Quick Answer: what a marketer needs from Google Analytics 4 when a report is due
Five days, about an hour a day, ending in one report: which channels produced the actions your business cares about. That is what a marketer needs from Google Analytics 4 when a deadline is real. Not fluency in the tool. One defensible answer, written down, in a report you can rerun next month without rebuilding it.
One number from our own database, queried on 2026-10-02. Learners here have built 1,630 learning paths: 326 for Python, 162 for SQL, 6 for marketing, 4 for Excel, 2 for analytics. Across all 5,790 video slots in those paths, 573 titles mention SQL and 6 mention analytics. Zero mention GA4 or Google Analytics. Zero mention Looker Studio or Data Studio. Zero mention Tag Manager. That zero is a real measured result, not missing data, and it is an observation about our own users rather than a claim about the world. It also matches how this skill lands: nobody signs up to learn the reporting tool, they get handed it with a date attached.
So this is not a channel list. It is five days with one job each, built from free videos, aimed at one deliverable: a report that names which channels produced the actions that matter, and that you can rerun next month without building it again.
Do this before you watch anything: check your data retention
Go into Admin, then Data Settings, then Data Retention, and write down what it says. For a standard Google Analytics property, user-level data retention can be set to 2 months or 14 months, and other event data has the same two choices, with longer ones reserved for Analytics 360. Do not assume which one yours is on. Large and XL properties are limited to 2 months either way. Changing it needs the Editor role, and Analytics waits 24 hours before implementing the change.
Two details decide whether this matters. The setting does not affect standard aggregated reports, including their primary and secondary dimensions; it only affects explorations and funnel reports. And raising the retention period applies only to data you have already collected and not already deleted, so it does not bring back anything that has aged out.
That is why this is step zero. Day 4 is one exploration, and an exploration reaches back only as far as your retention window allows. Sit on 2 months without noticing, spend Thursday learning Explore against a two-month window, promise a comparison you cannot build, and you repeat the week. Change it today and the 24-hour wait passes while you watch Day 1. The one reference you need: Google's data retention documentation.
The one decision that shapes the week: one answer or a standing report
Before Day 1, answer one question. Does this report come back next cycle, or is it one ask that dies when you send it? A one-off ends on Day 4, with a saved exploration and a screenshot. A report that comes back every month has to end on Day 5, scheduled, with its definitions written down.
Most people get this wrong in the same direction. The first request feels like a one-off, so they build it by hand, screenshot four charts into a slide, and present it. Then it comes back, and month two is another afternoon of clicking. By month four nobody can run it but you, and the definitions have drifted, because you made slightly different filter choices each time. A number that moves because your method moved is worse than no number.
So assume it comes back. Marketing reporting almost always does, because it is attached to a cycle rather than a question. One note before Monday: get the last report anyone sent on this, even if it was three numbers in an email, because your first job is explaining any difference between those figures and yours. If you have ever run a month-end report in a spreadsheet, our one-week Excel plan for data analysis is the same rhythm in another tool.
Day 1: make the property tell the truth
By the end of today you know which property you are actually reporting on, your own visits are out of the numbers, and the tag fires on the pages that matter. Day 1 buys you the right to trust anything you say later in the week. Every other day of this plan assumes it happened.
The concept that makes GA4 click: everything in it is an event, not a pageview-shaped hit with a session wrapped around it. So GA4 is two questions, not a hundred screens. Which events am I collecting, and which traffic am I counting? Today is the second one.
Three screens do the work. Admin, then Data Streams, shows what is actually sending data, which is how you find the stream somebody set up and forgot. Admin, then Data Settings, then Data Filters, excludes internal traffic, once you have defined your office and VPN addresses in the stream's tagging settings. Those same settings hold unwanted referrals, which is how a payment subdomain stops looking like a traffic source.
The mistake nearly everyone makes is trusting a filter that is not on. A data filter is created in one of three states, Testing, Active or Inactive, and you choose which. In Testing it only marks matching traffic with the "Test data filter name" dimension and leaves it in your reports. Read the state of every filter today.
