Free Salesforce Data 360 Consultant practice questions

Question 1

Ursa Major Solar's leadership asks the Data 360 consultant which business problems are strong candidates for Data 360 real-time capabilities. Which two problems should the consultant identify? Choose 2 answers.
A. The finance team runs a revenue report once per quarter for the board.
B. Every website visitor sees the same content, and marketing wants pages to change based on the visitor's interactions in the current session.
C. The legal team archives seven-year-old installation contracts once per year.
D. In-store consultants and mobile field technicians see different, inconsistent versions of the same customer's profile.
E. The marketing team sends one identical monthly newsletter to all subscribers.

Accepted Answer

Correct answer: B. Every website visitor sees the same content, and marketing wants pages to change based on the visitor's interactions in the current session. | D. In-store consultants and mobile field technicians see different, inconsistent versions of the same customer's profile.
Real-time Data 360 enables instant access to profiles so websites can dynamically modify content based on real-time interactions, and it resolves identity across channels so every touchpoint, from in-store reps to mobile service agents, sees a consistent profile with the latest information. A quarterly board report has no need for sub-second processing. An annual archiving process is a batch, compliance-driven task. A single identical newsletter involves no personalization or timing that would benefit from real-time data.

Question 2

Cumulus Financial's marketing team wants to use customers' ZIP codes as the primary attribute for deciding who receives premium credit card offers in Data 360. The consultant is concerned about ethical risk. Which two recommendations should the consultant make? Choose 2 answers.
A. Evaluate whether ZIP code acts as a proxy for a protected class, because in the US ZIP codes are often highly correlated with race.
B. Collect customers' race explicitly so the segment can be adjusted after it is built.
C. Give every marketer edit access to the segment so more people can review it for bias.
D. Favor intent- and behavior-based attributes, such as product interest and browsing activity, rather than relying only on demographics.
E. Purchase third-party demographic data to enrich every profile before deciding which attributes to use.

Accepted Answer

Correct answer: A. Evaluate whether ZIP code acts as a proxy for a protected class, because in the US ZIP codes are often highly correlated with race. | D. Favor intent- and behavior-based attributes, such as product interest and browsing activity, rather than relying only on demographics.
Treating sensitive data carefully means considering which attributes might be proxies for protected classes, and the guidance notes that ZIP codes in the US are often highly correlated with race and can create bias. It also recommends segmenting audiences on intent or behavioral attributes rather than only demographics. Explicitly collecting race adds sensitive data without a validated purpose and increases the risk of bias. Broad edit access violates the right of least privilege, under which only people who truly need data should have it and, by default, should read rather than change it. Buying more demographic data runs against collecting only what you need and choosing third-party partners carefully.

Question 3

Universal Containers (UC) plans to automate audience segmentation in Data 360. Marketing estimates the automation will save 5 hours per campaign cycle, UC runs 24 campaign cycles per year, the blended hourly rate is $60, and the one-time setup will take 40 hours at the same rate. Using the ROI approach recommended for justifying a Data 360 investment, what annual savings should the consultant report?
A. $7,200
B. $4,800
C. $9,600
D. $2,400

Accepted Answer

Correct answer: B. $4,800
The recommended calculation is ROI = (time saved x hourly rate x campaign cycles) minus setup cost, so (5 x $60 x 24) - (40 x $60) = $7,200 - $2,400 = $4,800 in annual savings. $7,200 is the gross time savings and ignores the setup cost that must be subtracted. $9,600 incorrectly adds the setup cost to the savings instead of subtracting it. $2,400 is only the setup cost itself, not the net savings.

