Build a Looker Studio Dashboard That Answers Real Questions

Why Most Looker Studio Dashboard Projects Fail Before They Start

A SaaS client came to me last quarter with a 14-page Looker Studio report their agency had built. It had scorecards, time series, a geo map, a funnel chart, and four different pie charts showing traffic by device category. It was gorgeous. And nobody opened it.

When I asked the marketing director what decision the dashboard helped her make, she paused for ten seconds. "I guess... we check if traffic went up?" That is not a decision. That is a reflex.

The dashboard was not broken technically. The GA4 connector was configured correctly. The data matched. The problem was architectural: nobody had asked what questions the report needed to answer before opening the editor. They started with charts, not decisions.

If you are new to the tool itself, I covered the fundamentals -- what Looker Studio is, how it compares to GA4's built-in reports, and where it fits alongside other BI tools -- in What Is Looker Studio? A Marketer's Guide. This post picks up where that one left off. We are building.

Start With Five Questions, Not Five Pages

Before you touch Looker Studio, write down the three to five questions your team asks every week. Not metrics. Questions.

Here are examples from actual client engagements:

• Which channels are driving qualified leads this month, and how does that compare to last month?
• Is our Google Ads spend producing conversions at an acceptable cost, or are we burning budget on non-converting campaigns?
• Which landing pages convert visitors, and which leak them?
• Are our email campaigns generating revenue, or just opens?
• Where in the funnel are we losing people -- top, middle, or checkout?

Each question implies specific metrics, dimensions, and comparisons. "Which channels drive qualified leads" requires sessions by source/medium, a conversion metric you trust, and a comparison period.

If you cannot map a chart to one of your questions, delete the chart. A Looker Studio dashboard earns attention by being useful, not by being decorative.

The Three-Page GA4 Dashboard Structure I Use With Clients

After building GA4 dashboards for over forty client properties, I have settled on a structure that works across ecommerce, SaaS, and lead-gen businesses. Three pages. No more until you have outgrown them.

Page 1: Acquisition Overview

This page answers: Where is traffic coming from, and is it the right traffic?

Component	Type	Data
KPI row	Scorecards	Sessions, Users, Conversions, Conversion Rate -- each with comparison to previous period
Channel trend	Time series	Sessions by default channel group, weekly
Source/medium table	Table with bars	Rows: source/medium. Columns: sessions, engaged sessions, key events, key event rate
Campaign filter	Drop-down	Filters the entire page by campaign name

The scorecard row at the top is the most-read element. Executives scan it and decide whether to keep scrolling or close the tab.

Page 2: Engagement and Content

This page answers: What are people doing on the site, and where do they drop off?

Component	Type	Data
Top landing pages	Table	Landing page path, sessions, engagement rate, average engagement time, key events
Event summary	Table	Event name, event count, users -- filtered to your key events
Scroll depth or engagement chart	Bar chart	Engagement rate by page or section
Device breakdown	Stacked bar	Sessions and conversion rate by device category

If you have set up GA4 events with proper parameters, this page becomes extremely useful. If your event tracking is sloppy, this page will expose it.

Page 3: Conversions and Revenue

This page answers: Are we making money, and from where?

Component	Type	Data
Revenue scorecard	Scorecards	Total revenue, transactions, average order value -- with comparison
Conversion by channel	Bar chart	Key events or purchases by default channel group
Campaign ROAS table	Table (blended)	Campaign name, Google Ads cost, GA4 revenue, calculated ROAS
Conversion trend	Time series	Weekly conversions with optional trendline

For the campaign ROAS table, you need to blend GA4 revenue data with Google Ads cost data. I cover the tricky parts of blending below.

Data Source Choices: GA4 Connector vs BigQuery

This is the single most consequential decision in your Looker Studio dashboard build. You have two primary options for GA4 data.

The native GA4 connector is fast to set up. You authenticate, pick your property, and your dimensions and metrics appear. The problem emerges at scale: the connector uses the GA4 Data API, which applies sampling when you exceed the query threshold. Query three months of data on a high-traffic property and you might be looking at sampled numbers without realizing it.

The BigQuery connector requires more upfront work -- you need to set up the GA4 BigQuery export and write SQL to transform nested event data into flat tables. The payoff: no sampling, no API quota limits, full control over attribution logic, and the ability to join GA4 data with CRM or backend data before it reaches Looker Studio.

My recommendation: start with the native GA4 connector if you have fewer than 500,000 sessions per month. Move to BigQuery when you need unsampled data, custom attribution, or blended data from non-Google sources.

A hybrid approach I use frequently: connect GA4 natively for the acquisition and engagement pages (where directional data is fine), and connect a BigQuery-sourced table for the conversions page (where accuracy matters).