Best for: Analytics Mania (~150K subscribers). The cleanest orientation pass, on 2026 screens that match yours. Start with: Analytics Mania, "Google Analytics 4 tutorial for beginners (2026 update) || GA4 tutorial for beginners || GA4 course", https://www.youtube.com/watch?v=-Y-RQtYLpHM
Best for: Loves Data (~150K subscribers). The same ground at a calmer pace. Start with: Loves Data, "Google Analytics 4 Tutorial (2026)", https://www.youtube.com/watch?v=YCCWXp408h0
Best for: Walker Analytics (3K+ subscribers). Admin-screen specific, and the admin screens are today's job. Start with: Walker Analytics, "Google Analytics 4 Admin Tutorial | GA4 Admin Tutorial", https://www.youtube.com/watch?v=cHLm2wb8hoI
Best for: MeasureSchool (~190K subscribers). A whole GA4 course in one 56-minute video; a refresher, not a shelf. Start with: MeasureSchool, "Full Google Analytics 4 Course for Beginners", https://www.youtube.com/watch?v=ZaBycjPJztk
Best for: Analytical Minds (~2K subscribers). A settings checklist, for a last pass before you call Day 1 done. Start with: Analytical Minds, "GA4 Top Features and Best Practices For Beginners | Lesson 14", https://www.youtube.com/watch?v=VnZi9zwF9KQ
On your own property today: list every data stream and data filter with its state. Define your internal addresses, activate the internal traffic filter, then load your own site and watch the Realtime report. You should see yourself arrive, and know tomorrow you will not.
Day 2: decide what counts as a conversion
By the end of today your property has two or three key events in it, agreed with whoever asked for the report, and they are being collected. A key event is GA4's name for a conversion. Until you pick yours, every acquisition report you build can only count visits, which is not what anyone is asking about.
The concept is smaller than the vocabulary suggests. Google renamed conversions to key events, and its own conversions versus key events documentation exists largely to explain the rename. A key event is an ordinary event you flagged as mattering. The flag is the whole mechanism: flag one, and your acquisition reports gain a column that counts it. That column is your report.
So the hard part of today is a conversation, not a setting. Ask the person who wanted this report what action, on the site, counts as a win. A form submitted. A demo booked. Write down two or three, no more. Flag everything and the key events column becomes a stranger version of the traffic count, which throws away the one column that gives your report a point.
The mistake is assuming the event already exists. The actions a marketer cares about are usually a form submission or a button click, and those often need their own tag in Tag Manager. Do it now: a key event that starts collecting on Thursday has nothing to show on Friday.
Best for: Loves Data (~150K subscribers). Treats key events as a reporting concept, which is what you need. Start with: Loves Data, "GA4 Key Events - How to Track Conversions in Google Analytics 4 (2025 Update)", https://www.youtube.com/watch?v=A546lw0g_4s
Best for: Google Analytics (~520K subscribers). The official lesson, in the wording Google's help uses. Start with: Google Analytics, "2.4 Understand & create key events in Google Analytics - Analytics Academy on Skillshop", https://www.youtube.com/watch?v=Ror4ubtzMq0
Best for: Analytics Mania (~150K subscribers). On the conversions-to-key-events rename, which makes older tutorials confusing. Start with: Analytics Mania, "NEW: Conversions and Key Events in Google Analytics 4", https://www.youtube.com/watch?v=w31Tn1x5rJk
Best for: Analytical Minds (~2K subscribers). For when the event does not exist yet and you need Tag Manager. Start with: Analytical Minds, "How To Implement Google Analytics 4 Tracking with GTM | Lesson 8", https://www.youtube.com/watch?v=7XQLhU2zWt8
Best for: Loves Data (~150K subscribers). Pairs key events with audiences, so you see where that goes later. Start with: Loves Data, "GA4 Tutorial (Part 5) - How to Set Up Key Events and Audiences", https://www.youtube.com/watch?v=d8go8MFk3ls
On your own property today: get a written answer on what counts as a win, flag those two or three events, then trigger each one on the live site. Watch the Realtime report until you have seen every one fire.
Day 3: tag your campaigns so the channel report resolves
By the end of today every link you control that points at your own site carries campaign parameters, spelled one agreed way. That is what makes the channel report resolve instead of dumping your work into Direct and Unassigned. This is the cheapest day of the week and the one that fixes the most.
Here is the mechanism, worth saying plainly because nobody explains it before handing you a UTM builder. GA4 knows where a visit came from only through what arrives with the request: the referring site, if there is one, and the campaign parameters in the URL. Default channel grouping is rules applied to those values. If utm_medium says cpc, the visit lands in a paid bucket. If it says email, it lands in email. If it says something nobody wrote a rule for, it lands in Unassigned, and if nothing arrives at all, it lands in Direct. Your channel report is not an opinion GA4 formed about your marketing. It is a lookup on strings you typed.
