DAMA CDMP Data Quality Specialist Quick Review
Quick Review for the DAMA International DAMA CDMP Data Quality Specialist (CDMP Quality) exam, with high-yield concepts, traps, and practice guidance.
Quick Review purpose
This Quick Review is for candidates preparing for the DAMA International DAMA CDMP Data Quality Specialist, exam code CDMP Quality. Use it to refresh the most testable Data Quality ideas before moving into IT Mastery practice, topic drills, mock exams, and detailed explanations.
The exam is likely to reward candidates who can do more than define terms. You should be able to distinguish similar concepts, choose appropriate quality controls, connect data quality to governance and business value, and reason through realistic data issue scenarios.
High-yield Data Quality mindset
Data quality is not simply “clean data.” It is the degree to which data is fit for its intended business purpose. The same data may be acceptable for one use case and unacceptable for another.
Core decision rule
Ask three questions first:
- What business process or decision uses the data?
- What quality requirement makes that use possible?
- Where in the data lifecycle should the issue be prevented, detected, corrected, or monitored?
If a question asks for the “best” answer, prefer the option that connects data quality to business use, measurable rules, ownership, and sustainable controls rather than one-time cleansing.
Essential vocabulary
| Term | What it means | Common exam trap |
|---|---|---|
| Data quality | Fitness of data for intended use | Treating quality as absolute instead of contextual |
| Data quality dimension | A category used to evaluate data quality | Assuming every dimension applies equally to every data set |
| Data quality rule | A testable statement of expected data condition | Confusing a business rule with its technical implementation |
| Data profiling | Analysis of data content, structure, patterns, and anomalies | Treating profiling as remediation rather than discovery |
| Data cleansing | Correcting, standardizing, or improving data values | Assuming cleansing solves root causes |
| Data validation | Checking whether data conforms to defined rules | Assuming valid data is always accurate |
| Data stewardship | Accountability for data meaning, quality, and use | Confusing steward responsibility with IT ownership of systems |
| Data owner | Business role accountable for data decisions and priorities | Treating ownership as purely technical custody |
| Data custodian | Role responsible for technical operation, storage, or protection | Confusing custody with business accountability |
| Data quality scorecard | Periodic reporting of metrics and thresholds | Reporting metrics without decisions or action plans |
| Root cause analysis | Investigation of why defects occur | Stopping at symptom correction |
Data quality dimensions
Different frameworks use slightly different dimension names. Focus on the concept and the business question each dimension answers.
| Dimension | Core question | Example | Trap |
|---|---|---|---|
| Accuracy | Does the data correctly represent the real-world object or event? | Customer date of birth matches verified source | Accuracy often requires comparison to an authoritative reference |
| Completeness | Is required data present? | Mandatory tax ID is populated | A field can be complete but wrong |
| Validity | Does the value conform to format, domain, or rule? | Country code is in approved ISO list | Valid does not always mean accurate |
| Consistency | Is the data the same across systems, records, or time? | Customer status is consistent in CRM and billing | Inconsistency may reveal timing, definition, or integration issues |
| Uniqueness | Is each real-world entity represented once? | No duplicate customer profiles | Duplicate detection often needs matching logic, not exact equality |
| Timeliness | Is the data available when needed? | Daily sales feed arrives before reporting cutoff | Timely data can still be inaccurate |
| Currency | Is the data up to date for its intended use? | Address reflects latest known residence | Currency depends on update expectations |
| Conformity | Does the data follow required standards or patterns? | Phone numbers use standard format | Conforming values may still be semantically wrong |
| Integrity | Are relationships and dependencies preserved? | Every order has a valid customer ID | Referential integrity is narrower than overall data integrity |
| Reasonableness | Does the value make sense in context? | Employee age is plausible | Requires business context and thresholds |
| Precision | Is the level of detail appropriate? | Coordinates recorded to required decimal places | More precision is not always better |
| Accessibility | Can authorized users obtain data when needed? | Analysts can access approved data set | Accessibility must be balanced with security and privacy controls |
Dimension distinctions candidates often miss
Validity vs. accuracy
A value can be valid but not accurate.
- Valid:
99999is a five-digit postal code format. - Not accurate: it is not the customer’s actual postal code.
Choose validity when the issue is about format, allowed values, or rule conformance. Choose accuracy when the issue is about truthfulness against reality or a trusted source.
Completeness vs. coverage
Completeness usually asks whether required values are present in records that exist. Coverage asks whether the population itself is sufficiently represented.
