Certified Data Architect — Free Practice Questions
10 free sample questions from a bank of 83, with the correct answers and explanations. No signup required — start practising right now.
1Universal Containers (UC) is replacing a home grown CRM solution with Salesforce. UC has decided to migrate operational (open and active) records to Salesforce, while keeping historical records in legacy system. UC would like historical records to be available in Salesforce on an as needed basis.
Which solution should a data architect recommend to meet business requirement?
Leverage real-time integration to pull records into Salesforce.
Build a swivel chair solution to go into the legacy system and display records.
Leverage mashup to display historical records in Salesforce.
Bring all data in Salesforce, and delete it after a year.
Answer: C
The short version
C — Mashup shows legacy history on demand. Operational records migrate to Salesforce while historical records stay in legacy and appear in Salesforce only as needed.
Key concepts in this question
Mashup: UI-level display combining Salesforce data with external data without storing it in Salesforce
Operational vs historical migration: Move only open/active records, leave history in place for on-demand access
Why C is correct
Mashup is the Salesforce-recommended pattern for displaying Salesforce-hosted and externally hosted data together without replication. It avoids storage cost and migration effort for history while satisfying the as-needed visibility requirement. Real-time data replication is unnecessary when users only need occasional reference views.
Why the others are wrong
A. Real-time integration to pull and store records replicates data and adds storage, sync, and consistency overhead contrary to the keep-history-in-legacy goal.
B. Swivel-chair requiring users to log into the legacy system is poor UX and not an architected Salesforce solution.
D. Bringing all data in then deleting after a year contradicts keeping history in legacy and still incurs full migration plus retention risk.
Certified-Data-Architect exam tip
History stays out + viewed as needed = mashup, not migration.
2Northern Trail Outfitters (NTO) has been using Salesforce for Sales and Service for 10 years. For the past two years, the marketing group has noticed a rise from 0% to 35% in returned mail when sending mail using the contact information stored in Salesforce.
Which solution should the data architect use to reduce the amount of returned mail?
Email all customers and ask them to verify their information and to call NTO if their address is incorrect.
Delete contacts when the mail is returned to save postal costs for NTO.
Have the sales team call all existing customers and ask to verify the contact details.
Use a third-party data source to update the contact information in Salesforce.
Answer: D
The short version
Answer D cleanses stale mailing addresses with a trusted third-party data source.
Key concepts
Contact data quality decays over time
Third-party address validation and enrichment
Returned mail cost and deliverability
Why
NTO return rate rose from 0 percent to 35 percent over years in Salesforce. Manual verification by email or sales calls does not scale and does not fix systemic decay. A third-party source updates addresses in bulk, restoring deliverability. That is why D is the settled answer.
Why the others are wrong
A asks customers to self-verify by email, low response and slow
B deletes contacts on bounce, destroys history and does not fix root cause
C uses sales calls for all customers, unscalable and costly
exam tip
For decaying addresses, think enrichment vendor first, manual outreach last.
3Universal Containers (UC) is migrating data from legacy system to Salesforce. During data analysis it was discovered that data types of fields being migrated do not match with Salesforce data types.
Which solution should a data architect use to ensure successful data migrations?
Migrate legacy data to a staging database for mapping then leverage an ETL tool to transform the data and load into Salesforce.
Migrate the legacy data leveraging an ETL tool to transform data types and load data into Salesforce.
Export legacy data into CSV files and leverage data loader to load data into Salesforce.
Export legacy data into the staging database and leverage stored procedures to transform data types before loading into Salesforce.
Answer: B
The short version
Answer B uses an ETL tool to transform legacy types and load into Salesforce.
Key concepts
Data type mapping legacy to Salesforce
Extract transform load
Successful migration with validation
Why
Fields do not match Salesforce types, so direct CSV or Data Loader load fails or corrupts data. An ETL tool transforms types, handles errors, and loads efficiently. That is why B is the settled answer for type mismatch.