Which makes consistency the entire skill. Campaign values are case sensitive, so Email and email arrive as two separate source and medium values, and newsletter, Newsletter and news-letter as three. The channel rules will usually still bucket them together, but your campaign reports fill up with near-duplicate rows that nobody can total. The fix is boring and it works: one shared sheet of allowed values for source and medium, one person who adds to it, every link built from it.
Two failures eat the rest of the day. Links that pass through a redirect or a shortener can arrive stripped, the most common reason a campaign you definitely tagged shows as Direct. And tagging internal links overwrites real campaign credit with your own homepage banner. Bot traffic is the third, padding every channel unevenly, which is why bot filtering sits here and not on Day 1: it is a reporting problem, not a setup problem.
Best for: Two Octobers (~1K subscribers). UTM discipline as an attribution problem, not a form. Start with: Two Octobers, "Using UTMs to Improve Attribution Tracking", https://www.youtube.com/watch?v=O6H8HDR43Gg
Best for: Clicks Don't Lie (3K+ subscribers). Goes from tagged links to the campaign reports themselves. Start with: Clicks Don't Lie, "How to Track & Analyze Campaigns in Google Analytics 4 (GA4)", https://www.youtube.com/watch?v=FAmkxJtwynQ
Best for: Clicks Don't Lie (3K+ subscribers). Email breaks most often, because another tool builds the links. Start with: Clicks Don't Lie, "How to Track & Analyze EMAIL Campaigns in Google Analytics 4 (GA4)", https://www.youtube.com/watch?v=xx3aA19ENk8
Best for: Apotheca Marketing (under 3K subscribers). Campaign reporting framed for a marketer, which is Friday's framing. Start with: Apotheca Marketing, "Marketing Campaign Reporting in GA4", https://www.youtube.com/watch?v=E4nnYJQJfPw
Best for: Two Octobers (~1K subscribers). Today's data-quality half: bots padding whichever channel they land in. Start with: Two Octobers, "Filtering Bot Traffic in GA4", https://www.youtube.com/watch?v=x_huuna14TM
Best for: Deciphering Digital (~10K subscribers). How GA4 assigns credit across touchpoints, in plain terms. Start with: Deciphering Digital, "Understanding Attribution Models in Google Analytics 4 (GA4)", https://www.youtube.com/watch?v=CDjtDTsb0Z8
Best for: Walker Analytics (3K+ subscribers). Conversion paths, the report that shows what your tagging bought. Start with: Walker Analytics, "Conversion Paths In Google Analytics 4 | GA4 Tutorial", https://www.youtube.com/watch?v=64AuhyVhw6o
On your own property today: open last month's traffic acquisition report and find the Unassigned and Direct rows. Build your allowed-values sheet, retag every link you control, then click one. It should arrive in the channel you intended.
Day 4: one exploration that answers what the standard reports cannot
By the end of today you have one saved exploration that answers your actual question: which campaigns produced your key events, and where people fell out on the way. Standard reports answer fixed questions well. Explore is where you ask a question nobody built a screen for, including yours.
Explore is a pivot builder over your events, and that is the only mental model you need. Dimensions go into rows to decide what you group by, metrics into values to decide what you measure, and segments restrict which sessions count at all. A free-form exploration is that grid. A funnel exploration is the same idea with ordered steps and a drop-off rate between them. A path exploration answers what happened next, a different question from how many.
Pick one question and build one thing. The useful default for a marketer is a free-form exploration with session campaign or session default channel group in rows, your key events as the metric, and a segment limited to new users if the report is about acquisition.
Three things go wrong. The retention window from step zero bites first: Explore cannot show what the property no longer holds, and this is the only part of GA4 where that setting applies. Second, explorations are private to you until you share them, so Thursday's work is invisible on Friday unless you act. Third, a row reading "(other)" means the dimension has more distinct values than the report can hold, and the fix is a narrower dimension.
