Study guide
15.2 Selecting Analytical Techniques for Engagement Objectives
CIA Part 2 Study Guide Exam
Track your progress as you read
Finished sections get a check mark.
1 Chapter 1: CIA Part 2 Exam Orientation
1.1 Current CIA Part 2 Exam Facts 1.2 Eligibility, Fees, and Pearson VUE Scheduling 1.3 Scoring, Results Timing, and the 2025 Syllabus Map
2 Chapter 2: Engagement Objectives and Scope
2.1 Topical Requirements in Objectives and Scope 2.2 Elements Used to Build Engagement Objectives 2.3 Scope Limitations, Stakeholder Requests, and Scope Changes
3 Chapter 3: Evaluation Criteria and Planning Context
3.1 Selecting and Testing Evaluation Criteria 3.2 Applying Topical Requirements When Planning 3.3 Strategy, Risk Management, and Performance of the Activity Under Review
4 Chapter 4: Technology, Continuity, and Finance in Planning
4.1 Cybersecurity, IT General Controls, Privacy, and Data Security 4.2 Business Continuity, Disaster Recovery, and Resilience 4.3 Finance and Accounting Concepts for the Activity Under Review
5 Chapter 5: Business Process and System Risks in Planning
5.1 Asset, Supply Chain, Inventory, Payables, Procurement, and Compliance Processes 5.2 Third-Party Process Risks 5.3 CRM, ERP, and GRC System Risks
6 Chapter 6: Engagement Approach and Project Management
6.1 Agile, Traditional, Integrated, and Remote Approaches 6.2 Project Management While Planning and Conducting Engagements
7 Chapter 7: Engagement Risk Assessment
7.1 Topical Requirements and Pervasive Financial, Operational, IT, Cyber, and Regulatory Risks 7.2 Emerging Risks and the Impact of Change 7.3 Methods to Evaluate and Prioritize Risks and Controls
8 Chapter 8: Structure, Environment, and Culture
8.1 Centralized vs Decentralized, Flat vs Traditional, In-Person vs Remote 8.2 Culture, Group Behavior, and Tone at the Top
9 Chapter 9: Control Design, Effectiveness, and Efficiency Procedures
9.1 Procedures to Evaluate Control Design 9.2 Procedures to Test Operating Effectiveness 9.3 Procedures to Test Control Efficiency
10 Chapter 10: Work Programs, Testing Methodologies, and Resources
10.1 Evaluating Adequacy of the Engagement Work Program 10.2 Testing Methodologies for Accounting, Finance, IT, Operations, and Cybersecurity 10.3 Financial, Human, and Technological Resources and Limitation Implications
11 Chapter 11: Information Sources and Preliminary Survey
11.1 Interviews, Observations, and Walk-Throughs 11.2 Data Analysis as an Information Source 11.3 Policies, Checklists, Risk-and-Control Questionnaires, and Self-Assessments
12 Chapter 12: Evidence Relevance, Sufficiency, and Reliability
12.1 Criteria for Evidence Quality 12.2 Independence, Corroboration, and the Control Environment 12.3 Evidence an Informed, Competent Person Could Rely On
13 Chapter 13: Technology Options for Findings and Conclusions
13.1 Artificial Intelligence, Machine Learning, and Robotic Process Automation 13.2 Continuous Monitoring, Dashboards, and Embedded Audit Modules
14 Chapter 14: Process Mapping and Data Analytics
14.1 Workflow Segments, Process Maps, Walk-Throughs, and RACI 14.2 Structured vs Unstructured Data and the Analytics Process 14.3 Diagnostic, Prescriptive, Predictive, Anomaly, and Text Analysis
15 Chapter 15: Analytical Review and Finding Evaluation
15.1 Ratios, Variances, Trends, and Benchmarking 15.2 Selecting Analytical Techniques for Engagement Objectives 15.3 Comparing Existing Conditions to Evaluation Criteria 15.4 Root Causes, Effects, and Significance of Findings
16 Chapter 16: Workpapers and Engagement Conclusions
16.1 Organizing Complete Workpapers with Sufficient Evidence 16.2 Linking Workpapers to Results; Retention, Regulation, and Policy 16.3 Aggregating Findings and Concluding on Governance, Risk, and Control
17 Chapter 17: Engagement Supervision
17.1 Supervision Throughout Planning and Fieldwork 17.2 Coordinating Assignments and Reviewing Workpapers and Conclusions 17.3 Evaluating Auditors' Performance
18 Chapter 18: Communication During the Engagement
18.1 Formal and Informal Communication Across Planning, Fieldwork, and Reporting 18.2 Situations That Require Escalation 18.3 Selecting Appropriate Stakeholders for Engagement Communication
19 Chapter 19: Final Review and Test Day
