CBDA Mock Questions Set 2: 35 New Scenario Questions with Answers

Finished set 1 and want fresh material? These 35 CBDA mock questions are all new scenarios, weighted like the official IIBA blueprint, with an answer key that explains every option and names the section of the Guide to Business Data Analytics behind it.

Last verified: October 2026 (IIBA CBDA page and CBDA Certification Handbook).

How to use this set: the real CBDA exam has 75 scenario-based questions in 2 hours, about 96 seconds each. Give yourself about 55 minutes for these 35 questions, answer all of them without looking at the key, then score yourself. Note which domain your misses fall into: that tells you what to study next.

This is set 2 of our CBDA mock question series. Earlier sets: set 1. The domain mix below follows the IIBA weights: Identify the Research Questions 20% (7 questions), Source Data 15% (5), Analyze Data 16% (5), Interpret and Report Results 20% (7), Use Results to Influence Business Decision Making 20% (7), and Guide Organization-level Strategy for Business Analytics 9% (4).

Identify the Research Questions (7 questions)

Question 1

A healthcare provider wants to improve the patient experience of its discharge process. The analyst is formulating research questions and wants at least one descriptive, one predictive, and one prescriptive question, mirroring the Guide's example. Which question is the prescriptive one in this set?

• A. Will the discharge experience improve if the network adds a follow-up call feature for patients?
• B. What are the underlying factors that most influence the patient discharge experience overall?
• C. Which measures should the network use in order to evaluate the patient discharge experience?
• D. How should the network classify each discharge as either a positive or a negative experience?

Question 2

Following the Guide's case-study workshop approach, an analyst at a hospital group is running the live problem-framing session. Which activity is the Guide's intended use of techniques such as root cause analysis, 5 Whys and the business model canvas during the workshop?

• A. To coordinate the workshop logistics and finalise the agenda beforehand
• B. To identify the problems to be addressed so they can be defined as research questions
• C. To build the production analytics pipeline that will serve the results
• D. To validate the accuracy dimension of the source data after analysis

Question 3

A SaaS analyst is told that stakeholder identification is the foundation for later communication planning in the analytics effort. For an expansion-revenue initiative, which question from the Guide's stakeholder checklist directly informs how and when to keep stakeholders engaged?

• A. Which gradient-descent learning rate should the predictive analytics model use during its training on the historical billing data?
• B. How many null values currently exist within the billing table, and which of its columns are required versus optional for the records?
• C. What is the projected annual recurring revenue uplift from the initiative, and what is the expected payback period on the investment?
• D. What communication methods and techniques are appropriate, and when should the stakeholders be communicated to during the work?

Question 4

A telecom analytics team must decide how formally to document its business data analytics approach. Some members want a strict template; others prefer visual models on a shared wiki. According to the Guide, what is the correct view on formality?

• A. A formally signed planning template is required in every case before any of the analytics work on the initiative is allowed to begin at all.
• B. There is no right or wrong answer on formality; some teams use a planning template while others build visual models on shared wikis.
• C. Documentation of the approach should be avoided wherever possible, so that the analytics team is able to move through the work much faster.
• D. Only once the team reaches the production stage of the work does any documentation of the chosen analytics approach actually become necessary.

Question 5

In the Guide's investment-banking case study, the analyst anticipates that adopting non-traditional analytics will feel like a big change for many trading-floor stakeholders. According to the workshop plan, how does she address this in the session?

• A. She provides a short explanation of the use of deep analytics to ease the change
• B. She removes the topic of deep analytics from the agenda to avoid resistance
• C. She defers all stakeholder communication until after the production rollout
• D. She replaces the workshop with a written memo sent only to senior management

Question 6

An analyst at a broadband provider is identifying stakeholders for an initiative on installation delays. A colleague argues they only need to talk to senior managers. Per the Guide, which addition to the stakeholder set is MOST important and why?

• A. The custodians and consumers of data, because identifying them is critical to the initiative
• B. Only the most senior executive, because that single view is sufficient to scope the work
• C. Only external customers, because internal roles add no value to stakeholder analysis
• D. No one further, because stakeholder identification can wait until reporting begins

Question 7

A public-sector agency is defining the future state for a model that flags benefit applications as fraudulent or not. Stakeholders worry that wrongly flagging legitimate applicants is far more harmful than missing some fraud. Which measurable success criterion BEST reflects defining the future state for this classification problem?