Why the others are wrong
A adds staging plus ETL, more robust but heavier than needed for type transform alone
C CSV plus Data Loader has no transform, so mismatches fail
D staging plus stored procedures adds custom code without ETL loading benefits
exam tip
Type mismatch equals transform first, ETL is the Data Architect default.
4Universal Containers has been a customer of Salesforce for 10 years. Currently they have 2 million accounts in the system. Due to an erroneous integration built 3 years ago, it is estimated there are 500,000 duplicates in the system.
Which solution should a data architect recommend to remediate the duplication issue?
Extract the data using data loader and use excel to merge the duplicate records
Utilize a data warehouse as the system of truth
Develop an ETL process that utilizes the merge API to merge the duplicate records
Implement duplicate rules
Answer: C
The short version
C — Merge 500k duplicates with an ETL merge process. Existing duplicates need remediation by merging, not just preventing future duplicates.
Key concepts in this question
Merge: Preserves one master record, re-parents children, deletes losers
Remediation vs prevention: Cleanup of existing dupes vs duplicate rules/jobs for new records
Why C is correct
With 500,000 existing duplicates across 2 million accounts, a programmatic ETL process using merge can match, survive, and re-parent at scale. Duplicate rules alone only warn or block new duplicates and do not fix the backlog. A warehouse as system of truth does not itself clean Salesforce records.
Why the others are wrong
A. Export to Data Loader plus Excel manual merging is infeasible and error-prone for 500k records.
B. Declaring a warehouse the system of truth does not remediate duplicates already in Salesforce.
D. Duplicate rules prevent future duplicates but do not merge the existing 500k duplicate set.
Certified-Data-Architect exam tip
Backlog of dupes = merge them; rules = stop new ones.
5Northern Trail Outfitters (NTO) runs its entire business out of an enterprise data warehouse (EDW). NTO’s sales team is starting to use Salesforce after a recent implementation, but currently lacks the data required to advance leads and opportunities to the next stage.
NTO’s management has researched Salesforce Connect and would like to use it to virtualize and report on data from the EOW within Salesforce. NTO will be running thousands of reports per day across 10 to 15 external objects.
What should a data architect consider before implementing Salesforce Connect for reporting?
Maximum page size for server-driven paging
Maximum external objects per org
OData callout limits per day
Maximum number of records returned
Answer: C
The short version
Answer C checks OData callout limits before reporting with Salesforce Connect.
Key concepts
Salesforce Connect virtualization
External objects and OData callouts
Daily limits under high report volume
Why
NTO plans thousands of reports per day across 10 to 15 external objects. Each report triggers OData callouts to the EDW. Daily callout limits throttle this pattern, so the architect must size limits first. That is why C is settled.
Why the others are wrong
A page size matters for paging but is not the daily volume blocker
B external object count is a packaging limit, not the reporting bottleneck
D max records returned is per-request, not the aggregate daily risk
exam tip
Connect plus high report volume always check callout limits first.
6Universal Containers (UC) has implemented a master data management strategy, which uses a central system of truth, to ensure the entire company has the same customer information in all systems. UC customer data changes need to be accurate at all times in all of the systems. Salesforce is the identified system of record for this information.
What is the correct solution for ensuring all systems using customer data are kept up to date?
Send customer record changes from Salesforce to the system of truth in real time.
Send customer data nightly to the system of truth in a scheduled batch job.
Have each system pull the record changes from Salesforce using change data capture.
Send customer record changes from Salesforce to each system in a nightly batch job.
Answer: C
The short version
Answer C lets each system pull changes from Salesforce with Change Data Capture.
Key concepts
Master data management with Salesforce as system of record
Change Data Capture eventing
Real-time sync to all consumers
Why
Salesforce is the record for customer data that must be accurate everywhere at all times. Push batch jobs lag and point-to-point pushes fan out complexity. CDC lets every system subscribe and pull changes reliably. That is why C is settled.
Why the others are wrong
A pushes to system of truth only, does not update all consumers
B nightly batch to truth is stale and one-directional
D nightly batch to each system is stale and brittle
exam tip
System of record plus always accurate everywhere points to CDC subscribe.
