split differently.
Tracking gapsAd blockers, cookie consent and phone calls mean some conversions are never recorded at all.
Offline salesIf the sale happens on the phone or in person, the platforms only know about it if you send it back to them.

None of the reports is wrong, exactly. They are answering different questions. But it means you can’t add the platform numbers together, and you can’t let each platform judge itself.

Which attribution model should a small business use?

The honest answer: the model matters less than having one place where every lead’s source is recorded. Here is the setup we recommend for service businesses:

1. Pick one source of truth: your CRM

Every lead should land in your CRM with where it came from. Add UTM tags to every ad, email and social link, and save the UTMs and the click ID (the gclid from Google, the fbclid from Meta) in hidden form fields. Save both the first touch (how they first found you) and the last touch (what they clicked before enquiring). Our free UTM builder makes the tagging part easy.

When the lead becomes a customer, you can see which channels brought in paying customers, not just form fills.

2. Ask “how did you hear about us?”

Add the question to your forms and booking calls. Our guide to how did you hear about us has the answer options to use. Some of your best channels leave no click to track: word of mouth, a podcast, a ChatGPT recommendation, a post someone shared in a group chat. People will tell you about those when no tracking tool can. Compare their answers with the tracked source. When they disagree, you learn something.

3. Use the simple models where they help

  • In GA4, paid and organic last click is easy to explain and good for day-to-day reports. Look at first-touch data (the first user source) when you want to see which channels find new people.
  • In Google Ads, data-driven is the default and is usually fine once the account has steady conversions. Last click is easier to explain if you are just starting.
  • In Meta, keep the default settings, but look at click-through results separately. View-through numbers are the most likely to include sales that would have happened anyway.

4. Send real sales back to the ad platforms

The platforms can only optimize for what they can see. If leads become customers on the phone or in your CRM, send those stages back: Meta’s Conversions API for Meta and offline conversion imports for Google Ads. Then the ads chase buyers, not form fillers.

5. Check against the bank account every month

Compare the total sales the platforms claim with the sales you actually made. Then work out your real customer acquisition cost per channel from your CRM, not from the ad platforms. If you want to know whether a channel really adds sales, pause it in one area or for a few weeks and see what happens to your leads. That kind of test shows what attribution models can only estimate.

Common attribution mistakes

  • Adding up platform reports. Every platform counts the shared customers, so the total is always too high.
  • Letting each ad platform grade itself. Meta and Google are not neutral judges of their own ads.
  • Cutting the ads that find new people. Under last click, awareness ads look useless because someone else gets the final click. Cut them and the branded searches often drop a few weeks later.
  • Giving branded search all the credit. People searching your business name usually found you somewhere else first.
  • Not tracking phone calls. For many local businesses, calls are most of the leads. Without call tracking, the channels that drive calls look weak.
  • Changing the model to make results look better. Pick a setup, keep it and compare like with like over time.

Attribution is the tracking step in our 7-step customer acquisition system. It is what lets you put money behind the channels that bring in customers and cut the ones that don’t. If AI tools are starting to send you visitors too, see how to track AI traffic in GA4.

Questions about attribution models

What is the best attribution model?

There isn’t one best model for everyone. For most small businesses, the best setup is the lead source saved in your CRM (first and last touch), a “how did you hear about us?” question, and the platforms’ own data-driven or last-click models used as a second opinion. That beats picking the perfect model in any single tool.

What is the difference between first touch and last touch attribution?

First touch gives all the credit to the first channel that brought someone to you. Last touch gives it all to the last thing they clicked before converting. First touch shows which channels find new people. Last touch shows which channels close.

What is data-driven attribution?

It is a model where Google Ads or GA4 uses your account’s data to estimate how much each touchpoint added to the chance of a conversion, instead of using a fixed rule. It is the default in Google Ads and GA4. It only sees touchpoints those tools can track, so it can’t credit things like word of mouth.

Why do Meta and Google both take credit for the same sale?

Each platform only sees its own ads. If someone saw your Meta ad and later clicked your Google ad, both platforms count the sale. Meta also counts some people who only saw the ad without clicking. That is why the platform totals add up to more than you actually sold.

What attribution model does GA4 use?

GA4 uses data-driven attribution by default for its conversion reports. You can change it in Admin under Attribution settings to paid and organic last click or Google paid channels last click.

Do I need attribution software?

Usually not at first. Most small businesses get further with UTM tags on every link, the source saved on every lead in the CRM and call tracking. Paid attribution tools start to make sense when you spend across many channels and have a long sales process with lots of touchpoints.

Want this set up properly, from the first click to the closed sale? See our tracking and analytics service or book a call.

Your Revenue GrowthStarts Here.

Growth questionnaire

Tell us where you want to go.

Five short steps, about 3 minutes. We read every answer and come back with what we would fix first.

Step 1 of 5

What kind of business do you run? *