Looker Studio Templates: When to Use Them, When to Build From Scratch

Google's template gallery and third-party Looker Studio templates are tempting. Copy one, swap in your data source, and you have a working dashboard in minutes. I have seen teams save hours this way. I have also seen templates cause more confusion than they solve.

When Looker Studio templates work

A GA4 dashboard template that shows sessions, users, top pages, and channel breakdown can work out of the box for a startup that just needs basic traffic visibility. Templates are also a decent starting point if you are learning the tool.

When templates fail

Templates break down when:

• The template was built for a different GA4 configuration. A template expecting ecommerce events will show blank charts if your property only tracks leads. A template using purchase revenue will not work if you track value through custom events.
• The metrics do not match your definitions. A template might define "conversions" as all key events. You might only care about demo requests. The chart looks populated but answers the wrong question.
• The template obscures your data quality problems. When every chart is pre-built, nobody asks whether the underlying data makes sense. A template full of numbers feels like progress even when the numbers are wrong.

The worst GA4 dashboard template mistake I encounter: teams adopt a template, never audit whether the GA4 property feeding it has clean conversion tracking, and spend months making decisions on data that was never validated. If you are not confident your tracking is clean, that is worth fixing before you build any dashboard. I typically start marketing measurement engagements with a tracking audit for exactly this reason.

Calculated Fields and Blending: Where Dashboards Get Powerful (and Fragile)

Calculated fields worth building

Looker Studio lets you create custom metrics using formulas. A few I add to nearly every Looker Studio report:

• Conversion rate: Key Events / Sessions -- because GA4's connector does not always expose this as a native metric for custom key events.
• Cost per conversion (blended): Cost / Key Events -- requires blending Google Ads cost with GA4 conversion data.
• ROAS (blended): Purchase Revenue / Cost -- same blending requirement.
• Non-bounce rate: Engaged Sessions / Sessions -- a cleaner engagement metric than raw engagement rate for some contexts.

When creating calculated fields, use the CASE function for segmentation. For example, grouping campaigns into branded vs non-branded:

CASE
  WHEN REGEXP_MATCH(Campaign, "(?i).*(brand|branded).*") THEN "Branded"
  ELSE "Non-Branded"
END

Data blending pitfalls

Blending in Looker Studio is a left outer join. The most common mistake: blending GA4 and Google Ads data on date alone without a campaign dimension, which produces a single aggregated row per day and makes campaign-level analysis impossible.

The second most common mistake: mismatched join keys. GA4 records campaign names from UTM parameters. Google Ads uses the campaign name from the Ads interface. If these do not match exactly -- and they often do not when UTMs are manually tagged -- your blended data will have rows that join to nothing.

My preferred approach: do the joining upstream. Export both GA4 and Google Ads data to BigQuery, join them in SQL where you can inspect the results row by row, and bring the clean, joined table into Looker Studio as a single data source. This eliminates the fragility of in-dashboard blending and gives you a Looker Studio report you can actually trust.

Five Mistakes That Ruin GA4 Looker Studio Dashboards

Across dozens of dashboard audits, these come up repeatedly:

1. No comparison period. A scorecard showing "12,456 sessions" means nothing without context. Always enable comparison (previous period or same period last year).

2. Defaulting to "Last 28 days." If your team reviews performance monthly, use calendar month date ranges. "Last 28 days" shifts with every view, which makes comparing screenshots from different meetings meaningless.

3. Mixing data sources without labeling. If one column comes from GA4 and another from Google Ads, someone will compare the two and ask why "conversions" shows two different numbers. They are different metrics from different platforms, and they rarely match exactly.

4. Building the dashboard before auditing the data. I cannot overstate this. If your GA4 tracking has known issues -- missing events, duplicate transactions, broken consent setup -- your Looker Studio dashboard will visualize the problems with pixel-perfect clarity. Fix the plumbing first.

5. No documentation page. Add a final page explaining which GA4 property powers the data, when it last refreshed, what "conversion" means in this context, and any known limitations. Dashboards outlive the person who built them.

Sharing and Maintenance

A dashboard that nobody opens is wasted work. A few practical points:

Set up scheduled email delivery. Looker Studio can email a PDF snapshot weekly or monthly. Even if your team prefers the interactive version, the email reminds people the dashboard exists.

Use viewer-mode filters, not edit-mode filters. Hard-coding a filter in edit mode means viewers cannot override it and may not realize it exists. Use interactive filter controls instead.

Review quarterly. Business questions change. The dashboard you built in Q1 may not answer Q3 questions. Schedule a quarterly check: are these still the right questions? Do the metrics still map to actual decisions?

FAQ

Not confident your GA4 data is accurate enough to dashboard? Let's fix that -- I will audit your tracking, validate the numbers against your backend, and build a reporting layer you can actually trust.