Cloud Digital Leader · Domain 2 of 6

Cloud Digital Leader Domain 2: Exploring Data Transformation with Google Cloud

Section 2 tests how data creates business value, how to pick the right Google Cloud storage or database product for a use case, and how Looker, BigQuery and streaming pipelines turn raw data into decisions.

Exam weight
~18%
Official objectives
3
Estimated questions estimate
9–11
Suggested study time estimate
7–10 hrs

About 13 min read Reviewed October 2026

Overview

Section 2 explains why data sits at the center of transformation and which Google Cloud products manage and analyze it. Objective 2.1 covers the value of data for real-time insight, trend spotting, strategy and AI, the difference between databases, warehouses and lakes, data types from first-party to semi-structured, the data supply chain, governance, and how openness reduces silos.

Objective 2.2 is product selection: Cloud Storage, Spanner, Cloud SQL, AlloyDB, Bigtable, BigQuery and Firestore, plus the Cloud Storage classes Standard, Nearline, Coldline, Archive and Autoclass, and paths for migrating or modernizing databases. Objective 2.3 covers Looker for self-service business intelligence, real-time streaming analytics, and pipeline products such as Pub/Sub, Dataflow and Managed Service for Apache Spark.

How it's tested

Most Section 2 items give a data type, an access pattern and a business need, then list four Google Cloud products. The winning answer fits all three. Watch for words such as global, relational, analytical, real time, rarely accessed or document, because each one rules products in or out quickly.

  • Pick the database

    Transactional, global, document or high-throughput key-value needs point to Cloud SQL, AlloyDB, Spanner, Firestore or Bigtable.

  • Pick the storage class

    Access frequency and retention time decide between Standard, Nearline, Coldline, Archive, or Autoclass when the pattern is unknown.

  • Analytics and streaming

    Dashboards, self-service BI or real-time event processing point to Looker, BigQuery, Pub/Sub or Dataflow.

Reasoning it through

A media company must keep raw video files for seven years for legal reasons and expects to read them less than once a year. Standard storage is built for frequent access and costs more to hold data. Nearline suits roughly monthly access, and Coldline roughly quarterly. Archive is the lowest-cost class for data accessed less than once a year, so it fits the retention need.

Objectives 3

Official objectives from the Google Cloud outline, in outline order.

