Looker Studio SEO Dashboard in 1 Hour

A Looker Studio SEO dashboard is a single report page that pulls your Search Console and GA4 data into one view: clicks, impressions, organic sessions, key events and the pages behind them, plus a block for traffic from AI assistants. You can build a working version in about an hour if you keep it to one page and a short list of fields.

This is for marketers and site owners who want something live instead of a monthly spreadsheet. It covers the connections, one blend, the AI referral block and the mistakes that make dashboards lie. Naming note: Google’s documentation now says “Looker Studio is now called Data Studio.” The steps are the same, so this post uses the older name that people still search for.

On this page

• What belongs on a one-page SEO dashboard?
• Connect Search Console and GA4
• How do you blend Search Console with GA4?
• Adding an AI referral segment
• Build the charts and controls
• Mistakes that make an SEO dashboard lie
• Getting your SEO dashboard working

What belongs on a one-page SEO dashboard?

A one-page dashboard needs four scorecards, two trend charts and two tables, and nothing else. Anything more turns into a wall of charts that nobody reads. The test for each block is whether a change in it would make you do something.

The diagram below shows the layout. Search Console supplies the visibility numbers (clicks, impressions, queries). GA4 supplies what happened after the click (sessions, key events). The AI block comes from GA4 only, because that is where assistant referrals show up.

Scroll the diagram sideways to see all of it.

If you already have a monthly format, use it as the spec. Our SEO report template lists the six sections a good monthly report carries, and the SEO KPIs worth tracking explains which of these numbers earn a place. The dashboard is the live version of the numbers sections; the commentary still gets written by a person.

Connect Search Console and GA4

You need two data sources, and the Search Console one comes in two flavors. In Google’s connector documentation, a Search Console data source uses either the Site Impression table (aggregated by site) or the URL Impression table (aggregated by page URL), and one data source can use only one of them.

Use this table to decide what to create:

For each one, Google’s steps are the same: sign in, choose Create, then Data source, pick the connector, authorize if asked, choose the site or property and table, and click Connect. Three details are worth knowing before you start:

• You must be signed in to the tool and to Search Console with the same Google account, and you need at least view permission on the property.
• For GA4, Google requires at least Read & Analyze permission on the property you connect.
• Search Console may hide some queries for privacy, which can show up as a blank query row. Filter nulls out of the queries table so it does not mislead.

Set the credentials option deliberately. Google’s connector page lets you choose the owner’s credentials, the viewer’s credentials or a service account. Owner’s credentials let a client see the dashboard without having Search Console access themselves, which is usually what you want for a shared link, but it also means anyone with the link sees your data. Think before you send it.

If you are new to the source data, the Search Console guide explains the Performance report these fields come from.

How do you blend Search Console with GA4?

A blend joins two data sources on a shared field, and for SEO the useful key is the landing page. Google’s documentation says each source in a blend must share one or more dimensions, called a join key, and that the blend keeps all records from the leftmost source plus matches from the sources to its right (a left outer join).

That gives you a table where each row is a page, with Search Console clicks and CTR beside GA4 sessions and key events. It answers the question neither tool answers alone: which pages earn clicks but fail to convert them into something useful.

The catch is that the two fields do not look alike. The URL Impression table aggregates by page URL, while GA4’s Landing page dimension is the page path of the first pageview in a session. A full URL will not equal a path, so the join returns nothing until you fix it. Google’s blend management page suggests a calculated field to align join fields that are formatted differently, so create one on the Search Console side that trims the domain and leaves the path, then join on that.

The diagram shows the join and where the mismatch sits.

Scroll the diagram sideways to see all of it.

To create the blend in the editor, Google lists three routes: select charts and right-click to choose Blend data, use +BLEND DATA in a component’s properties, or open the Resources menu and choose Manage blends. Put Search Console first so it is the left table. Check the totals afterward: a blend that silently drops rows is the most common quiet error in SEO dashboards.