Question 4

Cloud Kicks ingests about 13,000 new order rows per day into a DLO that already holds about 7 million rows. The data stream's Refresh History shows roughly 13,000 records added, but the DLO's Refresh History shows about 7 million records removed and 7 million added in the same window. Finance asks which numbers reflect billable processing. What should the consultant explain?
A. The DLO numbers drive billing, so the stream is reprocessing the full table and should use incremental mode
B. The data stream numbers reflect the business records processed; the DLO numbers reflect storage file rewrites
C. The DLO numbers reveal duplicate ingestion, so the consultant should open a case with Customer Support
D. Both sets of numbers are billed and added together, so the team should reduce the ingestion frequency

Accepted Answer

Correct answer: B. The data stream numbers reflect the business records processed; the DLO numbers reflect storage file rewrites
Data stream Refresh History is a business logic view of records processed from the source, while DLO Refresh History is a storage view of file-level operations. Data 360 uses Apache Iceberg, which often rewrites entire Parquet files, so committing a small change can show millions of records removed and added at the DLO level. Salesforce Help says to always use the data stream Refresh History metrics for billing and credit consumption. The DLO figures aren't evidence of full reprocessing or duplicate ingestion, and they aren't billed in addition to the data stream figures.

Question 5

Ursa Major Solar created a record-level security policy so that each of its four regional marketing teams sees only customers in its own region. In Data Explorer, a West region marketer correctly sees only West customers. However, when the same marketer builds a segment on the Unified Individual DMO, the segment population includes customers from every region. What explains this behavior?
A. The policy was scoped to the default data space, and segments always run in a separate system data space
B. Record-level policies aren't enforced when segments are created, because segmentation runs in system context
C. Record-level policies take 24 hours to propagate from Data Explorer to the segmentation engine
D. The marketer's Customize feature permission for Segmentation overrides record-level access policies

Accepted Answer

Correct answer: B. Record-level policies aren't enforced when segments are created, because segmentation runs in system context
Salesforce Help states that RLS policies aren't enforced when creating segments and calculated insights because these operations run in system context. If regional marketers must segment only their own customers, the consultant should partition the data another way, for example by associating filtered DLOs with regional data spaces. Segments aren't run in a separate system data space. Policies take effect within a few minutes of activation, not 24 hours. Feature permissions control whether a user can view or customize a feature, and they aren't the reason the policy doesn't filter segment populations.

Question 6

Universal Containers accidentally loaded a test file of 1,200 dummy customer profiles through an Amazon S3 data stream that uses the Upsert refresh mode. Some dummy profiles also have related engagement records in another DLO. The team wants all of the dummy data removed. Which three statements should the consultant share? Choose 3 answers.
A. Place a delete file in CSV or Parquet format that identifies the records in the same source directory the data stream reads from.
B. Switch the data stream to Incremental mode and add a Delete Record Flag column so the dummy records are removed on the next run.
C. Keep the data stream in the Upsert refresh mode, because deleting ingested records from an S3 data stream requires upsert.
D. Disable the primary key field on the data stream so that the dummy records are excluded from the DLO on the next refresh.
E. Delete the related engagement records separately, because deleting a profile record doesn't cascade to its related records.

Accepted Answer

Correct answer: A. Place a delete file in CSV or Parquet format that identifies the records in the same source directory the data stream reads from. | C. Keep the data stream in the Upsert refresh mode, because deleting ingested records from an S3 data stream requires upsert. | E. Delete the related engagement records separately, because deleting a profile record doesn't cascade to its related records.
To delete ingested records, the data stream must use the upsert refresh mode; for Amazon S3, Google Cloud Storage, and Azure connectors you create a delete file in CSV or Parquet format, and the records being deleted must be read from the same source directory the data stream ingests from. Data 360 doesn't support cascading deletes, so related engagement or profile records must be deleted from each table. The Delete Record Flag approach applies to incrementally refreshed, non-file-based data streams. Fields used as the primary key can't be disabled.

Question 7

Universal Containers' admin is creating a Snowflake data share target in Data 360 so Snowflake analysts can query shared profile objects. Saving the target fails with an authentication error, even though the Snowflake credentials were confirmed correct. The Snowflake account restricts access with a network policy. What should the consultant check first?
A. That the data share is already linked to a target, because a Snowflake target can be saved only for a linked share.
B. That Snowflake's network policy includes the Data 360 IP addresses, and the admin's own IP address, in its allowed IP list.
C. That the Snowflake analysts have accepted the share and created a database before the target is saved in Data 360.
D. That the legacy Snowflake connector is used, because Snowflake V2 targets don't support accounts with network policies.

