Digital Marketing

GA4 Conversion Optimization: The First 10 Metrics to Check

Talha Aslan 19 min read 4 views

Short answer

What are the first 10 metrics for GA4 conversion optimization?

GA4 conversion optimization starts with ten metrics: sessions, engagement rate, engagement time, key events, session key event rate, user key event rate, landing page breakdown, channel breakdown, funnel step completion and revenue. They run from volume to value, so each one frames how you read the next.

Keep reading for the full details and examples

The sheer number of reports sends many teams in the wrong direction. In practice, when you have hundreds of dimensions and metrics but no starting point, you spend hours inspecting tables and still reach no decision.

In this guide, you will see what each metric tells you, when it misleads, and the order in which to read it. Also, all numeric examples are illustrative, not client results. For definitions, we therefore rely on the official Google Analytics Help and Analytics Data API documentation.

We wrote this as a plan for your first week of GA4 conversion optimization. For each metric, you ask one question, write one note and move on. As a result, the GA4 interface stops feeling like a maze and starts working like a diagnostic path.

Why does metric order matter for GA4 conversion optimization?

Because each metric changes how you read the one before it. For example, if you read a rate without knowing traffic volume, you draw big conclusions from a tiny sample. Likewise, if you read revenue without the rate, you cannot tell whether the problem sits in traffic or on the site.

So it helps to group the ten metrics into three layers:

  • Volume layer: how many visitors and sessions you have.
  • Quality layer: how interested visitors act on the site.
  • Value layer: whether that behavior turns into goals and revenue.

At every layer, first ask whether you have enough data. If you do not, then do not interpret. Wait for a longer period instead, because a few days of movement is usually noise.

What should you verify before GA4 conversion optimization begins?

Even the best report suffers from a faulty setup. Before you optimize, confirm that your conversion events fire correctly, do not double count and measure the real goal. Otherwise all ten metrics tell the wrong story.

To check your setup, use our guide on 50 checks for a correct GA4 setup. If conversions look lower than expected, our article on missing GA4 conversions is the first stop.

Get the terminology straight too. Google Analytics Help describes a key event as an event that measures an action particularly important to the success of your business, and it says any event you collect can become a key event. Read the details on the key events help page. Interface labels can differ by language and version, so what matters is choosing the right event.

If your measurement is sound and the findings point at the page itself, the work often turns into a decision about web design services. Treat measurement and design as separate questions: measurement first, design second.

How does the summary table rank the ten metrics?

Start with the map. The table also shows the question each metric answers and the trap it falls into most often. We then go into detail in the sections that follow.

OrderMetricQuestion it answersCommon trap
1SessionsIs there enough traffic?Reading a mixed-source total
2Engagement rateDo visitors show interest?Low rate on one-page answers
3Average engagement timeDoes the page hold attention?Assuming long time is always good
4Key eventsHow many goal actions happened?Reading the count without volume
5Session key event rateWhat share of sessions reached the goal?Inflating it with many key events
6User key event rateWhat share of users reached the goal?Confusing it with the session rate
7Landing page breakdownWhich entry page converts?Judging a low traffic page
8Channel and source breakdownWhich traffic has quality?Ignoring unassigned traffic
9Funnel step completionWhere do users drop out?Defining the steps loosely
10Revenue and revenue per userDoes conversion earn money?Forgetting refunds and margin

1. Why do sessions come first in GA4 conversion optimization?

Sessions are the denominator behind every other rate. The Analytics Data API schema describes sessions as the number of sessions that began on your site or app, triggered by the session start event. In short, every rate calculation rests on this number.

Check two things with sessions. First, is there a sudden spike or collapse between periods? Second, is the volume big enough to make rates trustworthy? In a small sample, a single visitor also swings the rate hard.

Example scenario: a page gets twenty sessions per week and produces two conversions. So the rate looks like ten percent. If the next week brings one conversion fewer, the rate falls to five percent. However, that halving is not a real decline. It is the result of tiny volume, so look at monthly totals in such cases.