19.1 2025 vs 2019 Traps and Part 2 vs Part 3 Boundaries 19.2 Pacing, Test-Center Rules, and Official Results
15.2 Selecting Analytical Techniques for Engagement Objectives
Key Takeaways Explain This Topic AI
B5b tests whether you can match the analytical technique to the engagement objective: efficiency, completeness, fraud indicators, cutoff, and accuracy each need a different lens.
A reasonableness test builds an independent estimate from operational drivers and compares it to the recorded amount; it is the completeness-oriented cousin of ratio analysis.
Precision of the expectation limits how much assurance analytics can give; a wide, poorly sourced expectation cannot support a high-assurance objective.
When the difference exceeds the threshold and inquiry is uncorroborated, or when analytics cannot be precise enough for the objective, perform tests of details.
Fraud and existence objectives almost never rest on high-level ratios alone; analytics flag anomalies, then details (and targeted data interrogation) have to carry the assurance.
Last updated: August 2026
Match the Technique to the Engagement Objective
Section B5b asks you to determine appropriate analytical techniques to achieve engagement objectives. The syllabus does not want a favorite tool. It wants a match . Efficiency, completeness, existence, accuracy, cutoff, and fraud-or-abuse each fail in different ways, so each needs a different analytical lens—or a decision that analytics will not be enough and tests of details must carry the work.
Start from the objective written in the work program, not from the data you happen to have. If the objective is "assess whether warehouse labor is used efficiently," inventory turns and the current ratio are the wrong instruments even if they are easy to compute. If the objective is "assess whether recorded sales are complete," vouching recorded invoices to customer orders tests the wrong direction (occurrence, not completeness). The technique has to be capable of detecting the kind of misstatement or process failure the objective cares about.
Engagement objective Analytical technique that fits Independent inputs When it is not enough
Efficiency of a process Flexible budget / per-unit variances; output per input (lines per hour, claims per FTE, cost per shipment) Volume, standard rates, engineered standards, peer sites Standards are outdated, volumes are unreliable, or the unit of measure does not match the work
Completeness of recorded activity Reasonableness test from operational drivers; related-account relationships (shipments to sales to COGS to inventory) Bills of lading, production counts, meter reads, claim inventories Drivers themselves are incomplete, or the conversion factor (price, yield) is too rough
Existence or occurrence of recorded items Analytics are usually a weak primary test; aging or unusual-item scans can focus the sample Shipping logs, subsequent receipts, confirmations as details High-level ratios cannot prove a recorded item is real
Accuracy / valuation Margin analysis, aging, NRV trends, standard-to-actual rate variances Price lists, aging files, commodity curves Need details to test individual valuations or clerical accuracy
Cutoff Period-to-period spikes around period end; shipping versus billing lag Shipping dates versus invoice dates as a population Spike only scopes the issue; details test the boundary
Fraud or abuse indicators Unexpected relationships, round-dollar or split-item patterns, after-hours postings, duplicate keys Access logs, vendor master, payment file Ratios of AP turnover versus last year will not find duplicate payments; details and targeted interrogation must follow
Reasonableness Tests
A reasonableness test is the technique CIA items use when the objective is completeness or overall reasonableness of a total. You build an independent estimate of what the recorded amount should be, using data that does not come solely from the same recording process, and you compare that estimate to the book. Classic pattern: independently obtained shipment quantities × average selling price ≈ recorded sales. Other patterns: occupied rooms × average daily rate ≈ room revenue; production units × standard yield ≈ recorded output; badge hours × contractual rates ≈ recorded contract labor; metered usage × tariff ≈ recorded utilities.
