Data Studio tutorial: build your first report, step by step

What the rename changed, and what it did not

The release note of 16 April 2026 says it in one line: Google rebranded the product as Data Studio, and updated the home page to give one place for your reports and data sources. Google’s quick start guide, updated 26 October 2026, signs you in at datastudio.google.com and opens the home page on the Reports tab.

The editor’s menus kept their names. The quick start still uses Create, Add data, Add a chart and Share, the names older guides use. What moved since April is smaller, and all of it is in the same release notes:

One name did not change: Looker, without “Studio”, is a separate Google product, not another name for this one.

What Data Studio is, and whether it is free

Data Studio is Google’s report editor: you connect data sources, place charts and controls on a canvas, and share the result as a link that updates when the data does. Google calls it a no-cost tool (Welcome to Data Studio, October 2026), and the free edition is the one this tutorial uses.

What you get without paying, per Google’s own pages:

• Google’s connectors to BigQuery, MySQL, PostgreSQL, Microsoft SQL Server, Amazon Redshift, Cloud Storage, Google Sheets, Microsoft Excel, CSV upload, Google Analytics, Google Ads, Search Console, YouTube Analytics and the other Google Marketing Platform products (available connectors, October 2026). Partner connectors come from the gallery and may cost money; Google says so on its about data sources page.
• Viewing needs no Google Account. A browser is enough to open a shared report. Creating and editing needs a signed-in Google Account and a supported region, and a Google Workspace administrator may have to turn Data Studio on (what you need, October 2026).
• Real-time co-editing, scheduled email delivery, and embedding a report in another page.

Data Studio Pro is the paid version. It adds what a team needs once reports become shared property: content owned by the organization through a Google Cloud project rather than by the person who built it, team workspaces with Manager, Content Manager and Contributor roles, up to 200 delivery schedules per report and delivery to Google Chat, chart alerts, Gemini assistance, and Google Cloud Customer Care. Google offers a 30-day trial with no limit on users (about Data Studio Pro, October 2026). Nothing in this tutorial needs Pro, except where a step says so.

The home page is where you start. It lists your reports and data sources in one place; the Create button in the top left opens a new Report or a new Data source. Older courses also list Explorer under Create: Google deprecated it on 11 September 2025 and converted existing explorations to reports (release notes).

Data Studio vs Google Sheets

Data Studio does not replace a spreadsheet: it has no cells to type into. It reads data that lives somewhere else and draws it. That one difference decides most of the comparison.

Google’s help is plain about the first two rows: most data sources keep a live connection, and your data is not imported into Data Studio except temporarily, to answer a specific request (about data sources, October 2026).

In practice the two work together more often than they compete. The Google Sheets connector turns one worksheet into a data source; the data must be a table with a single header row and one data type per column, and by default Data Studio checks the sheet for new data every 15 minutes (connect to Google Sheets, manage data freshness, October 2026). Keep the sheet as the place where people type, Data Studio as the place where they look. If the sheet is itself filled from a Data Mart by the Google Sheets Extension, the sheet and the report can read the same definitions; more on that below.

Metrics, dimensions, and three more terms to know

Five words in Data Studio mean something narrower than they do elsewhere, and the last one matters for everything after the tutorial.

• Dimension. A set of values you group by: names, categories, dates, descriptions. Browser, Landing page and Campaign are dimensions; so are unaggregated numbers such as a price (dimension, April 2026).
• Metric. An aggregation applied to a set of values, by COUNT, SUM, AVG and the like. A metric has no set of values of its own, so you cannot group by it (metric, April 2026). Every metric carries an aggregation: Sum, Average, Count, Count Distinct, Max, Min, Auto or None. A field left at None in the data source is summed in reports, and Auto, which Data Studio applies to formulas that already aggregate, cannot be changed (aggregation, October 2026).
• Connector. What Data Studio uses to reach a system: the BigQuery connector, the Google Analytics connector, or a partner connector from the gallery.
• Data source. One connection through a connector, with its fields and the credentials it reads with. Sharing a report does not share its data source: what a viewer sees depends on the data source’s credentials (quick start guide).
• Calculated field. A field you write as a formula. Google has two kinds, decided by where you create them (about calculated fields): a data source calculated field is available in any report on that data source and needs edit rights on it; a chart-specific calculated field exists only in the chart you created it in.

Where a data source is created decides who can change its definitions. A data source you add while editing a report is embedded: anyone who can edit the report can edit it, and it is copied with the report. One you create from the home page is reusable across reports, and only the people you share it with can edit it (about data sources, October 2026). The aggregation matters as much as the field: a column of conversion rates set to Average gives a different number from a formula that divides total orders by total sessions, which is the whole of the example further down. The same split in Google Analytics is covered in metrics and dimensions in GA.