2.1 Describe the intrinsic role that data plays in an organization's digital transformationData value, databases vs. warehouses vs. lakes, data types, data supply chain, governance, openness.
What it means
This objective covers why data matters to transformation and the concepts leaders use to talk about it. You explain how data supports real-time insights, trend identification, strategic decisions and AI. You differentiate databases, data warehouses and data lakes, recognize first-, second- and third-party data alongside structured, unstructured and semi-structured data, describe the data supply chain and governance, and explain how openness and interoperability reduce silos and lock-in.
What to understand
A database handles day-to-day transactions, a data warehouse stores cleaned, structured data for analysis, and a data lake holds raw data in many formats at low cost. First-party data comes from your own customers, second-party data is another organization's first-party data shared with you, and third-party data is bought or aggregated from outside. The data supply chain runs from genesis to collection, processing, storage, analysis and activation. Governance sets policies for quality, access, privacy and lifecycle so data can be trusted. Open formats and interoperable tools let organizations combine data across systems without being locked in.
How it shows up
Expect definition questions and scenarios that ask which data type, storage concept or supply chain stage is involved.
Common confusion
Thinking a data lake and a data warehouse are the same; the lake stores raw data, the warehouse stores structured data for analysis.
2.2 Determine which Google Cloud data management products are applicable to different business use casesCloud Storage, Spanner, Cloud SQL, AlloyDB, Bigtable, BigQuery, Firestore; storage classes; database migration.
What it means
This objective is about matching Google Cloud data products to needs. You choose among Cloud Storage, Spanner, Cloud SQL, AlloyDB, Bigtable, BigQuery and Firestore; define relational, non-relational, object storage, SQL and NoSQL; compare the Standard, Nearline, Coldline, Archive and Autoclass options by cost and access; and describe ways to migrate or modernize an existing database.
What to understand
Cloud Storage is object storage for files and unstructured data. Cloud SQL is managed MySQL, PostgreSQL and SQL Server for regional transactional apps. AlloyDB is a PostgreSQL-compatible database for demanding transactional and analytical workloads. Spanner is a relational database with global scale and strong consistency. Bigtable is a wide-column NoSQL store for very high throughput, such as IoT or time-series data. Firestore is a NoSQL document database for mobile and web apps. BigQuery is a serverless data warehouse for analytics. Nearline, Coldline and Archive trade lower storage cost for higher access cost and longer minimum storage; Autoclass moves objects between classes automatically. Database migration can be a straight move to a managed service or a modernization to a cloud-native database.
How it shows up
A use case with data shape, scale and access pattern is given, and you pick the single best product or storage class.
Common confusion
Choosing BigQuery for a transactional app; BigQuery is for analytics, not high-frequency row updates.
2.3 Discuss how smart analytics, business intelligence tools, and streaming analytics can add value in different business use casesLooker self-service BI, BigQuery with Looker, real-time streaming, Pub/Sub, Dataflow, Managed Service for Apache Spark.
What it means
This objective covers turning stored data into decisions. You describe how Looker democratizes data access, the value of visualizing BigQuery data in Looker for real-time reports, dashboards and workflow integration, why streaming analytics matters, and which products modernize data pipelines, including Pub/Sub, Dataflow and Managed Service for Apache Spark.
What to understand
Looker is a business intelligence platform with a semantic model so business users can explore governed data and build dashboards themselves. Connecting Looker to BigQuery lets reports reflect current data without copying it elsewhere. Streaming analytics processes events as they arrive, so a business can react in seconds, for example to fraud or stock changes, instead of waiting for nightly batches. Pub/Sub is the messaging service that ingests and distributes event streams. Dataflow is a managed service for batch and stream processing pipelines. Managed Service for Apache Spark runs Spark and related open-source workloads without managing clusters by hand.
How it shows up
Questions describe slow reports, IT bottlenecks or a need to act on live events and ask for the product or approach that helps.
Common confusion
Confusing Pub/Sub with Dataflow; Pub/Sub moves messages, Dataflow transforms and processes them.

High-yield concepts

The terms, rules and formulas this domain turns into questions most often.

  • Database vs. warehouse vs. lake

    Transactions; structured analytics; raw data in any format.

    Remember
    Each serves a different stage of the data journey.
    Tested as
    Definition and scenario questions.
    Trap
    Putting raw logs straight into a warehouse definition.

  • Data supply chain

    Genesis, collection, processing, storage, analysis, activation.

    Remember
    Activation is putting insight into action.
    Tested as
    New list in the 2026 guide.
    Trap
    Stopping at analysis and forgetting activation.

  • First-, second-, third-party data

    Own data; a partner's own data; purchased or aggregated outside data.

    Remember
    Second-party is shared directly by a partner.
    Tested as
    Data sourcing questions.
    Trap
    Calling partner-shared data third-party.

  • Structured vs. unstructured vs. semi-structured

    Rows and columns; images, audio, text; tagged formats like JSON.

    Remember
    Cloud and AI unlock value in unstructured data.
    Tested as
    Data type questions.
    Trap
    Treating JSON as fully unstructured.

  • Spanner

    Relational database with global scale and strong consistency.

    Remember
    Global, consistent transactions point to Spanner.
    Tested as
    Database selection.
    Trap
    Choosing Cloud SQL for global multi-region writes.

  • Cloud SQL and AlloyDB

    Managed MySQL, PostgreSQL, SQL Server; high-performance PostgreSQL-compatible database.

    Remember
    Both are relational and transactional.
    Tested as
    Lift-and-modernize database scenarios.
    Trap
    Picking Bigtable for relational data.

  • Bigtable

    Wide-column NoSQL for very large, high-throughput workloads.

    Remember
    IoT and time-series data are classic fits.
    Tested as
    High-volume ingestion scenarios.
    Trap
    Using it when SQL joins are required.

  • Firestore

    Serverless NoSQL document database for mobile and web apps.