Three quality tests decide whether the reasonableness test can actually achieve the objective:
Independence of the driver. If you estimate sales from the same invoice file you are testing, you have a circular procedure. Pull quantities from the warehouse-management or transportation system, from bills of lading, or from a third-party logistics feed.
Stability of the conversion factor. Average price works if mix is stable or if you stratify. A blended average that hides a shift from high-price to low-price SKUs will miss incompleteness or bury overstatement.
Precision versus the threshold. If your estimate can only be good within plus or minus 20 percent, you cannot treat a 8 percent book-to-estimate gap as comfort, and you also cannot treat it as a finding. You do not have a sharp enough instrument. Either improve the model (better drivers, stratification) or move to tests of details.
Related-account relationships are a cousin of reasonableness tests. Sales, COGS, inventory, and freight should move in economically coherent ways. A jump in recorded sales with flat shipments, falling inventory that does not match throughput, or freight that does not move with cartons is an unexpected relationship. Coherence is not proof of completeness, but incoherence is a signal that completeness, cutoff, or classification may have failed.
Precision, Residual Risk, and When Analytics Are Insufficient
Analytics give assurance only to the extent the expectation is precise and independent . A high-assurance objective on a material account with a noisy expectation is a mismatch. In that case the correct technique is not "more ratio analysis." It is tests of details : vouching, tracing, confirmation, inspection, reperformance, and recalculation directed at the population that remains risky.
Treat this as a decision rule, not a vibe:
If the objective needs only a reasonableness view, and the expectation is tight, and the difference is inside the threshold, analytics can be the primary procedure for that objective.
If the difference exceeds the threshold , management's explanation must be corroborated with evidence. Vague commentary ("it was a busy quarter") is not corroboration.
If the explanation is uncorroborated, or only explains part of the gap, analytics are insufficient . Design tests of details for the unexplained portion.
If you never could build a precise expectation—missing driver data, unstable mix, unreliable system reports—do not pretend the procedure ran. Go to details (or change the approach and document why).
If the objective is fraud, existence of specific recorded items, or accuracy of individual transactions, start with the assumption that high-level ratios will scope work, not finish it.
North Hub again. Objective A: efficiency of outbound freight. Technique: freight dollars per shipment versus a flexible budget and versus South Hub. Independent inputs: shipment counts from the transportation-management system. If actual is $448,000 versus a $460,000 expectation, analytics can support a no-exception conclusion on efficiency for this procedure.
Objective B: completeness of recorded sales at the same company. Technique: shipments × average selling price, stratified by product family. If the estimate is $18.4 million and recorded sales are $16.1 million, a 12.5 percent shortfall against a 4 percent threshold is not something you "ratio away." Investigate cutoff, unbilled shipments, and missing invoice runs. If inquiry does not close the gap with evidence, trace from shipping documents to recorded sales (a test of details in the completeness direction).
Objective C: whether duplicate vendor payments indicate fraud. AP turnover versus last year can be perfectly stable while duplicates still flow. The matching technique is targeted anomaly work (same invoice number, same amount different invoice, split payments just under approval limits) plus details on the hits. Choosing AP turnover because it is "an analytical procedure" fails B5b: the technique does not achieve the objective.
On the exam, when the stem asks which procedure best achieves the objective, name the technique that could detect the failure mode. When it asks what to do after an unexplained difference, do not lower the criteria, do not call it fraud from the ratio alone, and do not accept an uncorroborated story. Say that analytics are insufficient and move to tests of details.