• A. Prioritize high precision, TP/(TP+FP), so that of the applications flagged as fraud, very few are actually legitimate.
• B. Maximize the raw count of applications processed per day regardless of the model's outputs.
• C. Pick whichever metric the engineering team finds easiest to compute from the logs.
• D. Use no numeric measure and judge success by stakeholder impressions after launch.

Source Data (5 questions)

Question 8

Northwind Telecom must decide whether to pull subscriber usage data directly from its centrally managed data warehouse or from a peripheral secondary source where the data has already been manipulated into a more usable form, but whose quality is questionable. According to Plan Data Collection guidance, what MOST influences the sourcing choice here?

• A. Whether the secondary source uses the same query language as the warehouse so that no SQL is ever needed
• B. Whether the convenience of the pre-massaged secondary source outweighs the quality risk versus a direct pull that costs more work
• C. Whether the data can be relabeled as unstructured so that it bypasses any data quality assessment entirely
• D. Whether the warehouse can be shut down so that all teams are forced to use the secondary source only

Question 9

In the worked Source Data case study, Adaku, a telecom expert, was asked to help predict real-time fraudulent traffic. Industry, business, and solution knowledge pointed her to one data source as most relevant. Which lesson does this BEST illustrate?

• A. The first data source chosen should never be changed during the work
• B. Industry and business knowledge are key competencies for identifying the most relevant data source
• C. Raw data is always sufficient for analysis without any transformation
• D. Surveys are the only reliable way to detect fraud

Question 10

A telecom analyst is planning data collection for a churn study. Some needed data is available only during specific time frames and other data is restricted by privacy rules. When planning data collection, what should the analyst consider EARLY?

• A. Only the volume of data, ignoring whether it can be obtained
• B. Deferring all questions of availability until after analysis
• C. Collecting only unstructured data to simplify the project
• D. The availability of the data, including time-frame and privacy constraints

Question 11

A SaaS team reviews the Source Data case study, where raw call detail records were not directly usable and had to be reshaped before they yielded insight. Which case study takeaway does this illustrate?

• A. Available data may need additional transformation to serve its intended use
• B. A structured plan alone guarantees that raw data is immediately analysis-ready
• C. Industry knowledge is unnecessary once the right data source is identified
• D. Successive experimentation should be avoided once a first method is chosen

Question 12

A SaaS analyst is asked to collect product-usage data. The platform already records clickstream and in-app events during normal use, with no analytics objective originally in mind. Which type of collection does this represent?

• A. Active data collection, since it requires a survey to gather
• B. Active data collection, since the goal is specific from the start
• C. Passive data collection, since it is unobtrusive data from day-to-day transactions
• D. Neither passive nor active, because event data cannot be reused

Analyze Data (5 questions)

Question 13

A public-sector analyst is drafting a data analysis plan and must explain to non-technical stakeholders why a particular algorithmic model was chosen for the research question. According to the Develop Data Analysis Plan task, what level of model understanding is expected of the analyst?

• A. Deep enough to derive each algorithm's mathematics from first principles.
• B. Enough to explain the analysis approach and why a model fits the research question.
• C. No understanding, since model choice is hidden from the analyst entirely.
• D. Enough to hand-code every model without any data scientist involvement.

Question 14

During exploratory analysis of a fraud dataset a finance analyst sees readings far outside the normal range. According to the Guide, what do such data outliers (noise) signal?

• A. That the research question itself must be discarded immediately.
• B. That the data may need to be excluded as part of assessing data quality.
• C. That the model results are confirmed and ready to report.
• D. That stakeholders should approve the budget before any cleansing.

Question 15

An energy retailer is preparing data and finds the customer city is stored as a full text string repeated in 10 meter-reading rows for one customer: 6 read 'Bristol', 4 read 'bristol'. The team wants to normalize this to reduce data redundancy and standardize the value. After standardizing case and storing the city once, how many distinct city values should represent this customer?

• A. 1 distinct value.
• B. 2 distinct values.
• C. 6 distinct values.
• D. 10 distinct values.

Question 16

A logistics firm forms a new data science team modelled on the Guide's Analyze Data case study, where data scientists struggled to understand business needs. Which approach did the business analyst use to bridge the gap?