Build your first Data Studio report

These steps follow Google’s quick start guide, with the menu names as Google prints them. Use Google’s sample data the first time: nothing to authorize, nothing to pay for.

Step 1. Create a report

1. Go to datastudio.google.com and sign in with your Google account.
2. In the navigation panel, click Create, then Report.

The report editor opens with the Add data to report panel already showing.

Step 2. Add data

The fastest route uses data Google provides:

1. In the Add data to report panel, click My data sources.
2. Select Sample Google Analytics Data.
3. Click Add, then confirm.

A table of fields appears on the canvas. Click Untitled report in the header and give the report a name.

To use your own BigQuery data instead, click Connect to data, select the BigQuery connector, click Authorize if asked, and choose a project, a dataset and a table. Google’s guide practises on a public one, the faa dataset’s us_airports table. Connecting to BigQuery needs a Google Cloud project with a billing account. The same panel lists every other connector, Google’s and partners’.

Step 3. Add a chart

1. In the toolbar, click Add a chart and pick a chart type.
2. Click the canvas where the chart should go, then move and resize it.
3. Add dimensions and metrics in the properties panel on the right, or drag fields from the data panel onto the chart.

Dragging a field straight onto an empty part of the canvas creates a chart from it. The properties panel has two tabs: Setup decides what the chart shows, Style how it looks.

Which chart to pick depends on the question. Google’s list of chart types (types of charts, October 2026), with what each is for:

Charts from third-party developers sit under Community visualizations and components in the toolbar. Google marks them Beta, and they need the data source to use the owner’s credentials (community visualizations, October 2026). For a marketing report, the first four rows cover most of what people ask; a funnel chart is worth knowing before you fake one with bars.

Step 4. Add a date range control

1. In the toolbar, click Add a control and choose Date range control.
2. Place it on the report.
3. In the properties panel, set a default range or keep Auto date range.

By default a date range control governs every chart on the page, but only charts whose data source has a date dimension (date range control). A chart that ignores the control has no date field to filter on.

The same Add a control menu holds the rest (about controls, October 2026):

• Drop-down list and fixed-size list: pick values of a dimension, such as Campaign or Country.
• Input box and advanced filter: type a value, or a condition with operators and regular expressions.
• Slider and checkbox: a numeric range, or a true or false field.
• Preset filter with buttons: one click applies a filter you configured.
• Dimension control: viewers swap the dimension a chart breaks down by, without editing it.
• Data control: viewers switch the dataset the report reads, so one report serves many accounts.

Three rules explain most “my filter does nothing” moments. A control filters at the data source level, so the field does not have to be in the chart, only in its data source. It matches fields by field ID, not display name. And across data sources it works automatically only when the field IDs match: true for two data sources from the same fixed-schema connector, such as Google Ads; not true for Google Sheets or BigQuery, where you unify the IDs yourself.

Controls are for what a viewer chooses. For what nobody should change, such as excluding test orders, use a filter property on the chart, page or report: only editors can set it, and it is always on (explore your data, October 2026). Charts can act as filters too: with cross-filtering, clicking a bar filters the rest of the page. It is on by default for most connectors and off for the ads connectors, where you turn it on per chart in the Chart interactions section of the properties panel (chart cross-filtering, April 2026).

Step 5. Check it in view mode

Click View. View mode shows the data and the controls the way a viewer will see them, and does not let anyone change the layout. Click Edit to go back.

Step 6. Share it, and schedule delivery

Click Share in the header, add people or groups, and set each one to Can view or Can edit. Before you send the link, check the data source: sharing the report does not share it, and what a viewer sees depends on the data source’s credentials.

The arrow next to Share opens the other options. Schedule delivery emails a PDF of the report on a schedule you set: choose recipients and pages, a start date and time, and a Repeat frequency, then click Schedule (schedule automatic report delivery). Get report link gives you a short link to the report (get a link to your report).

That is a working report. The rest of this guide is about what it is reading.

Reporting on more than one data source

One report can hold any number of data sources, each chart reading its own. Ad spend from Google Ads in one scorecard and sessions from Google Analytics in another is easy. Putting both in one chart, cost per session by campaign, needs a blend.