Also avoid trusting the total without splitting by source. For example, in the week you launch an ad campaign, sessions jump, but the quality of that traffic may differ from organic visitors. Therefore, view volume as a total first and by channel second. That way, you spot where growth comes from and how it affects conversion.

2. What does engagement rate tell you about visitor interest?

Engagement rate is the share of engaged sessions out of all sessions. Google Analytics Help defines an engaged session as one that meets at least one of three criteria: it lasts longer than 10 seconds, it has a key event, or it has two or more screen or page views. Bounce rate is the percentage of sessions that were not engaged. The definition comes from the engagement rate help page.

This metric is your first quality filter. Specifically, a low engagement rate means visitors arrived by mistake or the page did not meet their expectation. However, a page that answers everything on one screen can also look low, so read the metric together with conversion rate.

For more on this, see our guide on how to reduce bounce rate on a business website. We give no threshold values, because they vary by industry and page type.

The practical way to use this metric is to compare your own pages with each other. For example, compare the engagement rate of your service pages. If one sits far below the rest, then that page may have a message or design problem. In other words, you take the benchmark from your own site, not from outside numbers.

3. When does average engagement time mislead you?

Engagement time relates to how long users keep your site in the foreground. The Data API schema describes user engagement duration as the total time your website or app was in the foreground of users' devices. Then divide it by active users, and you get a gauge of attention depth.

A long time does not make a page good. For example, a confused visitor might wander around the page. Likewise, a short time is not always bad, because someone looking for a phone number can find it and leave fast.

So pair this metric with key event rate rather than reading it alone. When attention is high and conversion is low, the problem usually sits in the offer, the form or the trust signals. When attention is low, instead question the match between message and traffic.

Also judge time by page type. For instance, you expect a long time on a detailed guide and a short time on a contact page. Instead of one site average, read similar pages together, because an average hides different behaviors.

4. What do key events alone tell you?

The key event count shows how many times your important actions happened. The Data API schema defines it as the count of key events, and it notes that marking an event as a key event affects reports only from its creation date, not historic data.

That second note matters. So if you mark an event as a key event today, it does not rewrite earlier periods. So always note the date of a setting change when you compare periods.

The count alone is not enough. You only learn whether forty conversions a month is good or bad when you compare it with traffic volume. Therefore, move to the rates next. If you have several key events, decide in advance which one is the real business goal.

The choice looks hard, but one question settles it: if this action happens, does my business make money? A phone tap, a form submission and a purchase usually say yes. Scrolling or watching a video, on the other hand, is often a supporting signal. Track supporting signals separately, and calculate conversion rate only from the primary goal.

5. How do you read session key event rate?

Session key event rate is the percentage of sessions in which any key event fired. The Data API schema defines it exactly this way. It is the main rate in conversion optimization, because it shows how effective the site is regardless of traffic volume.

To verify this rate yourself, use a conversion rate calculator. First, divide conversions by sessions and compare the result with the GA4 report. If there is a gap, then investigate the definition difference.

Watch out for these points when you interpret the rate:

  • On a site with several key events, the rate can inflate, because any one of them counts.
  • If you make micro conversions key events, the rate rises but business value does not.
  • Between periods, the traffic mix may have changed.

For micro conversions, read our article on micro conversions in e-commerce.

Mind the timing as well. When you calculate a rate, take the numerator and the denominator from the same period. For example, dividing yesterday's traffic by today's conversions gives the wrong answer. So keep the report window fixed, and compare periods using windows of equal length.

6. How does user key event rate differ from the session rate?

User key event rate is the percentage of users who triggered any key event. The session rate counts sessions instead. The difference becomes clear in businesses with a long decision cycle.

Example scenario: a visitor returns on three different days and fills in the form on the third visit. On a session basis, for instance, one of three sessions converts. On a user basis, however, one user converted. So in a business with a long sales cycle, the user rate sits closer to reality.

