To use Google Analytics for marketing, define your key events first, then tag every campaign with consistent UTM parameters and work from four reports: Traffic acquisition, Landing page, Retention, and Explorations. GA4 rewards a measurement plan and punishes ad-hoc report browsing, so decide what decision each number supports before you open anything.
This is a strategy guide, not an install guide. If Analytics isn’t running yet, start by adding Google Analytics to WordPress, and if you plan to manage several tags, put Google Tag Manager on WordPress underneath it. One thing to accept up front: Universal Analytics is gone, sessions are no longer the center of the model, and bounce rate has been replaced by engagement rate. Skills that transferred from UA will get you about 60% of the way.
Set up measurement that maps to business goals
The most common GA4 failure is not a broken tag. It is a property collecting hundreds of thousands of page views with zero key events defined, so nobody can tell a good month from a bad one.
Before you open a report, write down the three to six actions that mean money. For a SaaS product that might be trial started, demo requested, and paid signup. For a local service business it is form submitted, phone click, and directions click. For a publisher it is newsletter signup and affiliate outbound click.
Then mark each one as a key event under Admin > Data display > Events. Standard properties allow up to 30 key events, which is far more than you need. Be ruthless: if everything is a conversion, nothing is. Our walkthrough on how to track a custom event in Google Analytics 4 covers the implementation side, including the custom-dimension step that most people forget.
The reports a marketer actually lives in
Acquisition: user acquisition vs traffic acquisition
These two reports look almost identical and answer completely different questions. The difference is dimension scope.
User acquisition uses user-scoped dimensions such as First user source / medium. It answers “which channel introduced this person to us?” That is the report for judging top-of-funnel investment, brand campaigns, and content that wins new audiences.
Traffic acquisition uses session-scoped dimensions such as Session source / medium. It answers “which channel started this visit?” That is the report for judging retargeting, email, and anything designed to bring known people back.
A retargeting campaign will look terrible in User acquisition (it rarely introduces anybody) and strong in Traffic acquisition. Comparing the wrong one is how good campaigns get cut.
Engagement and Landing page
Under Reports > Engagement, the Pages and screens report tells you what gets read. The Landing page report is the more useful one for marketers because it attributes sessions and key events to the entry point, which is what your ads and search results actually point at. Sort by key events rather than sessions and you’ll usually find one or two pages carrying the whole funnel.
Retention and cohorts
The Retention report answers whether new users come back. For subscription and content businesses this is the number that predicts next quarter. For a one-off purchase business it matters less. Build a cohort exploration when you want to compare, say, users acquired from organic search in March against paid social in March over the following eight weeks.
Explorations: free-form, funnel, path
Explorations are where GA4 stops being a dashboard and starts being an analysis tool. Three types earn their keep:
- Free-form — build any table or chart from any dimension and metric combination. This is where your custom dimensions finally become visible.
- Funnel exploration — define ordered steps and see drop-off between them. Tick Make open funnel if users can enter mid-way. This is the single best tool for finding where checkout or signup leaks.
- Path exploration — see what users did before or after a specific event, without pre-defining the steps. Good for discovering unintended journeys.
| Marketing question | The GA4 report that answers it |
|---|---|
| Which channel brings us new people most cheaply? | Reports > Acquisition > User acquisition, dimension First user source / medium |
| Which campaign drove last week’s conversions? | Reports > Acquisition > Traffic acquisition, dimension Session campaign |
| Which entry page should we send ad traffic to? | Reports > Engagement > Landing page, sorted by key events |
| Where exactly are people abandoning checkout? | Explore > Funnel exploration with your checkout events as steps |
| Do the users we acquired in Q1 still come back? | Reports > Retention, plus a Cohort exploration for channel splits |
UTM tagging, done properly
Channel reporting is only as good as your tagging. GA4 recognizes nine campaign parameters, but two of them are not reported in the interface yet.
| Parameter | What it holds | Use it? |
|---|---|---|
| utm_source | Where the click came from: newsletter, tiktok, partner-name | Always |
| utm_medium | The mechanism: email, cpc, paid_social, referral | Always |
| utm_campaign | The campaign name: spring_sale_2026 | Always |
| utm_id | A stable campaign ID, used to join cost data | Strongly recommended |
| utm_term | Keyword, for manually tagged paid search | Paid search only |
| utm_content | Creative or placement variant, for A/B splits | When testing creative |
| utm_source_platform | The buying platform, e.g. Google Ads, DV360 | Recommended |
| utm_creative_format | Format such as video or display | Not currently reported in GA4 |
| utm_marketing_tactic | Targeting tactic such as remarketing | Not currently reported in GA4 |
A tagged URL looks like this:
https://example.com/pricing/?utm_source=newsletter&utm_medium=email&utm_campaign=q3_pricing_push_2026&utm_id=q3pp26&utm_content=hero_buttonNow the rules. Write them down somewhere your whole team can see, because GA4 treats Email and email as two different mediums.