A blend joins up to five data sources, in Data Studio’s blend editor instead of SQL (how blends work, October 2026). You pick a table per data source, a join operator (inner, left outer, right outer, full outer or cross) and the fields that match, such as Campaign Name in both Google Ads and Google Analytics. Google’s help lists the limits, and each one shows up in real reports:

• Equality joins only. Field A = Field B works; a date that falls inside a range does not.
• Rows are grouped before the join. Each table is aggregated by the dimensions you include, so without a unique ID per record, identical rows collapse and the count comes out lower than the same join in SQL.
• Metrics stop being metrics. In the blend, a source’s metrics become unaggregated numbers, and the aggregation is decided again in the chart.
• A blend lives in one report. It cannot be reused in another report, and it takes its credentials and data freshness from the data sources underneath.

A blend is the right tool for a one-off question across two sources that share a clean key. For a report people will run every week, with spend from four ad platforms next to store orders, the join belongs before Data Studio, where it is written once. How that looks for GA4 conversions and ad cost is in blend GA4 conversions with ad spend.

Where a report on a raw connector stops

A connector brings the fields its source already has. Everything else, the report’s author defines: net sales after returns, conversion against the orders your store recorded rather than the ones an ad platform claims, ad spend next to store revenue. Each becomes a calculated field or a blend, kept in one chart or one data source. Two people answering the same question in two reports write two formulas, and nothing on either page says which one it used.

Here is how far apart two reasonable formulas land. The rows below are from the Devices view of the Digital marketing dashboard, built on our demo e-commerce store (synthetic data, 1 January to 21 October 2026, frozen copy taken 22 October 2026):

Put a scorecard for “conversion rate” above this table, and it can show either of two numbers:

• 2.38%, if the formula is total orders over total sessions: 5,790 / 243,299.
• 3.46%, if the scorecard averages the four rates in the last column, which is what a pre-computed rate column with an Average aggregation does.

The second number is 45% higher than the first, it is wrong, and it looks just as finished. In a report on a connector, the formula is whoever built the chart. In the next report, it is someone else.

When one connector is not enough: prepare the data in BigQuery

Some reports outgrow connectors and blends, because the join they need is wider than five tables, or the key is not clean, or the data is not in any connector:

• every paid social platform in one table: Facebook Ads, LinkedIn Ads and X Ads side by side;
• search ads from Google Ads and Microsoft Ads together;
• campaign performance measured against GA4 conversions or store orders;
• Search Console clicks next to the revenue those pages produced;
• attribution and cohort reports, which need user-level history no connector returns.

For these, put the data in a warehouse first, join it there, and point Data Studio at the result. Two things push in the same direction. The Google Analytics connector is sampled the way GA4 is sampled, and Data Studio does not tell you when it is (Google Analytics 4 sampling, October 2026); the daily GA4 export to BigQuery carries “all the raw, unsampled event data” (BigQuery Export, October 2026; see GA4 sampling and the GA4 to BigQuery guide). And outside Google’s own products, Google builds no connector for ad platforms, so their data needs a way in.

In OWOX Data Marts that way in is a Connector Data Mart: open-source connectors for Facebook Ads, TikTok Ads, LinkedIn Ads, X Ads, Microsoft Ads, Reddit Ads and Criteo, plus custom connectors described in a JSON manifest for any HTTP API. A connector loads the fields you pick into your own storage on a schedule you set with a Connector Trigger, and the data lands in your warehouse, not in OWOX (Sources, Connector-based Data Mart).

The join itself is a SQL Data Mart: a query over the tables already in your warehouse, with an Output Schema whose fields carry business names and descriptions (SQL-based Data Mart). Spend from every platform, unioned and matched to orders by campaign and date, becomes one Data Mart instead of one blend per report. The steps for the BigQuery side are in BigQuery reports in Data Studio.

The same question on a Data Model

The dashboard those rows came from is connected, not a snapshot. Claude built it through OWOX MCP for Claude and ChatGPT, it reads the demo e-commerce Data Model, twelve Data Marts and fifteen declared relationships published by an analyst, and it refreshes with every Report run. Conversion rate is defined once, in the Sessions Data Mart, and computed at whatever grain a view asks for: by device, by traffic type, by month. The page says so under the table.

What makes it connected rather than a snapshot:

• One definition, every view. The Funnel view reads Sessions, Orders, Purchases, Pageviews and Pages; the Devices view reads Sessions and Traffic Sources. Both take conversion from the same Data Mart, so the funnel and the device table cannot disagree.
• A Report, refreshed when it runs. On OWOX an AI dashboard is kept as a Report on Data Marts and refreshed every time the Report runs. The public page is a frozen copy taken 22 October 2026, because a public page cannot query a live project, and it says so at the bottom.
• Your AI builds it, not OWOX. Underneath is your Data Model: the Data Marts your analyst published, with their definitions and declared joins, from which OWOX builds every query.