Accepted Answer

Correct answer: B. That Snowflake's network policy includes the Data 360 IP addresses, and the admin's own IP address, in its allowed IP list.
Salesforce's data share troubleshooting guidance says an authentication error when setting up a Snowflake data share target is resolved by confirming that the Data Cloud IP addresses, and the local machine's IP address, are in Snowflake's list of allowed IP addresses through an IP networking policy. A target is created first and then linked to a data share, so linking can't be a prerequisite for saving it. Accepting the share and creating a database in Snowflake happens after the share is linked to the target. Salesforce directs customers to migrate from legacy Snowflake targets to Snowflake V2, not back to the legacy connector.

Question 8

Cumulus Financial streams card authorization events into Data 360 through the Ingestion API at about 40,000 events per hour. The fraud team wants only card-not-present authorizations, with merchant country codes standardized, written to a separate object that is mapped to the data model within minutes of each event arriving. What should the Data 360 consultant recommend?
A. A batch data transform with Filter and Transform nodes that is scheduled to run once each day and writes the cleaned authorizations to a target DLO.
B. A streaming data transform whose SQL statement uses a WHERE clause and CASE WHEN logic to write the cleaned authorizations to a target DLO.
C. A calculated insight that filters the authorization DLO to card-not-present events and standardizes the country codes as a dimension on a schedule.
D. An identity resolution ruleset with normalized match rules that standardizes the country codes and removes other events before they reach the DLO.

Accepted Answer

Correct answer: B. A streaming data transform whose SQL statement uses a WHERE clause and CASE WHEN logic to write the cleaned authorizations to a target DLO.
A streaming data transform runs continuously, picking up new or changed records from a source object, reshaping them with a SQL statement, and writing them to a different target object that can then be mapped to the data model; a WHERE clause limits which source records it processes. A batch data transform runs on a schedule, so a daily run can't meet a minutes-level requirement. A calculated insight produces aggregated metrics built from dimensions and measures, not a cleaned record-level copy of each event. Identity resolution matches and reconciles profiles from mapped data; it doesn't filter or clean engagement records before they land in a DLO.

Question 9

Ursa Major Solar wants its data graph, whose primary DMO is Unified Individual, to include a calculated insight for each customer's total installation spend. The insight aggregates spend with the source Individual DMO's Individual Id as its dimension, and the source Individual DMO isn't part of the data graph. The consultant can't add the insight to the graph. What should the consultant change?
A. Rebuild the insight so the Unified Individual primary key is one of its dimensions, and then add it at the root of the data graph.
B. Add the calculated insight as a child node under the Unified Link Individual object, because insights can be attached only below link objects.
C. Re-create the data graph as a real-time data graph, because only real-time data graphs can include calculated insights and streaming insights.
D. Write the insight's output to a custom DMO through a data stream, because data graphs can't reference calculated insight objects directly.

Accepted Answer

Correct answer: A. Rebuild the insight so the Unified Individual primary key is one of its dimensions, and then add it at the root of the data graph.
A calculated insight can be included in a data graph only when it's based on a DMO that's also in the graph, the DMO's primary key field must be configured as a dimension on the calculated insight object, and the insight can be added only at the root of the graph. Rebuilding the insight with the Unified Individual primary key as a dimension satisfies these conditions for this graph. Calculated insight objects can't be added as child nodes under another object. Calculated and streaming insights aren't limited to real-time data graphs, and data graphs reference calculated insight objects directly, so no custom DMO is needed.