7The data architect for Universal Containers has written a SOQL query that will return all records from the Task object that do not have a value in the whatID field:
SELECT ID, Description, subject FROM Task WHERE WhatID != NULL
When the data architect uses the query to select values for a process, a time-out error occurs.
What does the data architect need to change to make this query more performant?
Change the WHERE clause to filter by a deterministic defined value.
Change query to SOSL.
Remove description from the requested field set.
Add LIMIT 100 to the query.
Answer: A
The short version
Choose 1 — answer A. The timeout comes from a non-selective filter (WHERE WhatID != NULL forces a full table scan); the documented fix is a selective filter on an indexed deterministic field, not trimming the SELECT list.
Key concepts in this question
A query is selective only when a WHERE filter on an indexed field keeps matched rows under the selectivity threshold.
Negation operators like != (and NOT LIKE, leading-% LIKE) are documented index-selectivity exceptions: they make the filter non-selective.
Non-selective queries against large objects time out or are terminated; the fix must change the filter, not the field list.
Why A is correct
A changes the WHERE clause to filter by a deterministic defined value — i.e., a selective, index-friendly equality filter instead of the != NULL negation. Per the selectivity docs, a query is selective when a filter on an indexed field reduces rows below threshold, and deterministic (indexable) values are exactly what the optimizer can use; this removes the full table scan that caused the time-out.
Why the others are wrong
B (change query to SOSL): SOSL is full-text search across objects, not a fix for a non-selective structured filter; it cannot express this selective-row retrieval and does not cure the table scan.
C (remove Description from the field set): only shrinks payload per row; the time-out is caused by scanning the table under a non-selective != NULL filter, so dropping one long-text field does not change selectivity or prevent the time-out.
D (add LIMIT 100): caps rows returned but the engine must still find/scan under the same non-selective filter; LIMIT does not make the query selective and does not fix the underlying full-scan time-out.
Certified-Data-Architect exam tip
Query time-out + != / NOT IN / NOT LIKE / leading-% LIKE in the WHERE clause = selectivity problem. The fix is always a selective deterministic indexed filter — never trimming SELECT fields, switching to SOSL, or adding LIMIT.
8Northern Trail Outfitters is planning to build a consent form to record customer authorization for marketing purposes.
What should a data architect recommend to fulfill this requirement?
Use AppExchange solution to address the requirement.
Utilize the Authorization Form Consent object to capture the consent.
Use custom fields to capture the authorization details.
Create a custom object to maintain the authorization.
Answer: B
The short version
Answer B uses the Authorization Form Consent object for marketing consent.
Key concepts
Consent management standard objects
Authorization Form Consent
Auditable marketing authorization
Why
Salesforce provides Authorization Form Consent to capture who consented, for what, and when. Custom fields or custom objects lose standard compliance behavior. AppExchange is unnecessary. That is why B is settled.
Why the others are wrong
A AppExchange adds dependency for native capability
C custom fields are not auditable consent records
D custom object rebuilds what the platform already provides
exam tip
Consent requirement equals standard consent objects before custom build.
9Northern Trail Outfitters (NTO) plans to maintain contact preferences for customers and employees. NTO has implemented the following:
1. Customers are Person Accounts for their retail business.
2. Customers are represented as Contacts for their commercial business.
3. Employees are maintained as Users.
4. Prospects are maintained as Leads.
NTO needs to implement a standard communication preference management model for Person Accounts, Contacts, Users, and Leads.
Which option should the data architect recommend NTO to satisfy this requirement?
Create case for contact preferences, and use this to validate the preferences for Lead, Person Accounts, and Users.
Use Individual objects to maintain the preferences with relationships to Lead, Person Account, and Users.
Create custom fields for contact preferences in Lead, Person Account, and Users objects.
Create a custom object to maintain preferences and build relationships to Lead, Person Account, and Users.
Answer: B
The short version
Choose 1 — answer B. NTO needs one standard preference model spanning Person Accounts, Contacts, Users, and Leads, and Salesforce's standard answer is the Individual object / Consent Data Model, not a custom build.