The Net Sales Review does the same for revenue: its Summary reads net sales from the Purchases Data Mart and prints each month’s gross booked value beside it, so cancelled and returned orders stay visible instead of inflating the total. The full set is on the dashboards page, and how the AI side works is on AI dashboards.

Point Data Studio at the same Data Marts

You do not have to choose between the two. In OWOX Data Marts, Data Studio is a pull destination: you do not create a Report for it, you turn on Available in Data Studio on a Data Mart, and Data Studio asks OWOX for the data when someone opens or refreshes the report (Reports). Google Sheets and email are push destinations, delivered on a run or a schedule; Data Studio and Excel read the data themselves.

Setting it up, per the Data Studio destination docs:

1. In OWOX, create a Destination of type Data Studio and copy its JSON Config. It holds a secret key, so keep it private.
2. On the Data Mart’s Destinations tab, turn on the Data Studio row. The Data Mart must be Published.
3. In Data Studio, Step 2 of the tutorial changes: instead of a raw connector, find the OWOX Data Marts connector in the gallery, authorize it, paste the JSON Config, select the Data Mart and click Connect.

Everything after that is the tutorial you just did. What is different is what arrives in the data panel:

Three things to know before you rely on it. OWOX serves cached results for the Data Mart’s Cache Lifetime, 5 minutes by default and up to 12 hours, so Refresh data in Data Studio can return the cached rows until it expires. One request can return up to 1,000,000 rows and 50 MB; past that, use fewer Report Columns, filters or aggregation. And Data Studio still does the arithmetic on the page: a ratio like conversion rate is still a formula in the report, so write it once as a data source calculated field on the OWOX data source, total orders over total sessions, and have every chart use that field.

The Data Marts run on your warehouse, Google BigQuery, Snowflake, AWS Redshift, AWS Athena or Databricks, and the data stays there. The product side of the connection is on the Data Studio destination page.

Data Studio best practices that hold up

Most best-practice lists for this tool were written for an interface that has since moved. These hold in the current one, each checked against Google’s help in October 2026:

1. Write a shared formula in the data source, not the chart. A data source calculated field is available in every report on that data source; a chart-specific one is invisible to the next chart. Make the data source reusable, created from the home page, when more than one report needs the same field.
2. Use filter properties for rules, controls for choices. “Exclude test orders” is a filter property on the report, always on. “Show only France” is a drop-down list a viewer sets, and resets when the report is refreshed.
3. Set data freshness on purpose. A BigQuery data source checks for new data every 12 hours by default and can go as often as every minute; Google Ads, Search Console and the other Google ad and measurement products refresh every 12 hours, and that cannot be changed (manage data freshness). Shorter is fresher, and on BigQuery it also means more queries.
4. Check credentials before you share. Owner’s credentials let viewers see data they have no access to themselves; viewer’s credentials make each viewer authorize. Google’s advice is to share an owner’s-credentials report only with people you trust (about data sources).
5. Use conditional formatting where it applies. Font and background rules by value work on tables, pivot tables, scorecards and bar charts, under Style, Conditional formatting, Add formatting (conditional formatting rules). A red scorecard when clicks fall under a threshold is read faster than a number.
6. Turn cross-filtering on where people will click. It is off by default on the ads connectors, so a Google Ads chart will not filter its neighbours until you switch it on.
7. Do not trust GA4 numbers that might be sampled. Data Studio does not flag sampling. Compare the same date range in GA4, leave today out of the range, or report from the BigQuery export.

Two tips from older lists no longer hold. Explorer for quick, throwaway analysis was deprecated on 11 September 2025. And real-time widgets on Google Ads or Google Analytics data were never real time: those data sources refresh on the schedule in point 3.

What to do next

Build the tutorial report on Google’s sample data first; it takes the guesswork out of the menus. Then, for the numbers people will act on, ask one question before the next chart: who wrote this formula, and where else is it written? If the answer is “in this chart”, write it once, in a data source calculated field at least, and in a Data Mart if the same number also lives in a spreadsheet or an AI chat.

ReportingMarketing Dashboard Examples and Templates, by the Question They AnswerOctober 30, 2026→ReportingWhat is Data Visualization: Examples, and Tools for 2025March 20, 2025→ReportingWhat is Ad Hoc Reporting? How to Answer a One-Off QuestionOctober 26, 2026→

See all articles →