A GA4 Dashboard: Build One by Asking

Most GA4 dashboards fail by showing everything. The ones that work answer four or five recurring questions and handle the one number that cannot be added up.

A GA4 dashboard is useful when it answers four or five recurring questions, such as which channels bring sessions and revenue, how engaged those sessions are, which landing pages produce key events, and how monthly users are trending. The mistake to avoid is adding up users across periods. The rest is a matter of picking the right grain for each view.

Why GA4 dashboards sprawl

GA4 exposes hundreds of dimensions and metrics. A dashboard builder invites you to add them all, and the dashboard ends up a wall of numbers nobody reads. The discipline is to start from what people ask and work backward to the data. If no one has ever asked about it, leave it out.

Five views that cover most needs

1. Traffic by channel over time

Sessions per channel group per month, stacked. It shows mix shifts, such as organic declining while paid grows, that a single total hides.

2. Revenue and purchases by channel

Revenue, purchases and revenue per session by channel for the last quarter. Revenue per session is summed revenue over summed sessions, not an average of daily values. Revenue here is GA4's number, which will not equal your store's. See why conversions never match for why.

3. Engagement quality

Engagement rate by channel, shown with session counts. A channel with a high rate on 30 sessions is not a finding. GA4 calls a session engaged if it lasts at least 10 seconds, has a key event, or views two or more pages.

4. Landing pages that produce key events

Top landing pages by key events for the month, with sessions beside them. The ratio of key events to sessions tells you which pages convert, and the volume tells you which ones matter.

5. Monthly users

Total users and new users by month, on their own. This is the one view that must stay separate. Users are counted per period, so a three-month total made from monthly figures overstates the audience. Show them month by month and compare like periods.

Handling ratios and comparisons

Three rules keep a GA4 dashboard honest.

1. Sum, then divide. Engagement rate is total engaged sessions over total sessions. Key events per session is total key events over total sessions.
2. Always show the base. A rate next to the count it came from.
3. Compare against something. Last month and the same month last year. Seasonality makes a raw month-over-month change misleading in many businesses.

If you maintain a manual report, the same sections appear in the GA4 report guide.

Getting these views from Quiriz

Quiriz's Google Analytics connector is in beta. It reads one GA4 property, read-only, for the last 13 months, refreshes daily and is included in paid plans. The three datasets line up with the views:

• Daily traffic (day by default channel group) for views 1 to 3.
• Landing pages by month for view 4.
• Monthly users for view 5, with a note in the data not to add months together.

Because Google Analytics sessions, engaged sessions, new users, key events, purchases, revenue, engagement rate and key events per session are defined once, every view uses the same arithmetic. Ask:

• "How many sessions did we get from each channel every month this year?"
• "What was our revenue by channel last quarter, with purchases?"
• "Which landing pages brought the most key events last month?"
• "How many users did we have each month?"

The questions you ask every month can be kept as saved questions and rerun after each daily refresh. Answers show the table, so you can confirm a number against the GA4 interface in a minute.

An illustrative reading of the five views

Imagine the dashboard shows these (example numbers): organic sessions flat, paid social sessions up 40 percent, but engagement rate for paid social down from 62 percent to 48 percent on twice the sessions. Revenue per session for that channel fell as well. The first view says traffic grew, the third says it grew with weaker visitors, and the second confirms it did not pay back. Read together, the views tell you the new paid social audience is broader and less interested, which is a targeting conversation. Any one view alone would have looked like good news. That is the case for keeping a small set of views that are read together instead of a large set read separately.

One last habit: write the definition of each metric in a small notes block on the dashboard itself, including the property time zone and the date of the last refresh. Most arguments about GA4 numbers are really arguments about definitions, and a visible note ends them in seconds instead of in a meeting.

What this will not show

• Source and medium, and UTM campaign, are not in the data. Channels are GA4's default channel groups.
• Key events arrive as one total, not split by event name.
• The landing page view covers the top 5,000 pages per month by sessions, so very long-tail pages are left out and the page-level total is below the daily total.
• No demographics, device or geography.
• It is a set of answers rather than a fixed-tile screen for an office wall. If you need one, a BI tool or Looker Studio still fits.

For the questions behind these views, read asking GA4 questions in plain English. Teams running shops can see Quiriz for ecommerce.

Frequently asked questions