• A. Replace the data scientists with business analysts across the whole team
• B. Stop all communication with the data scientists until the model was fully finished
• C. Learn the data scientists' terminology and translate it into business terms meaningful to stakeholders
• D. Have business stakeholders write the model code directly themselves

Question 17

During data analysis at an investment firm, an analyst compares several predictors against asset prices and observes that most predictors have a straight-line relationship with price. Which interpretation of this observation is MOST accurate per the Perform Data Analysis task?

• A. Any change in the predictor leaves the price completely unchanged.
• B. A change in the predictor changes the price by a proportional amount.
• C. Price changes only in large, discrete jumps unrelated to the predictor.
• D. The predictor and price move in opposite directions by random amounts.

Interpret and Report Results (7 questions)

Question 18

A healthcare analyst must distinguish a true business insight from a first-level data indication when reporting model results. Which option correctly defines a derived insight as opposed to an adjacent concept?

• A. The first level of inference drawn from data using statistical tools or visualizations
• B. A business-meaningful statement that answers the so-what question for a stakeholder
• C. The act of deciding cadence and channels for each stakeholder audience
• D. The packaging and summarization of a completed report into a dashboard

Question 19

A healthcare analyst computes a single headline figure for a leadership review. Of 800 discharges, 760 had complete records. The analyst reports the completeness percentage as one figure. What is the percentage, and how should it be presented per best practice?

• A. 95 percent, presented as a highlighted text figure since a single metric is often clearer as text than as a graph
• B. 85 percent, presented inside a 3-dimensional pie chart purely for emphasis
• C. 95 percent, presented only as a complex dual-axis chart with a trend line
• D. 40 percent, presented as a multi-series stacked bar to fill out the report

Question 20

An energy analyst plans stakeholder communication for an iterative initiative. Which consideration does the Guide list for planning this communication?

• A. All stakeholders must receive the same level of technical detail
• B. Communication should occur only once, at project close
• C. Communications may report an intermediate or final result, since analytics initiatives are iterative
• D. Confidentiality concerns can be ignored during reporting

Question 21

Hiroshi presents weekly churn to a telecom CEO who has 12 minutes total for the meeting and is known to skim charts. The analysis covers two segments, three regions and four contract types over twelve weeks. Which reporting choice is the BEST primary deliverable for this audience?

• A. A 28-slide deck with one chart per segment-region-contract combination so the CEO can drill into any cell during the 12-minute slot at the meeting.
• B. A live BI dashboard with twenty interactive filters that the CEO can manipulate during the 12-minute meeting on a connected display in the executive boardroom.
• C. A raw data export of every weekly churn record so the CEO can perform any breakdown personally during the 12 minutes available before the next standing meeting.
• D. A single page combining one Big Idea sentence and one well-labelled chart that summarises the headline finding, with the detail available as appendix on request.

Question 22

A manufacturing analyst is matching reporting terms. Which option correctly defines Plan Stakeholder Communication, the term being tested, and not an adjacent task?

• A. Continually re-validating the stakeholder analysis as needs and personnel change
• B. Translating raw data patterns into useful business insight statements
• C. Deciding the channels, cadence, and depth of communication for each audience
• D. Connecting the entire analysis journey as a visual hierarchy of decision points

Question 23

A public-sector analyst is building the narrative for a completed transit-ridership analysis. According to the Guide, which set of questions should the analyst address when building that narrative from the findings?

• A. Which vendor offers the cheapest charting tool and what is its license cost
• B. How many slides the executive deck should contain and in what font
• C. What time of day the report email should be scheduled to send
• D. Where are the data gaps, what does this mean for the business, and how can the business improve

Question 24

Scenario: A SaaS marketing team modeled on the HiFive Ice Creams case, in which HiFive set out to measure the success of its social-media marketing and the data science team studied which platform to use. The team must report which single medium to prioritize for the limited marketing budget. In the HiFive case, what did the data science team report as the optimal medium and why?

• A. Twitter, because it allowed the cheapest cost per impression in the test
• B. Instagram, because it produced the most visually engaging ice cream content
• C. A blend of all platforms equally, because no single platform stood out in the study
• D. Facebook, because it had the highest regional adoption, which was conducive to the research problem

Use Results to Influence Business Decision Making (7 questions)

Question 25

A public agency wants its analytics insights to actually change policy. Per the domain case study, what is the prerequisite for insights to influence a decision?