Example:
- Customer records have email populated: completeness.
- All active customers are included in the data set: coverage.
Timeliness vs. currency
Timeliness is about availability by the required time. Currency is about whether the value is up to date.
Example:
- Yesterday’s file arrived before 8:00 a.m.: timely.
- The address in the file is three years old: not current.
Consistency vs. integrity
Consistency compares values across places or contexts. Integrity often concerns structural relationships and constraints.
Example:
- CRM says customer is active; billing says inactive: consistency issue.
- Order record references a nonexistent customer ID: integrity issue.
Data quality lifecycle
Data quality management is continuous. It is not a single clean-up project.
flowchart LR
A[Define business need] --> B[Identify critical data]
B --> C[Define quality requirements]
C --> D[Create rules and metrics]
D --> E[Profile and assess data]
E --> F[Prioritize issues]
F --> G[Remediate data and root causes]
G --> H[Implement controls]
H --> I[Monitor scorecards]
I --> C
Lifecycle exam logic
| Step | Main purpose | Best evidence |
|---|---|---|
| Define business need | Clarify why quality matters | Business process, decision, risk, or outcome |
| Identify critical data | Focus effort where value or risk is highest | Critical data elements, key reports, regulatory or operational dependencies |
| Define requirements | Translate need into expectations | Business definitions, thresholds, rules |
| Profile data | Discover actual condition | Patterns, null rates, outliers, duplicates |
| Measure quality | Quantify performance against rules | Metrics, scorecards, trend lines |
| Analyze root cause | Find why defects occur | Process gaps, system constraints, unclear definitions |
| Remediate | Correct existing data and causes | Cleansing, process change, control improvement |
| Monitor | Sustain quality over time | Dashboards, alerts, ownership, escalation |
Data quality requirements and rules
A strong data quality rule is specific, testable, tied to business meaning, and owned.
Weak vs. strong rules
| Weak statement | Better data quality rule |
|---|---|
| Customer data should be good | Active customer records must have a non-null customer type |
| Order dates should make sense | Order date must not be later than shipment date |
| Product codes should be valid | Product code must exist in the approved product reference table |
| Duplicate customers should be avoided | No two active customer records may share the same verified national ID |
| Data should be updated quickly | Trade records must be available in the reporting warehouse within the defined business cutoff |
Rule types
| Rule type | What it checks | Example |
|---|---|---|
| Domain rule | Value belongs to allowed set | Status is Active, Inactive, Pending, or Closed |
| Format rule | Value follows pattern | Email contains required structure |
| Range rule | Value falls within limits | Discount is between 0 and 100 percent |
| Mandatory rule | Required value is present | Policy number is not null |
| Cross-field rule | Values are logically compatible | End date is not before start date |
| Referential rule | Related record exists | Invoice references a valid customer |
| Uniqueness rule | Entity is not duplicated | One active employee ID per employee |
| Derivation rule | Calculated value matches formula | Total equals sum of line amounts plus tax |
| Temporal rule | Timing relationship is valid | Effective date precedes expiration date |
| Conditional rule | Requirement applies under conditions | Cancellation reason required when status is Cancelled |
Data profiling review
Data profiling helps reveal what is actually in the data before defining or refining controls.
Common profiling outputs
| Profiling output | What it reveals |
|---|---|
| Null count or null percentage | Completeness issues |
| Distinct value count | Cardinality, possible code values, uniqueness clues |
| Frequency distribution | Unexpected values, dominant values, outliers |
| Minimum and maximum | Range issues |
| Pattern analysis | Format inconsistency |
| Cross-column analysis | Logical conflicts |
| Duplicate analysis | Uniqueness and entity resolution issues |
| Referential analysis | Broken relationships |
| Outlier detection | Potential anomalies requiring investigation |
Profiling traps
- Profiling finds symptoms; it does not automatically determine root cause.
- A surprising value is not always an error. It may be a legitimate business exception.
- Profiling without business context can produce misleading conclusions.
- Technical profiling should be paired with metadata, definitions, lineage, and process knowledge.
- Do not remediate solely because a value is rare. Confirm business rules first.
Measurement and metrics
Data quality metrics should be understandable, repeatable, actionable, and tied to thresholds.
