For most service businesses, GA4's data-driven model is the best default, but no GA4 model can tell you which ad closed a job on its own. GA4 credits website conversions, not booked work. The model that holds up compares data-driven and last click attribution, then checks both against leads counted from your own forms and jobs marked won.
Here is the problem in one sentence: you can open GA4 on Monday, change one setting, and watch your Google Ads "results" shift without a single new customer calling. Same traffic, same leads, same jobs, different story. If you run a plumbing, roofing, HVAC, cleaning or restoration business, that should worry you, because budget decisions get made on whichever story the report happens to tell. This guide explains the attribution models Google Analytics gives you, what each one is good and bad at for jobs that take days to close, and how to stop guessing which one tells the truth.
TruLata is an AI marketing platform for service businesses that puts search, ads, content, outbound and email on one screen, refreshes the data when you open it, and counts leads from the business's own forms rather than from ad platform reports. That last point matters a great deal for attribution, and we will come back to it.
What is an attribution model in Google Analytics 4?
An attribution model is the rule GA4 uses to decide which marketing touchpoint gets credit for a conversion. A customer might find you through an organic search, come back two days later from a Google Ads click, and finally fill in your quote form after opening a reminder email. Three touchpoints, one conversion. The attribution model decides who gets the credit.
As MarTech's guide to attribution models in GA4 puts it, attribution models aim to answer which channels drive results, and the catch is that different models can give you different answers even when working with the same data. That is the whole issue in a nutshell. The model is not a measurement. It is an opinion about how to divide credit, applied consistently.
Why this matters more for service businesses than for online stores
An online store sees the click and the sale on the same website, often in the same session. A service business does not. The real sequence usually looks like this:
- A homeowner clicks an ad or a search result and reads your page.
- They leave, compare two competitors, and come back later.
- They submit a form or call.
- Your team quotes, follows up, and maybe books the job days later.
- The job is completed and paid, sometimes weeks after the first click.
GA4 sees the first three steps at best. It never sees the quote, the follow up, or the payment. So even a perfect attribution model in GA4 is answering a smaller question than the one you care about: which marketing produced form fills, not which marketing produced paid jobs.
Which attribution models does GA4 actually offer?
GA4 attribution is simpler than it used to be. Older versions of Google Analytics offered a long menu of rule-based models. GA4 currently offers three models you can set as your reporting default:
1. Data-driven attribution
Data-driven attribution uses machine learning on your account's conversion paths to assign partial credit to each touchpoint based on how much it appears to contribute. Google recommends it as the default. Instead of a fixed rule, it compares paths that converted with paths that did not and weighs channels accordingly.
Good for: businesses with enough conversion volume for the model to learn from, and for seeing the supporting role that organic search, social and email play before the final click.
Weak for: low-volume accounts. If you get a modest number of form fills a month, the model has very little to learn from, and its splits can be hard to explain or verify.
2. Paid and organic last click
This model gives 100% of the credit to the last channel the person clicked through from before converting, whether that channel was paid or organic. This is classic last click attribution.
Good for: clarity. You can explain it to anyone in ten seconds, and it is easy to sanity check against what customers tell you on the phone.
Weak for: anything that happened earlier. The ad that first introduced you to the homeowner gets nothing if they came back later through a branded search. For a deeper look at that blind spot, see our piece on why the final click doesn't tell you what actually closed the job.
3. Google paid channels last click
This model gives 100% of the credit to the last Google Ads click before the conversion. If there was no Google Ads touchpoint at all, it falls back to paid and organic last click. It exists mainly so GA4 numbers line up more closely with how Google Ads itself tends to report.
Good for: reconciling GA4 with your Google Ads account.
Weak for: honest budget decisions. By design, it leans toward crediting Google Ads whenever an ad click appears anywhere near the end of the path.
One rule that applies to all three: direct visits are excluded
GA4 models do not give conversion credit to direct visits (someone typing your URL or using a bookmark) when another channel exists on the path. In practice, a returning customer who first found you via an ad and later typed your address will still have that conversion credited to the ad. That is usually reasonable, but it is worth knowing when your direct traffic looks large and gets no credit.
What happened to first click attribution in GA4?
First click attribution, which gives all credit to the first channel that brought a person to your site, is no longer one of the reporting models you can select as your property default. But the first click view did not vanish entirely. GA4's user acquisition report uses "first user" dimensions (such as first user source and first user medium), which attribute the user to the channel that first acquired them, regardless of what they did afterward.
So if you want to know which channel introduces new customers to your business, look at the user acquisition report. If you want to know which channel tends to finish the job of getting a form fill, look at last click. Those are different questions, and both are worth asking.