Read the two rates side by side. If the session rate is low but the user rate looks reasonable, visitors decide over several visits. In that case, improving the return visit experience beats single session optimization. For both definitions, check the metric descriptions in the Data API schema.

In practice, write the session rate first and the user rate second. In practice, a big gap means visitors come back. Then an email reminder, return visit content and a clear second step can raise conversions. A small gap means you must win the decision mostly on the first visit.

7. How does a landing page breakdown show where to intervene?

The total rate gives the site average, but you intervene on a single page. So break session key event rate down by landing page. The page that earns the most sessions but converts poorly becomes your most valuable candidate for optimization.

Apply two filters when you read the breakdown. First, set aside pages without enough volume. Second, consider the purpose of the page: a blog post may not aim at form submissions, whereas a service page is where you expect direct conversion.

If you skip this distinction, you fix the wrong page. For instance, if you spoil the headline of an informative blog post because its rate looks low, you may weaken a page that earns traffic. Instead, the right approach improves the path to the related service on the blog page, and improves the offer and the form on the service page.

When a report shows a landing page labeled "not set", read GA4 landing page not set. For common page level mistakes, our sibling article on 7 common CRO mistakes in landing page optimization covers them.

8. How does a channel and source breakdown reveal traffic quality?

The same site gives different results to visitors from different channels. If you look at key event rate by channel or source, you see which traffic has quality. Then you point budget and effort in the right direction.

Here you need to understand channel grouping. Specifically, GA4's default channel group rules classify traffic by source and medium. As a result, when tagging is missing, traffic falls into "unassigned" or the wrong channel. We cover this in our default channel group guide.

To tag your own campaign links consistently, a utm builder makes the job easier. Also, inconsistent tags make the channel breakdown meaningless. As you compare rates per channel, also remember volume.

For example, direct traffic and organic search traffic carry different intent. A direct visitor already knows you, so the rate usually behaves differently. Instead, compare each channel with its own previous period. Do not stack channels on top of each other to pick a "winner," and you will read the data fairly.

9. How does funnel step completion find the conversion leak?

A funnel lists the steps users take toward a goal. In e-commerce, the order is usually product view, add to cart, begin checkout and purchase. The Data API schema defines add to carts, checkouts and purchases as separate metrics.

You look for the leak in step completion. For example, if the move from product page to cart is weak, the problem may be price, product information or trust. Meanwhile, if the move from cart to checkout is weak, suspect shipping information or form friction.

On service sites, the funnel is shorter: landing page, form view, form submission. A short funnel does not make the leak easy to find, though, because alternative paths such as a phone tap exist. Therefore, track both the form and the call as separate steps. For setup, see our GA4 funnel exploration guide.

When you read the funnel, think of the order as a user journey, not a single session. For example, some users add to the cart on one day and buy three days later. A short window can therefore make the real loss look bigger than it is. Choose the window length to fit your business model, and define the steps precisely, or the leak will appear in the wrong place.

10. How do revenue and revenue per user show the value of conversion?

The last metric tells you whether conversion earns money. According to the Data API schema, purchase revenue is the sum of revenue from purchases minus refunded transaction revenue. Total revenue, in addition, covers purchases, subscriptions and advertising.

Read revenue together with conversion rate. For example, if the rate rises while revenue does not, you may be attracting low value products or discount hunters. If the rate holds and revenue grows, then basket value has risen.

Revenue per user is just as useful. By channel or landing page, it also shows which traffic actually pays. A channel with a low rate but large baskets can beat a channel with a high rate and small baskets. As a result, this comparison directly affects how you split budget and effort.

In service businesses, GA4 may not show revenue directly. In that case, match form conversions with closed deals. Also remember margin: revenue is not profit. So keeping that in mind shapes the direction of your optimization.

In what order should you read the ten metrics?

Build a fixed reading flow. If you open metrics at random, then each table tells its own story and you reach no conclusion. The order below moves from volume to value.