Common metric structures
| Metric | Plain-language calculation |
|---|---|
| Completeness rate | Number of records with required value / number of applicable records |
| Validity rate | Number of records passing rule / number of records tested |
| Defect rate | Number of failed records / number of records tested |
| Duplicate rate | Number of duplicate records / number of records assessed |
| Timeliness rate | Number of deliveries meeting cutoff / number of expected deliveries |
| Accuracy rate | Number of verified accurate records / number of verified records |
| Issue aging | Time since issue was opened or detected |
| Remediation rate | Number of resolved issues / number of opened issues in period |
Metric quality checklist
A useful data quality metric should have:
- A clear business definition.
- A defined population.
- A clear numerator and denominator.
- An owner.
- A measurement frequency.
- A threshold or target.
- An escalation path.
- A link to business impact.
- A way to distinguish severity, trend, and recurrence.
Data quality scorecards and dashboards
Scorecards communicate quality status. They should support decisions, not just display numbers.
| Scorecard element | Why it matters |
|---|---|
| Data domain or data set | Shows scope |
| Critical data element | Focuses attention on important data |
| Dimension | Clarifies type of quality issue |
| Rule tested | Makes measurement repeatable |
| Threshold | Defines acceptable performance |
| Current score | Shows present condition |
| Trend | Shows improvement or deterioration |
| Severity | Helps prioritize response |
| Owner | Enables accountability |
| Action plan | Connects measurement to remediation |
Scorecard trap
A dashboard with many metrics but no ownership, thresholds, or action process is weak data quality management. For exam scenarios, the stronger answer usually links metrics to accountability and continuous improvement.
Issue management
Data quality issues should be logged, triaged, assigned, investigated, resolved, and monitored for recurrence.
Issue management workflow
- Detect issue through profiling, control failure, user report, audit, or monitoring.
- Log the issue with evidence, affected data, and business impact.
- Classify by dimension, severity, domain, and affected process.
- Assign ownership to appropriate business and technical roles.
- Analyze root cause rather than only correcting visible records.
- Remediate data where appropriate.
- Fix process or control to prevent recurrence.
- Validate resolution with retesting.
- Monitor recurrence through ongoing metrics.
Severity decision factors
| Factor | Higher severity when… |
|---|---|
| Business impact | Decisions, revenue, operations, or customers are affected |
| Compliance or risk exposure | Required reporting or risk controls depend on the data |
| Scope | Many records, systems, or processes are affected |
| Criticality | Critical data elements are involved |
| Recurrence | Issue repeats after prior fixes |
| Detectability | Issue is hard to detect before impact |
| Timeliness | Issue affects urgent reporting or operations |
Root cause analysis
Root cause analysis separates symptoms from causes.
Common root causes
| Root cause category | Examples |
|---|---|
| Process design | Data captured too late, no verification step, unclear handoff |
| System design | Missing validation, weak constraints, poor interface design |
| Integration | Transformation error, mapping mismatch, timing conflict |
| Metadata | Ambiguous definition, inconsistent code meanings |
| Governance | No owner, no standard, unresolved policy conflict |
| Training | Users misunderstand required values or procedures |
| Incentives | Speed rewarded over accuracy, no accountability for defects |
| Source quality | External feed or upstream system sends poor data |
| Change management | New field, product, or process not reflected in rules |
| Manual workarounds | Spreadsheet edits, rekeying, informal corrections |
Root cause vs. remediation
| Situation | Symptom fix | Root cause fix |
|---|---|---|
| Null values in required field | Populate missing values | Make field mandatory at capture and train users |
| Invalid product codes | Replace invalid codes | Align source system dropdown with approved reference data |
| Duplicate customer records | Merge duplicates | Improve matching at onboarding and stewardship review |
| Late data feed | Reload file manually | Define SLA, monitoring, alerts, and escalation |
| Conflicting definitions | Reconcile report values | Approve business glossary definition and align transformations |
Prevention, detection, and correction controls
Data quality controls can be placed at different points in the lifecycle. Prevention is usually preferable when feasible, but detection and correction remain necessary.
| Control type | Purpose | Example | Best used when |
|---|---|---|---|
| Preventive | Stop defect creation | Required field, dropdown list, referential constraint | Data can be validated at entry or integration |
| Detective | Identify defects after creation | Profiling rule, reconciliation, exception report | Defects may occur despite controls |
| Corrective | Repair defects | Cleansing, standardization, deduplication | Existing data must be fixed |
| Compensating | Reduce risk when ideal control is not available | Manual review of high-risk records | System change is not immediate |
| Monitoring | Track quality over time | Dashboard, scorecard, trend alert | Ongoing assurance is needed |
Control selection rule
Prefer the control closest to the source of defect creation if it is practical, business-aligned, and does not create unacceptable process friction.