Question 10

Cumulus Financial maps rewards card numbers from two sources to the Party Identification data model object (DMO). Both data streams map the Party Identification Type as Rewards, but the mobile banking stream maps the Identification Name as 'CF Rewards' while the card processor stream maps 'Rewards Card'. After identity resolution runs, profiles that share a rewards card number across the two sources still aren't matched. What should the consultant fix?
A. Map the same Identification Name value in both streams, because identification numbers are matched only when the type and name values also match.
B. Map the rewards card numbers to Individual Id instead, because Party Identification records are ignored unless an email address is also mapped.
C. Map the Party field in both streams to the rewards card number, because Party must hold the identifier value that is being matched.
D. Configure key qualifiers on both data lake objects, because Party Identification records from two sources aren't compared without them.

Accepted Answer

Correct answer: A. Map the same Identification Name value in both streams, because identification numbers are matched only when the type and name values also match.
Data 360 matches Party Identification records that have the same Identification Number only when they also share the same Party Identification Type and Identification Name values, so the two streams must map a consistent name, for example with the same constant value. Party Identification works as an identifier in its own right and doesn't depend on an email address being mapped. The Party field is a foreign key that references Individual Id (Party = Individual.Id), so it must hold the individual's key, not the card number. Key qualifiers prevent key collisions when sources are harmonized into one DMO, but they aren't required for party identifiers to match.

Question 11

Cumulus Financial plans a household ruleset to group co-applicants for cross-sell campaigns. The design team proposes three match rules in the household ruleset: one on last name and address, one on a shared loan application number stored as a party identifier, and one on a shared phone number. Some customers own both a primary residence and a vacation home. Which two points should the consultant raise? Choose 2 answers.
A. All three match rules can be used, but each household can contain only one individual from each source system.
B. Household rulesets must be based on the Individual DMO directly, so the individual ruleset's output can't be reused.
C. A household ruleset supports only one match rule, so the team must choose last name and address or the party identifier.
D. Each individual can belong to only one household, so owners of two homes are assigned to the most recent address.
E. An individual can be matched into more than one household, so an owner with two street addresses can join two households.

Accepted Answer

Correct answer: C. A household ruleset supports only one match rule, so the team must choose last name and address or the party identifier. | E. An individual can be matched into more than one household, so an owner with two street addresses can join two households.
Only one match rule can be added to a household ruleset, and Salesforce recommends basing it on either last name and address or a party identifier such as a loan application number, so the team must pick one. Individuals can be matched into multiple households, for example one household per street address, so vacation-home owners can appear in two households. Nothing limits a household to one individual per source system; each unified household simply contains one or more individuals. Household rulesets are built on the Unified Individual output of an individual ruleset, not directly on the Individual DMO.

Question 12

Cumulus Financial wants an AI Models model that predicts which of three outcomes each renewing policyholder will choose: Renew As-Is, Renew with Changes, or Cancel. No model for this has been trained anywhere else. Given AI Models' supported model types, which approach should the consultant recommend?
A. Build a single native model using AI Models' multi-class classification type, since it directly supports three or more outcome labels
B. Use a single regression model to output one continuous score and have downstream users manually bucket it into the three outcomes
C. Decompose it into two chained binary classification models, since AI Models supports only regression and binary classification
D. Recommend against using AI Models entirely, since Data 360 has no way to approximate a prediction with more than two outcomes

Accepted Answer

Correct answer: C. Decompose it into two chained binary classification models, since AI Models supports only regression and binary classification
AI Models supports only two model types, regression and binary classification, so a three-outcome prediction has to be decomposed into a sequence of binary classification models — such as Renews vs. Cancels, and then As-Is vs. with Changes among the renewals — to stay within what is supported. There is no native multi-class classification type in AI Models, so building a single model with three or more outcome labels directly is not possible. A regression model predicts a continuous numeric value; it is not the documented approach for a categorical, multi-outcome business decision, and manually bucketing a score is not a supported or recommended AI Models pattern. Because decomposing the problem into binary classification models is a workable path within AI Models' supported types, recommending against using AI Models entirely goes further than necessary.