• A. The insights must be effectively communicated to build shared understanding
• B. The insights must come from the largest available dataset
• C. The insights must be kept confidential from stakeholders
• D. The insights must be presented only as raw tables

Question 26

At an energy utility, an analytics-driven process change has been agreed and the analyst is acting as change manager. What BEST describes this role per Manage Change?

• A. Building the predictive models that produced the original insights
• B. Overseeing the transformation of analysis results into implemented policies and procedures
• C. Approving the project budget on behalf of the executive sponsor
• D. Writing the SQL queries that extract the source data for analysis

Question 27

A government analyst presents a predictive proof of concept that detects anomalous pricing in a procurement platform. What problem is this analytics initiative primarily designed to surface?

• A. Pricing deviations that result from incorrect data input
• B. Employee scheduling conflicts
• C. Network latency in the platform
• D. Customer satisfaction trends over time

Question 28

A public-sector data team frames its implementation plan. Which single phrase BEST describes what Develop Implementation Plan accomplishes?

• A. It selects the research question to investigate
• B. It defines the F1 evaluation criterion for the model
• C. It writes the final press communication
• D. It plans how the recommendation gets executed

Question 29

A public agency realizes its analytics effort only delivers value at a specific point. According to Manage Change, when does the organization realize value from its analytics efforts?

• A. When the predictive model first achieves a high accuracy score
• B. When the research question is initially approved by sponsors
• C. When change is implemented and results become policies and procedures
• D. When the raw data is first loaded into the warehouse

Question 30

A retail analyst is well positioned to serve as change manager for an analytics-driven change. Why does the Guide consider business analysts well suited to this role?

• A. They ensure continuity between the analytics work and its implementation
• B. They control the project budget and approvals
• C. They write the predictive model code themselves
• D. They replace the need for stakeholder agreement

Question 31

In the case study for this domain, an analytics team in the UK giftware industry succeeded largely because of one practice. What was the starting point that let their insights influence decisions?

• A. Replacing the executives who disagreed with the data
• B. Developing a shared understanding of the analytics results and the insights from them
• C. Increasing the marketing budget before any analysis was reviewed
• D. Outsourcing the decision entirely to an external consultancy

Guide Organization-level Strategy for Business Analytics (4 questions)

Question 32

MedNorth Health, a regional hospital network, is standing up its first analytics group. Leadership has approved budget but the head of analytics, Priya, must decide how to build the team and keep talent. The roles available include data analyst, experience designer, data journalist, and business analysis professional. Priya has a strong business analysis professional who also has sufficient analytics and business knowledge. According to the Guide, what does this imply for team structure?

• A. Roles can never be combined under any circumstances, so a completely separate person is strictly mandatory for every single defined role
• B. The strong business analysis professional must be removed from the team because role overlap between people is simply not allowed at all
• C. Multiple roles can be merged, since a business analysis professional with enough analytics and business knowledge can also act as SME, data analyst, and data journalist
• D. The team structure is fixed entirely by external law and so it cannot reflect this organization's own capabilities in any way at all

Question 33

A manufacturer wants to align its new data strategy to the company's business objectives and outcomes, viewing the organization from several perspectives at once rather than a single financial lens. Which technique does the Guide recommend as useful for this alignment?

• A. Balanced Scorecard.
• B. Risk Analysis.
• C. Collaborative Games.
• D. Business Model Canvas.

Question 34

Atlas Capital is forming a single analytics team that operates as one unit supporting all other business units in decision-making, and it plans to upskill talent so the team stays cohesive. Which model is this, and what advantage does the Guide associate with it?

• A. A centralized model, often realized as a Centre of Excellence, where upskilling talent is an advantage because it forms a cohesive team
• B. A decentralized model, whose stated advantage is that absolutely no shared standards are ever needed across any of the units
• C. A hybrid hub and spoke model, whose defining feature is the complete absence of any central coordinating analytics unit at all
• D. A proof-of-concept pilot model, whose stated advantage is the permanent isolation of analytics work from the other business units

Question 35

FleetLink, a logistics provider, has run several analytics pilots embedded in separate business units and now wants an organization-level approach. The leadership team is deciding both how to organize the analytics teams and how to manage data as the number of initiatives grows. As the number of analytics initiatives grows, FleetLink has no single source of truth and struggles to track data and data requests. Which approaches does the Guide suggest to meet data needs in this situation?