Data cleansing and standardization
Cleansing improves existing data, but it should not be confused with sustainable quality management.
Cleansing techniques
| Technique | Use |
|---|---|
| Parsing | Split values into components, such as full name into first and last name |
| Standardization | Convert values to consistent format |
| Normalization | Reduce variation in representation |
| Validation | Confirm values meet rules |
| Enrichment | Add missing or improved values from trusted sources |
| Matching | Identify records referring to same entity |
| Deduplication | Remove or merge duplicate records |
| Survivorship | Choose best value among duplicates |
| Correction | Replace wrong values with verified values |
Cleansing traps
- Cleansing without business rules can corrupt data.
- Deduplication can incorrectly merge different real-world entities.
- Enrichment may introduce licensing, trust, lineage, or currency issues.
- Automated correction should be monitored, especially for high-risk data.
- Cleansing projects should feed lessons back into process and control improvement.
Matching, deduplication, and survivorship
Duplicate management is a common scenario area because it requires judgment.
| Concept | Meaning |
|---|---|
| Exact matching | Records match on identical values |
| Fuzzy matching | Records are similar enough to be potential matches |
| Deterministic matching | Rules-based matching using defined fields |
| Probabilistic matching | Statistical likelihood that records refer to same entity |
| Match threshold | Score above which records are considered matches |
| Clerical review | Human review of uncertain matches |
| Survivorship rule | Rule for choosing which value remains after merge |
| Golden record | Consolidated representation assembled from trusted sources |
Survivorship examples
| Rule | Example |
|---|---|
| Most recent | Use latest address update |
| Most trusted source | Use verified government ID source |
| Most complete | Keep record with most populated fields |
| Source priority | CRM overrides web form for customer name |
| Manual stewardship | Steward decides for ambiguous high-impact records |
Trap
A “golden record” is not automatically true or permanent. It depends on matching logic, source trust, governance decisions, and ongoing updates.
Reference data, master data, and data quality
Data quality issues often arise from poor control of reference or master data.
| Data type | Quality relevance |
|---|---|
| Reference data | Provides valid code sets and classifications |
| Master data | Represents key business entities such as customer, product, supplier, or employee |
| Transaction data | Records business events and depends on accurate master/reference data |
| Metadata | Defines meaning, rules, lineage, and context |
Examples
| Issue | Likely related area |
|---|---|
| Invalid country code | Reference data management |
| Same customer appears three times | Master data management |
| Revenue report uses unclear “active customer” definition | Metadata and governance |
| Orders reference missing product IDs | Referential integrity and master data |
| System A and System B use different status codes | Reference data alignment and integration |
Metadata and business glossary
Metadata is essential for data quality because it clarifies meaning, lineage, rules, and responsibility.
Metadata that supports data quality
| Metadata type | Example |
|---|---|
| Business definition | What “active customer” means |
| Data owner | Business accountable role |
| Data steward | Role responsible for quality coordination |
| Valid values | Approved code set |
| Data lineage | Source-to-target flow |
| Transformation rule | How source field is converted |
| Quality rule | Expected condition |
| Threshold | Acceptable quality level |
| Security classification | Access and handling requirement |
| Retention information | How long data is kept according to policy |
Exam trap
If a scenario shows inconsistent reporting because teams define the same term differently, the best answer is usually not “clean the database.” It is to resolve definitions, governance, metadata, and lineage.
Governance and accountability
Data quality requires business accountability and cross-functional coordination. IT can implement controls, but business stakeholders define quality requirements and acceptable thresholds.
Typical responsibilities
| Role | Typical responsibility |
|---|---|
| Data owner | Accountable for data domain decisions, priorities, and acceptable quality |
| Data steward | Coordinates definitions, rules, issue resolution, and quality monitoring |
| Data custodian | Operates technical environment and implements technical controls |
| Data governance council | Resolves cross-domain priorities, standards, and escalations |
| Data quality analyst | Profiles data, develops metrics, investigates issues |
| Business process owner | Ensures processes capture and use data correctly |
| Data architect | Designs structures and integration patterns that support quality |
| Application owner | Supports system-level validation and workflow changes |
Accountability trap
Do not assign all data quality responsibility to IT. Technical teams often implement solutions, but data quality requirements, definitions, priorities, and acceptance criteria must be business-led.