  1. Check sessions and decide whether the data suffices.
  2. Measure traffic quality with engagement rate and engagement time.
  3. Read the key event count and the two key event rates.
  4. Break the rate down by landing page and find the weakest page.
  5. Break the same rate down by channel and separate traffic quality.
  6. Build the funnel and find the leaking step.
  7. Check revenue and revenue per user, then compare them with margin.

At the end of each step, write a one sentence note. That way, instead of ten tables, you end up with one diagnostic sentence.

What does a first reading look like for an example service site?

Consider an example scenario: a small service site collects requests through a form. Monthly traffic is sufficient, but requests fall below expectation. The team only says "conversion is low."

First, we check session volume: it is enough. Engagement rate looks reasonable. Session key event rate, however, is low. Then, in the landing page breakdown, we see that blog traffic carries most sessions, while very few people move on to the service page.

At this point, calculate your conversion rate for the service page by entering its sessions and form submissions. If the result matches GA4, then the problem is not measurement but routing. In other words, blog posts lack a clear path to the service page. So the fix is to add contextual links and a call to action to the posts. This is an example scenario, and results change with your own data.

Which finding points to which problem?

The table below connects each finding to a likely cause. These are not final diagnoses, but they help you ask the right question.

FindingLikely problemFirst check
High sessions, low engagement rateWrong traffic or expectation mismatchChannel and ad message
High engagement, low key event rateWeak offer, form or trustCall to action and form
One channel with a very low rateTagging or targeting issueUTM and channel group
Add to cart but no checkoutCheckout frictionShipping info and form step
Conversions but low revenueLow value products or heavy discountsProduct and basket value
Low session rate, fair user rateLong decision cycleReturn visit experience

Use the table like a checklist, not like a verdict. Also, the same finding can point to more than one cause. For example, low engagement can come from the wrong channel or from a slow page. So for each row, run the cheapest check first, then move to the more expensive change. As a result, you do not waste effort.

What do you do when GA4 and Google Ads numbers do not match?

First, do not panic. The two systems run on different measurement models and attribution logic, so identical numbers are not expected. A gap can also signal a faulty setup, so look for the cause.

The check order goes like this: make sure you use the same period, the same time zone and the same conversion definition. Next, see whether the event you marked as a key event in GA4 matches the conversion imported into Google Ads. We explain the details in why GA4 and Google Ads numbers do not match.

For an optimization decision, treat one source as the reference and use the other for direction. Your choice depends on where the decision happens: GA4 may suit on site improvement, and Google Ads may suit bidding.

What optimization plan should you follow in the first 30 days?

The goal of the first month is reliable measurement, not winning. First, in week one, verify the setup and settle your key events. Second, in week two, read the ten metrics once and note your baseline. Then in week three, pick the weakest page and channel. Finally, in week four, make a single change and wait for the result.

A calm 30 day plan keeps GA4 conversion optimization from turning into random tweaking. Two rules matter in this plan:

  • Make one change at a time, or you cannot tell what worked.
  • Collect enough data before you judge the outcome.

For comparison, an ab test calculator helps you see whether a difference is meaningful. So we do not give target numbers here. Every site starts from a different point, so set current values from your own history.

Which mistakes should you avoid with these ten metrics?

Most GA4 conversion optimization mistakes start with definitions. The first mistake is reading a rate without volume. Then comes comparing periods without noting a settings change. Finally, there is counting every key event with the same weight.

Other common mistakes include:

  • Reporting a micro conversion as if it were the main goal.
  • Ignoring unassigned traffic in the channel breakdown.
  • Ignoring refunds and margin when you read revenue.
  • Changing a page because of one day of movement.
  • Mistaking a measurement problem for an optimization problem.

The last item is the most expensive mistake. For example, if an event fires twice, the rate looks higher than reality, and you may crown a page that was already working. If it never fires, however, the opposite happens. So before you change anything, confirm once more that the setup works.