Which GA4 model actually works for service jobs?
Honest answer: use data-driven as your default, read last click beside it, and treat neither as the final word. Here is how to think about it by business situation.
If you get a lot of form fills each month
Data-driven attribution has enough paths to learn from. Use it as your default and expect it to give some credit to the search, social and email touches that happen early. When data-driven and last click mostly agree, you can trust the direction of the numbers.
If you get a modest number of form fills each month
Most local service businesses live here. Keep data-driven as the default if you like, but lean on last click attribution and first user source for decisions, because they are explainable. If you cannot explain why a channel got credit, you should not move budget because of it.
If most of your customers call instead of filling in a form
No GA4 model fixes this on its own. If calls are not tracked as conversions, every model is dividing up an incomplete picture. Get calls tracked first, then worry about the model.
If your jobs take days or weeks to close
This is where GA4 attribution reaches its limit. The model chooses which channel gets credit for the lead. It cannot tell you which lead became a paid job. For that you need lead outcomes recorded somewhere, which is the subject of our guide to counting which ads brought in real jobs.
How do you compare attribution models in GA4 without getting lost?
GA4 lets you change the reporting attribution model at the property level, and the change applies to your current and historical reports. That means you can test how much your story depends on the model before you commit to it. A practical routine:
- Step 1: Confirm your key events are real leads. A key event should be a form submission or a tracked call, not a page view of your contact page. Garbage conversions make every model wrong.
- Step 2: Pick one month and note the credit per channel under data-driven attribution.
- Step 3: Switch to paid and organic last click and note the same numbers. Then switch back.
- Step 4: Look at the gap. If paid search gets far more credit under data-driven than under last click, ads may be doing early introduction work. If it gets far less, ads may be mostly catching people who were already coming back.
- Step 5: Check first user source in the user acquisition report to see who is introducing new people.
- Step 6: Compare GA4's lead count with your actual lead count from your form inbox or CRM. If they disagree, fix tracking before trusting any model.
Do this once a quarter, or any time you are about to make a significant budget change. It takes less than an hour, and it tells you whether your decision depends on a model setting.
Why do GA4, Google Ads and your lead inbox never agree?
Because they are counting different things with different rules. Google Ads typically counts conversions using its own attribution, its own conversion windows and its own definition of a conversion. GA4 counts using whichever model you selected. Your form inbox counts actual submissions, including people who blocked cookies, switched devices, or were never tagged.
None of those numbers is a lie, but only one of them is the actual count of people who asked you for work: the one from your own forms. That is why TruLata uses lead tracking counted from your own forms as the base number, then lets your team mark the outcome of each lead. Platform conversion counts become a reference, not the source of truth.
Common mistakes that make every model wrong
- Counting duplicate form fills. A thank-you page refresh can fire a second conversion.
- Counting spam. Bot submissions look like conversions to every model.
- Not tracking calls. For many trades, calls are most of the leads.
- Mixing up micro events and leads. A button click is not a lead.
- Reading one model and acting immediately. Always check a second view before cutting a channel.
What should you actually decide with GA4 attribution?
Use GA4 attribution for directional questions, and use your job records for money questions.
GA4 attribution is good at answering:
- Which channels introduce new visitors who later become leads?
- Does organic search support my ads, or compete with them?
- Did a landing page change shift which channel finishes the conversion?
GA4 attribution cannot answer:
- Which campaign produced the jobs that paid the most?
- Which leads were spam, out of area, or price shoppers?
- Whether a lead from an ad closed or went to a competitor.
For the second list, you need the lead's source tied to its outcome. That is a record your team keeps, not a setting in GA4. When those outcomes live beside your channel data, the attribution debate gets much shorter: you can see which sources produce won jobs, whatever any model says about the click.
How does TruLata handle attribution for service businesses?
TruLata reads GA4 so you do not have to open it, pulls ad spend and cost per lead from your ad accounts, and counts leads server side from your own forms. Each lead carries its source, and your team sets its outcome. Everything sits on your TruLata dashboard, stamped with when it was pulled, so you know how fresh the numbers are before you act on them.
The result is a simple comparison you can make every week: what GA4 credits, what the ad platform claims, and what your own forms and job outcomes show. When all three agree, spend with confidence. When they disagree, you know exactly where to look. If you want a broader framework for jobs with long close times, our guide to the marketing attribution model for service businesses goes further.
Ready to see which ads actually close jobs?
Stop flipping attribution settings and hoping one of them is right. Open the live demo of TruLata and see your GA4 data, ad spend and form-counted leads on one screen, refreshed when you open it. Sign up at login.trulata.com/signup, connect your accounts, and get an honest answer to which marketing brings in the work.