Key concepts in this question
The Individual object is Salesforce's standard record for a person's privacy and communication preferences.
The Salesforce Consent Data Model ties consent to the Individual and relates to leads, contacts, person accounts, and users.
Standard model vs. custom object/custom fields: the question explicitly asks for a standard communication preference management model.
Why B is correct
B recommends the Individual object with relationships to Lead, Person Account, and Users. That is exactly the documented standard: the Individual object manages preferences for a person across multiple objects (lead-turned-contact/person account, plus a user record), and the Consent Data Model explicitly covers leads, users, person accounts, and contacts. It gives one standard model instead of fragmented per-object fields.
Why the others are wrong
A (Case for preferences): wrong object and pattern. Cases track service issues/slos, not a person's standing communication preferences; validating preferences off a Case is not the documented standard model.
C (custom fields on each object): fragments preferences across Lead, Person Account, and User with no unified person-level view; it is custom, duplicated effort, and the opposite of the requested standard model.
D (custom preference object with relationships): rebuilds as custom what Salesforce already provides as standard (Individual + consent objects); more maintenance, no standard consent semantics, and not what the docs recommend.
Certified-Data-Architect exam tip
When a Data Architect question says standard preference/consent model across leads, contacts, person accounts, and users, pick the Individual object / Consent Data Model — never a custom object or per-object custom fields.
10A national nonprofit organization is using Salesforce to recruit members. The recruitment process requires a member to be matched with a volunteer opportunity. Given the following:
1. A record is created in Project__c and used to track the project through completion.
2. The member may then start volunteering and is required to track their volunteer hours, which is stored in VTOTime__c object related to the project.
3. Ability to view or edit the VTOTime__c object needs to be the same as the Project__c record.
4. Managers must see total hours volunteered while viewing the Project__c record.
Which data relationship should the data architect use to support this requirement when creating the custom VTOTime__c object?
Master Detail field on Project__c to VTOTime__c showing a list of VTOTime__c Records in a related list.
Lookup field on VTOTime__c to Project__c with formula filed on Project__c showing Sum of hours from VTOTime__c records.
Lookup Field on Project__c to VTOTime__C displaying a list of VTOTime__C in a related list.
Master Detail Field on VTOTime__c to Project__c with rollup summary field on Project__c showing sum of hours from VTOTime__c records.
Answer: D
The short version
Choose 1 — answer D. Same view/edit access as the parent plus a SUM rollup means one thing: a master-detail field on the child (VTOTime__c) pointing to the parent (Project__c), with a roll-up summary on Project__c.
Key concepts in this question
In master-detail, the detail (child) inherits sharing/security from the master; detail has no independent sharing.
The relationship field always lives on the detail (child) object, pointing at the master.
Native roll-up summary fields (SUM/COUNT/MIN/MAX) only work from a detail to its master over a master-detail relationship — never over lookup, never via formula.
Why D is correct
D puts the Master-Detail field on VTOTime__c (the child) related to Project__c (the master) and adds a roll-up summary on Project__c summing hours. The documented inheritance gives VTOTime__c exactly Project__c's view/edit access, and the roll-up gives managers total hours on the Project__c record. Direction, sharing, and rollup all match.
Why the others are wrong
A (Master-Detail field on Project__c to VTOTime__c): backwards. The relationship field must live on the detail/child; putting it on the parent reverses master and detail and cannot produce the required child-inherits-parent sharing plus parent rollup.
B (lookup on VTOTime__c + formula on Project__c for the SUM): fails twice — lookup children keep independent sharing (violates same-access rule), and a formula field cannot aggregate child records; native SUM rollups require master-detail.
C (lookup on Project__c to VTOTime__c with related list): wrong direction and wrong type — parent-side lookup to the child, no sharing inheritance, no rollup, so neither requirement 3 nor requirement 4 is met.
Certified-Data-Architect exam tip
Same-sharing-as-parent + rollup on the parent = master-detail field on the CHILD pointing to the parent, roll-up summary on the PARENT. Any option with the field on the parent or a lookup/formula combo is automatically wrong.