- Lowercase everything. No exceptions.
- Underscores inside a value, never spaces.
spring_sale, notSpring%20Sale. - Keep
utm_mediumto a short closed list you agree on once:cpc,paid_social,email,affiliate,referral,display. - Put the date or quarter in the campaign name so you can compare year over year.
- Never UTM-tag internal links on your own site. It restarts the session and wipes the original source.
- Do not manually tag Google Ads URLs. Leave auto-tagging on and let the
gcliddo the work.
Default channel groups and why “Unassigned” appears
GA4 sorts traffic into rule-based default channel groups: Organic Search, Paid Search, Organic Social, Paid Social, Email, Referral, Direct, Display, Affiliates, SMS, and more. Google added an AI Assistants channel, which pulls ChatGPT, Gemini, Copilot and similar referrers out of the generic Referral bucket. If you publish content, that channel is worth watching now.
Unassigned is the catch-all GA4 uses when no channel rule matches the event data. In practice it usually means one of four things: a utm_medium value that doesn’t match any rule (utm_medium=newsletter is a classic — use email), a source with no medium at all, consent-mode-modeled traffic, or a hit that arrived without campaign data because of a broken redirect. Fix the tagging and Unassigned shrinks. Google’s default channel groups reference lists the exact matching rules.
Attribution: what GA4 will and won’t tell you
GA4 now offers only three attribution models, chosen under Admin > Data display > Events > Key events attribution:
- Data-driven — a machine-learned model that distributes credit across the path using both converting and non-converting journeys. This is the recommended default.
- Paid and organic last click — all credit to the last non-direct click.
- Google paid channels last click — all credit to the last Google Ads click.
Comparisons, segments and audiences
Three tools that sound similar and are not. Comparisons sit at the top of standard reports and let you overlay up to four filtered views side by side, for example mobile vs desktop. They are temporary and report-scoped. Segments live inside Explorations and can be user-, session- or event-scoped, including sequence conditions. Audiences are persistent lists built under Admin > Data display > Audiences, and they are the only one of the three you can push to Google Ads.
Audiences are forward-looking: they start collecting members when you create them, though GA4 backfills up to 30 days of qualifying users for Google Ads export. Maximum membership duration is 540 days. Create the audiences you think you’ll want long before the campaign, because you cannot retroactively populate them.
Link Google Ads and Search Console
Both links live under Admin > Product links, and both need Editor on the GA4 property.
The Google Ads link gives you three things: import key events as Ads conversions, share audiences for remarketing, and see Ads clicks, impressions and cost inside GA4 reports (allow about 48 hours). You also need admin access on the Ads side.
The Search Console link adds Google Organic Search Queries and Google Organic Search Traffic reports. Here’s the gotcha nobody mentions: the report collection is unpublished by default. After linking, go to Library at the bottom of the left navigation and publish the Search Console collection, or you will link it successfully and see nothing. Data appears within 48 hours and covers a maximum of 16 months.
Build one Looker Studio dashboard
Looker Studio is free and connects to GA4 natively. The point is not a beautiful dashboard; it is removing the reason to log into GA4 for routine questions.
- Create a blank report and add the Google Analytics connector, then pick your property.
- Add a date range control and a comparison to the previous period.
- Scorecards across the top: Sessions, Engaged sessions, Key events, Total revenue.
- A time series of key events by week.
- A table of Session default channel group with Sessions, Key events and key event rate.
- A table of Landing page with Sessions and Key events, filtered to your top 20.
- Schedule an email delivery every Monday morning.
The GA4 limits marketers should know about
Four quirks will make your numbers look wrong if you don’t know they exist.
Data thresholding. GA4 withholds rows when a report includes demographics, interests or search-query data and there aren’t enough users to keep individuals unidentifiable. Thresholds are system-defined and cannot be adjusted. Widening the date range usually brings the rows back.