Best for: MeasureSchool (~190K subscribers). The free-form lesson, which is the shape you will ship. Start with: MeasureSchool, "How to Use GA4 Explorations | Free-Form Exploration (Lesson 5)", https://www.youtube.com/watch?v=TixOck2oCXg
Best for: Analytics Mania (~150K subscribers). The 2026 funnel walkthrough, for the drop-off question. Start with: Analytics Mania, "Funnel exploration in Google Analytics 4 (2026) || Funnel report in GA4", https://www.youtube.com/watch?v=1bK72QxzbJY
Best for: Analytics Mania (~150K subscribers). Path exploration, worth seeing once to know what it answers. Start with: Analytics Mania, "Path exploration in Google Analytics 4 (2026) || Path reports in GA4", https://www.youtube.com/watch?v=tX22B9qJkpU
Best for: Walker Analytics (3K+ subscribers). A practical pass over the Explore panels. Start with: Walker Analytics, "Google Analytics 4 Explore Reports Tutorial | GA4 Explorations", https://www.youtube.com/watch?v=97vIczmT8WI
Best for: Google Analytics (~520K subscribers). Short, official, and named the way Google's help names things. Start with: Google Analytics, "2.8 Use Explore for advanced analysis in Google Analytics - Analytics Academy on Skillshop", https://www.youtube.com/watch?v=MwlRg4B7X0c
Best for: Google Analytics (~520K subscribers). Segments, which turn a generic grid into an answer about one audience. Start with: Google Analytics, "Using Segments in Google Analytics", https://www.youtube.com/watch?v=cc8TC41mafY
On your own property today: build one free-form exploration with channel or campaign in rows and your key events as the metric, name it so a stranger could read it, save it and share it. Then write the one sentence it proves. If you cannot, it is not finished.
Day 5: ship it so it reruns without you
By the end of today the report leaves your hands on a schedule without you touching it, and a colleague can open it and read the answer. Day 5 is the difference between having done a piece of analysis once and owning a report. Skip it and you rebuild this week every month.
There are two honest routes and you need one. The cheap route is a scheduled email of a GA4 report: minutes of work, enough when the audience is two people who want the same four numbers. The durable route is Looker Studio connected to your GA4 property, which costs an afternoon and gives you a date range control, a filter people can use themselves, and a layout you decide. If the audience reaches past your team, build the second one.
What makes a standing report survive is treating it as a contract rather than a picture. Same question, same definitions, same cadence, every time. So write the definitions onto the report itself: which key events are counted, which date range, which segment, what you excluded. Four lines of small text at the bottom prevents most of the arguments a recurring report generates.
The usual mistake is twenty tiles. A dashboard with twenty tiles has no point of view, so nobody reads it and nobody notices when it breaks. Build the one chart that answers the question, one table underneath with the channel detail, and the definitions.
Best for: Clicks Don't Lie (3K+ subscribers). The cheap route done properly, straight out of GA4. Start with: Clicks Don't Lie, "Schedule Email Reports in GA4 | Automatic reports in Google Analytics 4", https://www.youtube.com/watch?v=mkX1QQu9Zhs
Best for: Analytics Mania (~150K subscribers). The durable route, current for 2026; follow this one if you follow only one. Start with: Analytics Mania, "Looker Studio tutorial for Google Analytics 4 (2026) || Build reports and dashboards", https://www.youtube.com/watch?v=2IeVO5WT_Ok
Best for: Loves Data (~150K subscribers). The build narrated slowly rather than demonstrated fast. Start with: Loves Data, "Google Analytics Dashboards Tutorial (Step-by-Step)", https://www.youtube.com/watch?v=40bPBomj9ik
Best for: Apotheca Marketing (under 3K subscribers). For when paid spend has to sit next to GA4 outcomes. Start with: Apotheca Marketing, "Google Ads + GA4 Dashboard in Looker Studio", https://www.youtube.com/watch?v=c8ywYTx2lws
Best for: Walker Analytics (3K+ subscribers). The failure modes, watched before you ship rather than after. Start with: Walker Analytics, "6 Reasons Why Your Looker Studio Dashboard Is Broken - And How To Fix Them!", https://www.youtube.com/watch?v=ybFhTjXsypE
Best for: Deciphering Digital (~10K subscribers). On making the report say something people act on. Start with: Deciphering Digital, "Using Looker Studio for Reporting: Actionable Insights from GA4", https://www.youtube.com/watch?v=kW_0p3AxM8Q
Best for: Analytical Minds (~2K subscribers). The plain case for why Looker Studio exists at all. Start with: Analytical Minds, "Introduction to Looker Studio and Why You Need It | Lesson 1", https://www.youtube.com/watch?v=Znd9SEKfmgs
On your own property today: ship it. Schedule the delivery or publish the dashboard, send it to one person who was not involved, and ask what it says. You are done when the next send happens without you opening GA4.
What to skip this week
Skip server-side tagging, the BigQuery export, custom dimensions beyond the one you need, comparing attribution models, and ecommerce reporting if your site does not sell anything. All five are real, several are valuable later, and none gets a report out of your hands by Friday.