Question 13

Northern Trail Outfitters wants to define and calculate customer lifetime value, most-viewed product categories, and CSAT scores from its entire unified digital estate, spanning Sales, Service, and Commerce data, at the profile, segment, and population level. What should the consultant recommend to build these multidimensional metrics?
A. Build a calculated insight that aggregates the metrics at the profile, segment, or population level.
B. Build a segmentation container that filters unified individuals by each metric's threshold value.
C. Create a real-time data graph that recalculates the metrics for every incoming event.
D. Configure a data action that fires whenever a data model object field changes, since data actions can perform their own aggregate calculations without a separate insight.

Accepted Answer

Correct answer: A. Build a calculated insight that aggregates the metrics at the profile, segment, or population level.
Calculated insights are purpose-built to define and calculate multidimensional metrics, such as lifetime value, most-viewed categories, and CSAT, at the profile, segment, and population level across a unified data set. A segmentation container filters unified individuals using existing attributes and insights to build an audience; it doesn't define new aggregate metrics itself. Real-time data graphs power real-time insights that recalculate on every event in milliseconds, which fits low-latency personalization, not this kind of large-scale historical metric. A data action triggers a downstream response when data changes; it doesn't calculate metrics.

Question 14

An external inventory-planning system at Northern Trail Outfitters needs to run custom SQL queries against Data 360 objects over REST, from outside Salesforce. What should the consultant recommend the integration use?
A. The Data 360 Connect REST API's SQL Query APIs, to submit and retrieve results for custom SQL over REST.
B. Insights Builder, since it can expose calculated insights as a REST endpoint automatically.
C. A standard Salesforce CRM connection, since it lets external systems query Data 360 directly, the same way a companion connection would expose query access.
D. Data Explorer, embedded into the external system's UI.

Accepted Answer

Correct answer: A. The Data 360 Connect REST API's SQL Query APIs, to submit and retrieve results for custom SQL over REST.
For custom SQL queries executed as REST calls from an external system, the documented path is the Data 360 Connect API's SQL Query APIs. Insights Builder authors insights inside Data 360; it doesn't expose a generic external SQL REST channel. A standard Salesforce CRM connection ingests CRM data into Data 360 and returns data actions to that org; it isn't a query mechanism for arbitrary external systems. Data Explorer is a Salesforce-hosted visual tool, not something an external system embeds to run custom SQL.

Question 15

Ursa Major Solar wants a segment that accepts a different filter value, such as a specific product SKU, at execution time when it's invoked through an API call from a broadcast flow, without permanently storing membership data in a Segment Membership DMO. Which segment type meets this requirement?
A. A standard segment with the Don't Refresh schedule option selected.
B. A waterfall segment built on the same DMO as the runtime filter target.
C. A dynamic segment.
D. An Einstein-generated segment created from a natural-language description.

Accepted Answer

Correct answer: C. A dynamic segment.
A dynamic segment runs segment queries without persisting data in Segment Membership DMOs and lets attribute filters use placeholders that accept dynamic values at execution time, run through an API call such as a broadcast flow, which matches every part of this requirement. A standard segment with Don't Refresh selected still persists membership data on publish and doesn't support runtime placeholder values; it simply delays when the schedule starts. A waterfall segment processes a priority-ordered list of existing batch segments and doesn't accept runtime filter parameters either. An Einstein-generated segment uses generative AI to suggest a segment definition from a description at build time; it doesn't provide runtime placeholders evaluated per API call.

Question 16

Northern Trail Outfitters' support team wants to know the instant a customer's Case Priority field changes to Critical on a data model object record, so they can immediately route the case to a senior agent, rather than waiting for the next scheduled segment refresh. Which Data 360 feature should the consultant recommend?
A. Configure a data action that sends a change event to a target when the Case Priority condition is met.
B. Publish a segment on a daily schedule that filters cases with Critical priority.
C. Build a Data Shares connection so the support system can query the data model object directly for periodic reporting rather than immediate routing.
D. Add a related list enrichment to the case record page showing Data 360 priority history.