Adding an AI referral segment

For AI traffic, chart sessions by the AI Assistant channel in GA4’s default channel group, then add a small custom field for anything it misses. Google Analytics’ release notes say that from May 13, 2026, a new AI Assistant channel appears in the default channel group, with visits from recognized assistants tagged with the medium “ai-assistant”. Google’s channel definitions say it covers sources like ChatGPT, Gemini, Deepseek, Copilot and Grok.

Two limits matter for a dashboard:

• AI Overviews and AI Mode are not in this channel. Google’s definitions keep them in Organic Search. So the AI block shows assistant referrals, not exposure in Google’s own AI features.
• History is unclear. The release note does not say whether older data is reclassified. Do not read a jump on the chart as a jump in demand until you have checked the date the channel started appearing.

Then there is the gap in what Google says it recognizes. The documentation names five assistants; it does not promise a complete list. Look at your Session source values in GA4 and, if a source you care about sits in Referral, add a calculated field in the data source. A searched CASE with REGEXP_MATCH does it:

CASE
  WHEN REGEXP_MATCH(Session source, ".*(perplexity|claude).*") THEN "AI referral (other)"
  ELSE "Other"
END

REGEXP_MATCH returns true only when the whole value matches the pattern, which is why the pattern starts and ends with .*. Use real hostnames from your own data rather than a copied list, and note the date you add each one. Our guide to tracking AI traffic in GA4 and Search Console covers the hostnames and why referrer-stripping means these counts are a floor. Google’s connector page also says GA4 segments and comparisons are not supported, so a calculated field or a chart filter is how you do this here, not a saved GA4 segment.

Build the charts and controls

With the sources ready, the build is mostly placing the eight blocks. The diagram below sets out the hour as a plan. The minute counts are an estimate to plan by, not a benchmark.

Scroll the diagram sideways to see all of it.

Place the controls first, then the charts:

1. Date range control. In edit mode, use the control icon in the toolbar, choose the date range control and set a default range. By Google’s documentation it affects every chart on the page unless you set a date range on a chart itself or group the control with specific charts. Only charts whose data source has a date dimension respond.
2. Scorecards. Clicks from the Search Console source. Sessions and key events from GA4, filtered to the Organic Search channel. AI Assistant sessions filtered to that channel. Key events are the GA4 measure for actions you mark as important, such as form submissions, so pick the one that matches a lead.
3. Trend charts. Time series by week, not by day. Small AI numbers are noise day to day.
4. Tables. Top landing pages from the blend, sorted by clicks, and top queries from the site-level source with nulls excluded.
5. Comparison range. Set a comparison date range on the scorecards and time series so each shows change against the previous period or the same period last year.

Keep one chart per question. If you want a country or device split, make it a second page rather than crowding the first.

Mistakes that make an SEO dashboard lie

Most wrong dashboards fail on definitions, not on layout. These are the ones worth checking before you send the link.

The last row matters most. A dashboard that counts only clicks understates AI visibility, because an answer that names your brand can send no session at all. The metrics that actually measure AI visibility and the walkthrough for measuring brand visibility in ChatGPT cover what to track outside the dashboard.

Also check the first build against the source tools. Open Search Console and GA4 for the same dates, compare the headline numbers, and expect small differences. If a number is wildly off, the usual cause is a join or filter, not the data.

Getting your SEO dashboard working

Start with the one page above, ship it, and add detail only when someone asks a question the page cannot answer. A dashboard earns its keep by being opened, and a smaller one gets opened.

Two next steps follow. First, write the monthly commentary beside it, using our SEO reporting checklist to keep the routine consistent. Second, decide how you will measure what clicks cannot show: whether assistants name you, which sources they cite, and where competitors appear instead. That measurement sits outside GA4 and Search Console, and it is the work behind our AI visibility audit, which gives you a baseline to put beside the dashboard’s traffic numbers.