Cardinality. A dimension with more than 500 unique values in a single day is high-cardinality. Once a report exceeds its row limit, everything past it collapses into a single (other) row. Page path and item ID are the usual culprits. Don’t register free-text parameters like search terms or full URLs as custom dimensions unless you truly need them.
Sampling in Explorations. Standard properties sample above 10 million events for a query; Analytics 360 goes up to 1 billion. Check the data quality icon at the top of the exploration for the percentage used. A sampled funnel with 40% of data is directional, not exact.
Data retention. This one costs people a year of history. The default for event data on standard properties is 2 months, not 14. Go to Admin > Data collection and modification > Data retention and change it to 14 months on day one. It only affects Explorations and funnel reports, not the standard aggregated reports, and it is not retroactive.
Troubleshooting the numbers that look wrong
Too much traffic showing as Direct
Direct is GA4’s shrug. It appears when no referrer arrives: clicks from PDFs, desktop email clients, in-app browsers, QR codes, HTTPS-to-HTTP hops, and untagged links. Some direct traffic is real. A sudden jump in it almost always means a campaign shipped untagged or a redirect is stripping parameters.
Self-referrals in the Referral channel
If your own domain appears as a referral source, a cross-domain journey is breaking the session, typically a third-party checkout, booking tool, or payment page. Add every domain in the journey to the unwanted referrals list and configure cross-domain measurement, both under Admin > Data collection and modification > Data streams > Configure tag settings.
Spam and bot traffic
GA4 filters known bots and spiders automatically and you cannot turn it off. What still gets through is scripted traffic hitting your real pages. Use internal traffic filters for your own offices, and if you see a burst from one hostname that isn’t yours, add a Hostname filter in your reports rather than trying to block it at collection.
GA4 disagrees with Google Ads, Meta or TikTok
Expect this, permanently. The platforms count a conversion against the day of the click or impression; GA4 counts it on the day the conversion happened. Ad platforms credit view-through and in-app conversions that a website tag never sees. GA4 uses data-driven or last-click attribution while each ad platform uses its own model, and every platform double-counts a user who converted after clicking two different ads. Differences of 10 to 30% are normal. Use GA4 to compare channels against each other on a consistent basis, and each ad platform to optimize inside itself. If you want a closer match on paid social, server-side tracking helps, which is why setting up the TikTok Pixel alongside the Events API narrows the gap.
Frequently asked questions
What is the difference between user acquisition and traffic acquisition in GA4?
User acquisition uses user-scoped dimensions and shows the channel that first brought a person to your site. Traffic acquisition uses session-scoped dimensions and shows the channel that started each individual visit. Use the first to judge prospecting, the second to judge retargeting, email and returning-visitor campaigns.
Why does GA4 show fewer conversions than my ad platform?
Because they count differently. Ad platforms attribute conversions to the click date and include view-through conversions; GA4 records them on the conversion date and only sees what its own tag observes. Consent banners, ad blockers and Safari’s cookie limits widen the gap further. A 10 to 30% difference is expected, not a bug.
How long does Google Analytics keep my data?
Standard properties default to 2 months of event-level data, with 14 months as the maximum. Analytics 360 adds 26, 38 and 50-month options. Change it under Admin > Data collection and modification > Data retention. Standard aggregated reports are unaffected, but Explorations and funnels are limited by this setting.
Is Google Analytics enough on its own for marketing measurement?
For most small and mid-sized businesses, yes, provided key events are defined and campaigns are tagged. What it does not do well is offline conversions, long B2B sales cycles that end in a CRM, and cross-device journeys without a user ID. Those need CRM data or a warehouse, not another analytics tool.
Do I need Google Tag Manager to use Google Analytics for marketing?
Not strictly. But once you are running paid social, remarketing and conversion tracking on more than one platform, GTM is the only sane way to deploy tags without a developer ticket for every change. It also makes it far easier to respect a consent banner across all of your tags at once.
Wrapping up
Most GA4 accounts do not need more data. They need three or four key events that map to revenue, a tagging convention everybody follows, retention set to 14 months, and one dashboard that answers the weekly questions. That setup takes an afternoon and outperforms months of clicking around reports.
My honest advice on where to spend your effort: get the key events and UTM discipline right first, then link Google Ads and Search Console, then build the Looker Studio view. Skip attribution-model debates until your tagging is clean, because a data-driven model fed by inconsistent utm_medium values tells you nothing useful. And if you change one setting today, make it data retention — every day you leave it at 2 months is a day of history you cannot get back.