Server-side tagging. Moving your tags off the browser onto a server you control. The serious answer to browser restrictions and tag bloat, and an infrastructure project, not a setting. It stops being skippable when measurement is visibly degraded and somebody has budget.
The BigQuery export. Piping raw GA4 events into a data warehouse to query them directly. The right move once your questions outgrow the interface, but a different job: it needs SQL and somebody to own a pipeline. If that appeals, our SQL reporting plan for CRM managers is the first step, and our guide to going from data analyst to analytics engineer shows where that road ends.
Custom dimensions, beyond the one you need. Each one is something to maintain and something a future report can get subtly wrong. Create the one your key event requires and no others.
Comparing attribution models. It feels sophisticated and it is the wrong order. Model comparison redistributes credit for key events, so until Day 2 gives you key events worth crediting, you are comparing two ways of dividing nothing.
Ecommerce reporting, if you do not sell online. Much GA4 material assumes a purchase event, a product list and revenue. If your site generates leads instead, that is vocabulary you do not need, and it sends you hunting for reports your property will never populate.
Two things are month two. Audiences, where key events get reused for remarketing: MeasureSchool, "Watch Me Create 13 GA4 Audiences", https://www.youtube.com/watch?v=FqD9YPjHw7Q and Analytical Minds, "How to Build Custom Audiences in GA4 for Remarketing | Lesson 13", https://www.youtube.com/watch?v=3u5KZRj1R9Y . And traffic from AI assistants: Clicks Don't Lie, "How to see AI traffic in Google Analytics 4 (Track AI in GA4)", https://www.youtube.com/watch?v=CywCLvPmGDc and Two Octobers, "Reporting on AI Traffic in GA4", https://www.youtube.com/watch?v=fpo1_E9UizI .
Frequently Asked Questions
What does a marketer actually need to know in Google Analytics 4?
Four things, in order. Which traffic counts as real. Which actions count as key events. How campaign tags make the channel report resolve. And how to build one exploration that joins those together. Everything else in the interface is for a different job than reporting on your own marketing.
Can you learn Google Analytics 4 in a week?
Enough to ship one report, yes. Five days at about an hour each covers data filters, key events, campaign tagging, one exploration and a scheduled delivery. It will not make you fluent, and it will not cover server-side tagging or the data warehouse. It gets one defensible report out on time.
What is the difference between a conversion and a key event in GA4?
Google renamed conversions to key events in Google Analytics 4, and its own documentation now compares the two terms directly. In practice a key event is an ordinary event you have flagged as one that matters, so reports can count it. The flagging is the only thing that makes it special.
Why does my tagged campaign traffic show up as direct in GA4?
Usually the parameters did not survive the click. A redirect stripped them, the link was shortened, someone pasted the untagged version, or the destination bounced the visitor to another URL. Direct is what GA4 shows when nothing on the arriving request says where the visit came from.
Should I change my GA4 data retention setting before I start?
Go and look at it first, because a standard property has two options and you cannot assume which one yours is on. Raising it applies to data you have already collected and not deleted, so it will not bring back what aged out. Changes take 24 hours, and you need the Editor role.
Does LearnPath give a certificate for finishing a Google Analytics 4 learning path?
Yes, on the Pro plan; free accounts do not get one. LearnPath builds a path of free videos, generates a quiz from each transcript, and gives Pro learners a certificate when the path is finished. Not one of the 5,790 videos queued here mentions Google Analytics, so treat it as proof you finished your own plan.
Proving you did it
At the end of the week you should be able to put three things on a screen. The saved exploration, shared, named so a stranger could read it. The scheduled report or published dashboard, with its definitions on it. And one sentence you can say in a meeting: these channels produced these key events last month, measured this way, and here is what I would spend differently. That sentence is the deliverable.
Be honest about what five days bought. Not fluency. Somebody who has run GA4 for three years will ask you something you cannot answer. What you have is one report you understand all the way down.
LearnPath turns a topic into an ordered path of free YouTube videos and generates a quiz from each video's transcript, so you find out whether the watching stuck. On the Pro plan, finishing a path issues a completion certificate; free accounts do not get one. That certificate proves you did the work in a plan. What proves you can do the job is a report that went out again the following month without you rebuilding it.
Zero of the 5,790 videos queued in paths here mention Google Analytics, Looker Studio or Tag Manager. That is a gap in timing, not ambition: nobody schedules the reporting tool, so it gets learned in the week the report is due. Give it five dated days instead. Build a work-ready Google Analytics 4 path before your next reporting cycle.