• A. Permanently stop and freeze all new analytics initiatives across the company until a single source of truth has finally been built
• B. Delete large parts of the historical data so that there is simply less data left for the analytics team to have to track over time
• C. Forbid every business unit from ever requesting any organizational data again so that the volume of data requests falls to zero
• D. A just-in-time approach or a self-service strategy can be employed to cater to data needs in the absence of a single source of truth

Answer key for CBDA mock questions set 2

Each explanation gives the reason for the correct option, explains why the other options are wrong, and names the section of the Guide to Business Data Analytics to reread.

1. Correct answer: A. Following the Guide's customer-experience example, a question asking whether the experience would improve by adding a feature (here, a follow-up call) is the prescriptive analytics question. Why others wrong: B - identifying the factors that influence experience is descriptive; C - identifying the measures for evaluating experience is descriptive; D - classifying each discharge as positive or negative is predictive. Source: Guide Section 2.1.5

2. Correct answer: B. In the worked case study, analysis techniques such as root cause analysis, 5 Whys and the business model canvas are used in the workshop to identify the problems to be addressed, which are then concretely defined as research questions. Why others wrong: A - coordinating logistics and the agenda is pre-workshop preparation; C - building a production pipeline is downstream delivery, not problem identification; D - validating data accuracy belongs to a later data-validation activity, not this workshop step. Source: Guide Section 2.1.8

3. Correct answer: D. Among the questions the Guide says analysts seek to answer before an initiative starts are which communication methods and techniques are appropriate and when stakeholders should be communicated to, which directly informs communication planning. Why others wrong: A - the learning rate is a modelling parameter, not a stakeholder question; B - null values in a table are a data-quality concern; C - revenue uplift projection is a business-case figure, not part of the stakeholder checklist. Source: Guide Section 2.1.2

4. Correct answer: B. The Guide states there is no right or wrong answer as to the degree of formality of the approach: some organisations formally document decisions using a planning template, while others build visual models captured on shared wikis. Why others wrong: A - a formal signed template is not required in every case; C - the Guide describes documenting decisions, not avoiding documentation; D - planning is part of every domain and is iterative, not confined to the production stage. Source: Guide Section 2.1.6

5. Correct answer: A. The case-study workshop plan states that because adopting new ways to conduct trades is a big change, the analyst provides a short explanation on the use of deep analytics. Why others wrong: B - removing the topic contradicts the plan, which deliberately introduces it; C - deferring communication until after rollout reverses the early-engagement intent of the workshop; D - the case study uses a facilitated workshop, not a memo to senior management only. Source: Guide Section 2.1.8

6. Correct answer: A. The Guide stresses that it is critical to understand the custodians and consumers of data when identifying the relevant stakeholders, using techniques such as brainstorming, interviews, and reviewing process flows and organisational charts. Why others wrong: B - relying on one senior view ignores the differing needs of stakeholder groups; C - excluding internal roles drops the very data custodians and consumers the Guide highlights; D - deferring stakeholder identification to the reporting stage contradicts the Guide's before-the-initiative timing. Source: Guide Section 2.1.2

7. Correct answer: A. defining the future state establishes measurable objectives, and for a fraud classification problem the Guide lists precision = TP/(TP+FP); when false positives (flagging legitimate applicants) are the most harmful error, precision is the appropriate measure of success. Why others wrong: B - throughput ignores the classification quality the future state must measure; C - choosing a metric by ease of computation does not reflect the business context that governs success; D - the future state requires measurable objectives, not impressions alone. Source: Guide Section 2.1.4

8. Correct answer: B. Plan Data Collection notes that how much data must be manipulated influences sourcing, and that a direct pull with extra work may be acceptable if post-massaged secondary data is questionable on quality, so an assessment of data quality helps determine the best source. Why others wrong: A - matching query language is not the stated driver of the sourcing decision; C - relabeling data as unstructured does not remove the need for a quality assessment; D - shutting the warehouse down is not part of the sourcing guidance. Source: Guide Section 2.2.1