Accepted Answer

Correct answer: A. Configure a data action that sends a change event to a target when the Case Priority condition is met.
Data actions monitor data model objects and calculated insight objects in Data 360 and send near-real-time change events to a target such as Salesforce Platform Event, Webhook, or Marketing Cloud Engagement when a defined condition is met, exactly what Northern Trail Outfitters needs to route the case the instant Case Priority becomes Critical. A scheduled segment publish only runs on a daily or other fixed cadence and does not fire the moment the field changes. A Data Shares connection is built for zero-copy analytical and machine learning access to data, not for pushing event-driven alerts that route a case. A related list enrichment surfaces historical Data 360 data on a record page; it does not detect a change or trigger any real-time routing action.

Question 17

Cloud Kicks wants to activate device-based advertising identifiers for an external platform activation. Which two contact points require the Device object to be mapped, in addition to Contact Point App, before they can be used in the activation? Choose 2 answers.
A. Mobile Advertiser ID (MAID).
B. Over-the-top (OTT) ID.
C. Email Address.
D. Phone Number.
E. WhatsApp.

Accepted Answer

Correct answer: A. Mobile Advertiser ID (MAID). | B. Over-the-top (OTT) ID.
Both Mobile Advertiser ID and Over-the-top ID require the relationship chain Contact Point App.Device to Device.Device Id, in addition to Contact Point App.Party to Individual.Individual Id, along with Device fields like Device Id, Advertiser Id, and OS Name, because these contact points are device-centric advertising identifiers. Email Address only requires the Contact Point Email.Party to Individual.Individual Id relationship with Subscriber Key and Email Address fields; it has no Device object dependency. Phone Number similarly only requires the Contact Point Phone.Party to Individual.Individual Id relationship with phone-specific fields, not a Device mapping. WhatsApp requires the Contact Point OTT Service.Party to Individual.Individual Id relationship with Subscriber Key, Username, and Country; it also doesn't require the Device object mapping that MAID and OTT need.

Question 18

Cumulus Financial builds a segment filter where Product_Category equals "Loan," but the source system stores the value as "loan" in lowercase for many records. Will those lowercase records be included in the segment?
A. Yes, because segment filter values aren't case sensitive, so "Loan" matches "loan."
B. No, the consultant must add a second filter value for "loan" in lowercase to capture those records.
C. No, filter values must exactly match the case stored in the source system, or the container returns zero results.
D. Yes, but only because the field's data type is Text and a calculated insight normalizes the case automatically.

Accepted Answer

Correct answer: A. Yes, because segment filter values aren't case sensitive, so "Loan" matches "loan."
Segment filter values in Data 360 are not case sensitive, so a filter for "Loan" already matches records stored as "loan," "LOAN," or any other casing. There's no need to add a second value or duplicate condition for a different case, since case differences don't affect whether a record matches. Filter values don't need to exactly match the source system's casing, which is exactly why the container still returns the lowercase records without any additional configuration. This behavior applies to the filter itself and doesn't require building a calculated insight to normalize text case first.

Free Salesforce Data 360 Consultant
Practice Questions

What is the Data 360 Consultant exam like?

Questions by exam section

Solution Positioning

Question 1 · Medium · Choose 2

Question 2 · Harder · Choose 2

Question 3 · Medium

Data 360 Setup and Administration

Question 4 · Medium

Question 5 · Harder

Data Source Connection and Ingestion

Question 6 · Medium · Choose 3

Question 7 · Medium

Question 8 · Easier

Harmonization and Unification

Question 9 · Harder

Question 10 · Harder

Question 11 · Harder · Choose 2

Data Activations and Utilization

Question 15 · Medium

Question 16 · Easier

Question 17 · Harder · Choose 2

Question 18 · Easier

Where to go after these 18 questions

Ready for a full Data 360 Consultant practice exam?

Free Salesforce Data 360 ConsultantPractice Questions