    Remember
    Real-time sync for app data.
    Tested as
    App backend scenarios.
    Trap
    Choosing it for enterprise analytics.

  • BigQuery

    Serverless data warehouse and analytics engine.

    Remember
    No infrastructure to manage; analyze at scale with SQL.
    Tested as
    Analytics and BI scenarios.
    Trap
    Using it as an OLTP database.

  • Cloud Storage classes

    Standard, Nearline, Coldline, Archive by access frequency; Autoclass automates moves.

    Remember
    Colder classes cost less to store and more to read.
    Tested as
    Retention and archive scenarios.
    Trap
    Picking Archive for data read weekly.

  • Looker

    BI platform for governed self-service analytics.

    Remember
    Democratizes data access for business users.
    Tested as
    BI and dashboard scenarios.
    Trap
    Assuming it is a storage product.

  • Pub/Sub and Dataflow

    Event ingestion messaging; managed batch and stream processing.

    Remember
    Together they form a streaming pipeline.
    Tested as
    Real-time analytics scenarios.
    Trap
    Swapping their roles.

Side by side

Concepts candidates mix up, compared on the point that decides the answer.

Google Cloud databases

ConceptMeaningKey differenceCommon trap
Cloud SQL Managed MySQL, PostgreSQL, SQL Server Regional relational workloads, easy migration Picked for global strongly consistent writes
AlloyDB PostgreSQL-compatible high-performance database Faster transactional and analytical queries on PostgreSQL Confused with a NoSQL store
Spanner Globally scalable relational database Strong consistency across regions Overlooked for global financial ledgers
Bigtable Wide-column NoSQL Massive throughput, low latency, no SQL joins Used for relational reporting
Firestore NoSQL document database Mobile and web app data with real-time sync Picked for data warehousing

Cloud Storage classes

ConceptMeaningKey differenceCommon trap
Standard Hot data accessed often No minimum storage duration Used for long-term archives
Nearline Data read about once a month 30-day minimum storage Used for daily-access content
Coldline Data read about once a quarter 90-day minimum storage Confused with Nearline
Archive Data read less than once a year 365-day minimum, lowest storage cost Chosen for disaster recovery data needed often

Data pipeline products

ConceptMeaningKey differenceCommon trap
Pub/Sub Asynchronous messaging for events Ingests and distributes streams Expected to transform data
Dataflow Managed batch and stream processing Transforms and enriches data Expected to store messages
Managed Service for Apache Spark Managed Spark and open-source processing Fits existing Spark jobs Chosen when no Spark skills or code exist

Exam traps

Each trap is an answer that looks right. Learn the clue that should change your mind.

  1. Global relational

    What candidates pick
    Choose Cloud SQL for a ledger that needs consistent writes across continents.
    Why it looks right
    Cloud SQL is the familiar relational option.
    Why it's wrong
    Cloud SQL is regional; Spanner provides relational transactions with strong consistency at global scale.

    Clue to watch for Global users, multi-region writes, consistency.

  2. Warehouse as app database

    What candidates pick
    Choose BigQuery to back an order entry application.
    Why it looks right
    BigQuery is the flagship data product.
    Why it's wrong
    BigQuery is built for analytics; transactional apps need a database such as Cloud SQL, AlloyDB or Spanner.

    Clue to watch for Many small inserts and updates from an application.

  3. Archive for warm data

    What candidates pick
    Choose Archive to save money on backups restored every week.
    Why it looks right
    It has the lowest storage price.
    Why it's wrong
    Frequent reads make retrieval costs high and the 365-day minimum applies; Nearline or Standard fits better.

    Clue to watch for Stated access frequency.

  4. BI equals storage

    What candidates pick
    Choose Cloud Storage when a marketing team wants self-service dashboards.
    Why it looks right
    The data has to live somewhere.
    Why it's wrong
    The need is self-service analysis; Looker on BigQuery delivers governed dashboards to business users.

    Clue to watch for Business users, dashboards, waiting on IT.

  5. Batch for real time

    What candidates pick
    Recommend nightly batch reports for fraud detection.
    Why it looks right
    Batch reporting is well understood.
    Why it's wrong
    Fraud must be caught as events happen; streaming with Pub/Sub and Dataflow enables near real-time action.