For the general framework, see our article on the most important CRO metrics. This guide focuses on the GA4 reading order, so we do not repeat the general CRO tips covered there.

When should you get web design help after the metrics show a site problem?

Data often tells you where the problem is: a weak landing page, an unclear offer, a long or confusing form. If these findings need code or design changes, a team that works with measurement knowledge makes a difference.

Of course, you can make simple fixes yourself: clarify a headline, make the call to action visible, cut form fields. However, when page architecture, speed and mobile experience cause problems together, you need professional design work. In short, the important thing is to decide design by measurement findings, not by taste.

We explain how a conversion focused redesign works in what is conversion focused web design. We promise no results, but we recommend working with measurable goals.

How do you start with a quick action list?

The short list below is enough to read the ten metrics in your own account for the first time. Also, plan each item as one day of work.

  • Write down your key events and your main business goal.
  • Record the last forty five days of sessions and their fluctuation.
  • Separate engagement rate and engagement time by page type.
  • Put session and user key event rates side by side.
  • From the landing page and channel breakdowns, pick the one weakest page and one channel.
  • Draw the funnel and mark the biggest leaking step.
  • Read revenue together with margin.

After you finish these seven steps, you hold a diagnostic note. Then the next step is to test a single change. If you need help on the design layer, Talha Aslan and our team can work from your measurement findings through business website design built around your goals.

Which official sources should you use to verify the definitions?

GA4 metric names and definitions can change over time. For engagement rate and engaged sessions, the Google Analytics engagement rate page is the most reliable source. For the key event definition, see the key events page. You find metric names and descriptions in the Analytics Data API schema.

The GA4 conversion optimization advice in this article rests on definitions from those pages. Still, verify interface labels in your own account, and when a menu changes, simply look for the relevant section of the panel.

Core guide: What Is Digital Marketing? Its types, examples and how it is done, on one page with an interactive film that travels through the brain.

Frequently Asked Questions

Which metric should I check first in GA4 conversion optimization?
Start with sessions, because every other rate uses them as the denominator. Then read engagement rate, the key event count and session key event rate. If volume is too small, do not interpret the numbers yet. Wait for a longer period, then break the rate down by landing page and channel as the next step.
Is a key event the same as a conversion in GA4?
Google Analytics Help now calls an event that measures an action particularly important to your business a key event. The purpose matches the older conversion idea, but a Google Ads conversion is a separate concept. Definitions can differ between the two systems, so state in your report which one you use.
Does a low engagement rate always mean a problem?
No, not always. An engaged session lasts longer than 10 seconds, includes a key event, or has at least two page or screen views. A page that answers everything on one screen can look low. Read the rate together with conversion rate, channel and page purpose before you decide.
What is the difference between session and user key event rate?
Session key event rate is the percentage of sessions in which a key event fired. User key event rate is the percentage of users who triggered one. The two drift apart when people decide over several visits. For long decision cycles, the user rate is closer to reality, and reading both is safest.
Why do GA4 and Google Ads conversion counts differ?
The systems use different measurement models, attribution logic and timestamps. One side may also count a key event while the other counts an imported or tag based conversion. Align the period, time zone and definition first. Then decide which source you will use for which decision.
How much data do I need before I start conversion optimization?
There is no fixed number of months, because the needed volume depends on traffic and how often conversions occur. The general principle is to collect enough data that one visitor cannot swing the rate. With low traffic, read monthly totals and judge a single change only after enough data builds up.
  • ga4
  • conversion optimization
  • key events
  • engagement rate
  • landing pages
  • funnel analysis
  • ga4 metrics
Share:
Talha Aslan

Google Partner digital marketing expert. Hands-on with SEO, Google Ads, web design and e-commerce projects since 2012; every post here comes from that experience.

Next project

Get a quote for your project.

Your brief goes straight to Talha Aslan and team: strategy led by Talha, delivery by an experienced team. We reply the same day and send the scope, timeline and price as a written quote within 1 business day.