9. Correct answer: B. The case study key takeaways state that industry, business, and solution knowledge are key competencies that helped identify the most relevant data, with Adaku's knowledge of SIMbox fraud pointing her to the call detail record. Why others wrong: A - the case shows successive refinement of data and methods, so sources do change; C - the takeaways note the available data was not directly useful in raw form and needed transformation; D - the case relied on call detail records and predictive analytics, not surveys. Source: Guide Section 2.2.6

10. Correct answer: D. When planning data collection, analysts consider what data is needed and the availability of the data, and the Guide notes some data may be unavailable due to privacy rules while other data is only available during specific time frames. Why others wrong: A - planning weighs availability and validation, not volume alone; B - availability is a planning consideration handled before sourcing, not deferred until after analysis; C - planning determines how much structured versus unstructured data is feasible, not a blanket choice of unstructured only. Source: Guide Section 2.2.1

11. Correct answer: A. A case study key takeaway is that available data may not be directly useful and undergoes additional transformation to serve its intended use, as when the CDR was reshaped using business knowledge. Why others wrong: B - the takeaways say raw data often needs transformation, not that a plan makes it immediately ready; C - the takeaways stress industry, business, and solution knowledge as key, not unnecessary; D - the takeaways credit successive experimentation for surfacing more accurate methods rather than avoiding it. Source: Guide Section 2.2.6

12. Correct answer: C. Passive data collection is unobtrusive data gathered from users in their day-to-day transactions, available without an analytics objective in mind, such as web and mobile data, which matches the clickstream and in-app events. Why others wrong: A - active collection seeks information through means like surveys, which is not how this event data is obtained; B - the data was recorded without an original analytics objective, so it is not active collection driven by a specific goal; D - such data often exists already and is curated or transformed for research questions, so it can be reused. Source: Guide Section 2.2.3

13. Correct answer: B. The Guide says analysts understand the mathematical techniques and algorithmic models in sufficient detail to explain the analysis approach to business stakeholders, including why a particular model may be chosen for the research question. Why others wrong: A - the Guide says analysts may not require a rigorous understanding of the algorithms; C - the analyst does engage with model choice to explain it, so it is not hidden; D - the data scientist decides and conducts the modelling, so the analyst need not hand-code every model. Source: Guide Section 2.3.1

14. Correct answer: B. The Guide says the data scientist looks for data gaps or redundancy that signal cleansing may be needed, or data outliers (noise) that signal data may need to be excluded, as part of assessing data quality. Why others wrong: A - outliers prompt data treatment, not abandoning the question; C - finding outliers during exploration is the opposite of confirmed, report-ready results; D - budget approval is unrelated to what outliers signal in exploration. Source: Guide Section 2.3.3

15. Correct answer: A. Preparing data includes normalizing to reduce data redundancy plus standardization; standardizing case merges 'Bristol' and 'bristol' into one value, and storing the city once for the customer yields a single distinct value. Why others wrong: B - 2 reflects the un-standardized casing that normalization removes; C - 6 counts only the capitalized rows, ignoring standardization; D - 10 counts every redundant row, which is exactly what reducing redundancy eliminates. Source: Guide Section 2.3.2

16. Correct answer: C. In the case study, Haru collaborated closely with data scientists to learn terminology, understood the relevance of their questions, and translated this into business terms meaningful to stakeholders so the team built a shared understanding. Why others wrong: A - the approach was collaboration, not replacement; B - the approach increased communication rather than halting it; D - stakeholders writing model code is not part of the described bridge. Source: Guide Section 2.3.7

17. Correct answer: B. The Guide states that when predictors have a linear relationship with price, some change in the predictor changes the price by a proportional amount in a straight-line relationship. Why others wrong: A - a linear relationship means price does respond to the predictor, not stay unchanged; C - a straight-line relationship is proportional, not discrete jumps; D - linear here describes proportional movement, not random opposite movement. Source: Guide Section 2.3.4

18. Correct answer: B. Deriving insights means translating raw results into business-meaningful statements where an insight answers the so-what question for a stakeholder. Why others wrong: A - the first level of inference drawn with statistical tools is the Analyzing Data step, not yet a true business insight; C - deciding cadence and channels describes Plan Stakeholder Communication; D - packaging and summarizing a report describes Document and Communicate Findings. Source: Guide Section 2.4.4