    Clue to watch for Seconds matter, live events.

  6. Unknown access pattern

    What candidates pick
    Pick a single fixed storage class when access patterns are unpredictable.
    Why it looks right
    A fixed class seems simpler to budget.
    Why it's wrong
    Autoclass moves objects between classes automatically based on access, avoiding guesswork.

    Clue to watch for Unpredictable or changing access.

Answer framework

Work through these steps when a Section 2 question asks you to choose a data product.

  1. Decide whether the need is transactional, analytical, file storage, or real-time event processing, because that splits the product list immediately.
  2. For transactional needs, check scale and model: regional relational fits Cloud SQL or AlloyDB, global relational fits Spanner, document data fits Firestore, high-throughput key data fits Bigtable.
  3. For analytics, choose BigQuery for the warehouse and Looker when business users need governed self-service dashboards.
  4. For files and objects, choose Cloud Storage and pick the class by stated access frequency, or Autoclass when the pattern is unknown.
  5. For live events, combine Pub/Sub for ingestion with Dataflow for processing, and use Managed Service for Apache Spark when existing Spark jobs must move.

Practice question

A logistics company collects location pings from 200,000 delivery vehicles every few seconds and needs to store them with very low latency for later lookups by vehicle and time. The data does not require SQL joins. Which Google Cloud product is the best fit?

Want more like this? The Cloud Digital Leader practice test has questions for every domain.

Checklist

0 of 12

Tick each one you can do without notes. Your ticks stay in this browser.

Study strategy

Split this section into concepts and products. For concepts, write definitions for databases, warehouses and lakes, the data types, and the six stages of the data supply chain, then attach one business example to each. These items are new or expanded in the August 2026 guide and are easy marks once memorized.

For products, build a one-page selection table with columns for data model, scale, typical use case and the clue word that points to it. Include Cloud SQL, AlloyDB, Spanner, Bigtable, Firestore, BigQuery and Cloud Storage. Add a second table for storage classes with access frequency and minimum duration. Quiz yourself by covering the product column and reading only the clue.

Finish with analytics: sketch a simple pipeline from Pub/Sub to Dataflow to BigQuery to Looker and say what each step adds. Plan about 7 to 10 hours, and spend most practice time on product-selection scenarios, since they are the most common item type here.

FAQ

How much of the Cloud Digital Leader exam is Exploring Data Transformation with Google Cloud?

Exploring Data Transformation with Google Cloud is about 18% of the exam in the August 2026 guide, roughly 9 to 11 questions, up from about 16% in the previous version.

Which data products does the guide name?

Cloud Storage, Spanner, Cloud SQL, AlloyDB, Bigtable, BigQuery and Firestore for data management, Looker for business intelligence, and Pub/Sub, Dataflow and Managed Service for Apache Spark for pipelines.

Do I need to write SQL for this exam?

No. You should know what SQL and NoSQL mean and that BigQuery uses standard SQL, but the exam does not ask you to write queries.

What is Autoclass?

Autoclass is a Cloud Storage feature that automatically moves objects between storage classes based on how often they are accessed, which helps when access patterns are unpredictable.

What is new in this section compared with the previous guide?

AlloyDB, Autoclass and Managed Service for Apache Spark were added, along with data types, the data supply chain and openness. Cloud Spanner and Cloud Bigtable are now listed simply as Spanner and Bigtable.

How do I tell Spanner and Cloud SQL apart quickly?

Look for scale and geography. A regional app with a standard relational engine points to Cloud SQL. Global users needing consistent transactions across regions point to Spanner.

Is BigQuery multicloud?

Yes. BigQuery Omni lets organizations analyze data stored in other clouds, and the guide lists it in Section 4 among products that run in hybrid and multicloud environments.

Other Cloud Computing certifications

Where the same subject shows up in other exams. Cards marked "Same subject" open the matching domain.

Test your Domain 2 knowledge

See whether you can pick the right database, storage class and analytics product from short business scenarios.

Take the Cloud Digital Leader practice test Check your Cloud Digital Leader readiness

Keep going