19. Correct answer: A. 760 of 800 is 95 percent, and best practice notes that when only a single metric is involved, simple, strategically placed text is often more effective than a graph for figures like percentages. Why others wrong: B - 760 of 800 is 95 percent not 85 percent, and a 3-dimensional pie is discouraged; C - a single value does not warrant a complex dual-axis chart; D - 40 percent miscalculates the figure and stacked bars add needless complexity. Source: Guide Section 2.4.1

20. Correct answer: C. The Guide says stakeholder communications can involve an intermediate or final result since analytics initiatives are inherently iterative, with formality and detail varying among stakeholders. Why others wrong: A - the Guide says level of detail may vary among stakeholders, not be uniform; B - it advises maintaining an appropriate level of communication throughout, not only at close; D - it lists maintaining the appropriate level of privacy and confidentiality as a consideration. Source: Guide Section 2.4.2

21. Correct answer: D. D4 Determine Communication Needs of Stakeholders prescribes audience-fit reporting - executives need headline plus chart plus an appendix on request, as captured by The Big Idea and Business Visualizations. Why others wrong: A - 28 slides over-serves an audience that has 12 minutes; B - interactive dashboards are a self-serve tool, not a fit for a skim-only audience; C - raw exports place the analytical burden on the executive. Source: Guide Section 2.4.3, Section 3.19

22. Correct answer: C. Plan Stakeholder Communication is the task of deciding channels, cadence, and depth for each audience as part of the communication approach. Why others wrong: A - re-validating the stakeholder analysis describes Validate Understanding of Stakeholders; B - translating patterns into insights describes Derive Insights from Data; D - connecting the journey as a visual hierarchy describes the Data Journeys and Orchestration technique. Source: Guide Section 2.4.2

23. Correct answer: D. When building the narrative, the Guide says questions such as where are the data gaps, what does this mean for the business, and how can the business improve are addressed. Why others wrong: A - vendor pricing is not part of building the findings narrative; B - slide count and font are formatting choices, not the narrative questions cited; C - email send timing is a logistics detail, not a narrative-building question. Source: Guide Section 2.4.5

24. Correct answer: D. In the HiFive case study, the data science team responded that based on their study Facebook had the highest regional adoption, which was conducive to the research problem, so they selected Facebook as the optimal medium. Why others wrong: A - the case did not cite Twitter or cost per impression as the deciding factor; B - Instagram and visual engagement are not the reason given; C - the team selected one optimal medium rather than an equal blend. Source: Guide Section 2.4.7

25. Correct answer: A. The case study takeaway is that insights can influence decisions but they need to be effectively communicated, starting with a shared understanding. Why others wrong: B - dataset size is not the stated prerequisite; C - confidentiality from stakeholders prevents shared understanding; D - raw tables alone do not build understanding. Source: Guide Section 2.5.5

26. Correct answer: B. In the change-manager role the analyst oversees transforming analysis results into implemented policies and procedures, since implementing change is where the organization realizes value. Why others wrong: A - model building is upstream analysis work, not change management; C - budget approval sits with the sponsor, not the change manager; D - data extraction is a data-preparation task, not the change role. Source: Guide Section 2.5.3

27. Correct answer: A. The initiative detects anomalous pricing in order to identify pricing deviations that result from incorrect data input. Why others wrong: B - scheduling conflicts are unrelated to pricing anomaly detection; C - network latency is not what the pricing model surfaces; D - satisfaction trends are a different analysis, not the anomaly target. Source: Guide Section 2.5.2

28. Correct answer: D. Develop Implementation Plan plans how the recommendation gets executed, providing the bridge from analytics to delivery. Why others wrong: A - selecting the research question is an upstream step; B - defining the evaluation criterion is part of the analytics, not the plan; C - communication writing is a separate concern from execution planning. Source: Guide Section 2.5.2

29. Correct answer: C. Manage Change states implementing change is the end goal and it is where the organization realizes the value from its analytics efforts. Why others wrong: A - a high model score is an interim result, not realized value; B - approving the question is the start, not value realization; D - loading data is preparation, far from implemented value. Source: Guide Section 2.5.3

30. Correct answer: A. Business analysts are well suited as change managers because they can ensure the continuity between the analytics work and implementation. Why others wrong: B - budget control is not the cited reason; C - writing model code is not why they fit the change-manager role; D - stakeholders still agree on the changes, so the role does not replace that agreement. Source: Guide Section 2.5.3

31. Correct answer: B. The key takeaway is that insights influence decisions only when communicated effectively, and the starting point is developing a shared understanding of the analytics results and their insights. Why others wrong: A - replacing executives is not part of the case study; C - raising budget before review contradicts the evidence-led approach; D - the team used its own insights to influence decisions rather than outsourcing them. Source: Guide Section 2.5.5

32. Correct answer: C. The Guide says roles have overlapping and complementary skills and multiple roles can be merged to create the right team structure, giving the example that a business analysis professional with sufficient analytics and business knowledge can play SME, data analyst, and data journalist and support data architecture, engineering, and experience design. Why others wrong: A - the Guide explicitly allows merging roles, not one person per role; B - overlap is treated as useful, not a reason to remove the role; D - the Guide says the right structure is guided by the organization's capabilities, not fixed by law. Source: Guide Section 2.6.2

33. Correct answer: A. The Guide describes the Balanced Scorecard as useful for describing a balanced view of the organization from different perspectives and for aligning the data strategy to business objectives and outcomes, which matches the stated goal. Why others wrong: B - Risk Analysis addresses uncertainty and threats, not multi-perspective objective alignment; C - Collaborative Games support group elicitation, not this alignment purpose; D - the Business Model Canvas describes the business model rather than aligning the data strategy across perspectives. Source: Guide Section 2.6.4

34. Correct answer: A. The Guide defines the centralized model as the analytics team operating as a single unit supporting other business units, with an analytics Centre of Excellence as a good example, where upskilling talent may be an advantage as it forms a cohesive team. Why others wrong: B - decentralized embeds analytics in each unit and is not described as one cohesive single unit; C - hybrid is a mix of centralized and decentralized, so it is not defined by the absence of a central unit; D - a pilot has limited strategic impact and isolation is not described as its advantage. Source: Guide Section 2.6.1

35. Correct answer: D. The Guide states that as initiatives increase, tracking an organization's data and data requests becomes complex, and in the absence of a single source of truth a just-in-time approach or a self-service strategy can be employed to cater to data needs. Why others wrong: A - halting all initiatives is not the suggested response; B - deleting historical data is not advised and harms the data asset; C - forbidding data requests contradicts catering to data needs. Source: Guide Section 2.6.3

How did you score?

• 30 to 35 correct: strong result. Check the few misses and consider booking the exam.
• 25 to 29 correct: close to the level most candidates aim for. Reread the Guide sections for your weakest domain.
• Below 25: go back to the Guide chapters where most of your misses fall, then take another set.

IIBA does not publish a passing score, so treat these bands as a practice guide, not an official rule.

How other candidates use practice questions

Candidates who share their CBDA preparation online often recommend practice questions and sample exams to reinforce the concepts. Because the exam is scenario-based, they advise focusing on how to apply each idea rather than memorising definitions. Others describe steady daily study with the Guide and its structure, for example using mind maps of each domain. Mock sets like this one fit that approach: short, timed, and tied to a Guide section you can reread.

More CBDA practice

• Learn the format, fees and domain weights in our CBDA exam guide.
• Plan your preparation with the 6-week CBDA study plan.
• Drill one domain at a time: research questions, source data, analyze data, interpret and report results, business decision making and analytics strategy. For a mixed warm-up, try the scenario questions set.
• Take a full-length timed test in the free CBDA exam simulator.

FAQ

Are these real CBDA exam questions?

No. All questions are original. We write them from the Guide to Business Data Analytics, and we do not use exam dumps. The goal is to practise the way the exam asks questions, not to memorise answers.

What score do I need to pass the CBDA?

IIBA does not publish a passing score. Many prep providers use about 70% as a working target, so on a 35-question set aim for 25 or more correct answers.

Do I need to take the sets in order?

No. Each set is independent and has its own mix of questions, so you can start with any of them. Taking them in order simply gives you a steady way to track progress.

Sources

• IIBA: Certification in Business Data Analytics (CBDA)
• IIBA: Certification fees
• IIBA: CBDA Certification Handbook
• IIBA Sydney chapter: Business Data Analytics study group
• Managing Projects the Agile Way: How to study for and pass the CBDA exam
• Techcanvass: CBDA certification journey of Ivan
• Adaptive US: Mastering the CBDA exam