Data quality and the data lifecycle
Quality can be affected at every stage.
| Lifecycle stage | Quality concerns |
|---|---|
| Creation or capture | Input validation, user training, source controls |
| Acquisition | Supplier quality, external feed checks, contract expectations |
| Integration | Mapping, transformation, reconciliation, timing |
| Storage | Constraints, referential integrity, metadata |
| Usage | Fit for purpose, interpretation, access, reporting logic |
| Sharing | Standard definitions, formats, security, lineage |
| Archiving | Retention, historical integrity, accessibility |
| Disposal | Controlled deletion and auditability where relevant |
High-yield idea
The earlier a defect is prevented, the less costly it usually is to fix. However, the best exam answer should still consider feasibility, business impact, and process design.
Data quality in analytics and reporting
Analytics failures often stem from poor definitions, inconsistent transformations, incomplete populations, or stale data.
| Problem in report or model | Likely data quality issue |
|---|---|
| Totals differ between dashboards | Definition, lineage, transformation, or timing inconsistency |
| Model performs poorly for a customer segment | Coverage, completeness, bias, or representativeness issue |
| Report includes inactive products | Reference/master data or filter logic issue |
| Trend line changes unexpectedly | Source change, late-arriving data, or transformation change |
| KPI cannot be reconciled | Lack of lineage, unclear metric definition, aggregation mismatch |
Candidate mistake
Do not assume every reporting discrepancy is a data warehouse defect. It may originate in source systems, business definitions, integration timing, transformation logic, or report filters.
Data quality and data integration
Data movement can create or reveal quality defects.
Integration quality checks
| Check | Purpose |
|---|---|
| Record counts | Confirm expected volume moved |
| Control totals | Reconcile numeric totals |
| Hash totals or checksums | Detect changes or transfer errors |
| Referential checks | Confirm relationships remain valid |
| Domain checks | Confirm valid codes after transformation |
| Mapping validation | Confirm source-to-target logic |
| Exception handling | Capture rejected or suspect records |
| Latency monitoring | Confirm timeliness |
| Lineage documentation | Show how values were produced |
Trap
A successful file load does not mean the data is fit for purpose. Technical completion and data quality are related but different.
Data quality costs and business value
Data quality initiatives should be justified by business impact, risk reduction, efficiency, and trust.
Cost categories
| Cost type | Example |
|---|---|
| Prevention cost | Training, validation controls, standards |
| Appraisal cost | Profiling, monitoring, audits |
| Internal failure cost | Rework, manual correction, process delays |
| External failure cost | Customer impact, incorrect reporting, operational failure |
Business value examples
- Reduced operational rework.
- More reliable reporting.
- Improved customer experience.
- Better regulatory or risk reporting support.
- Reduced duplicate processing.
- Faster integration and migration.
- More trustworthy analytics.
- Improved process automation.
Prioritization
Not all data defects deserve equal effort. Prioritize based on criticality, impact, and feasibility.
Prioritization matrix
| High impact? | Easy to fix? | Typical action |
|---|---|---|
| Yes | Yes | Fix quickly and monitor |
| Yes | No | Escalate, plan remediation, implement compensating controls |
| No | Yes | Fix if low effort and no adverse effects |
| No | No | Defer, monitor, or accept risk |
Better prioritization considers
- Critical data elements.
- Business process dependency.
- Customer or stakeholder impact.
- Risk and control implications.
- Number of affected records.
- Severity of errors.
- Frequency and recurrence.
- Root cause complexity.
- Remediation cost.
- Availability of preventive controls.
Critical data elements
A critical data element is important enough that poor quality can materially affect business outcomes, risk, reporting, or operations.
How to identify critical data
Look for data used in:
- Key business decisions.
- Executive or regulatory reporting.
- Customer-facing processes.
- Financial calculations.
- Risk models or controls.
- Operational workflows.
- Master data relationships.
- Integration keys.
- Contractual or service obligations.
- High-volume automation.
Trap
A field is not critical merely because it exists in a database. Criticality comes from business use and impact.
Data quality strategy
A mature data quality program is proactive, governed, measured, and continuously improved.
Strategy components
| Component | Purpose |
|---|---|
| Scope and priorities | Focus on highest-value domains and data |
| Governance model | Define accountability and decision rights |
| Standards | Promote consistent rules, definitions, and controls |
| Metrics and thresholds | Make quality measurable |
| Issue management | Provide repeatable resolution process |
| Tooling | Support profiling, rules, monitoring, and remediation |
| Communication | Build awareness and transparency |
| Training | Improve data capture and stewardship behavior |
| Continuous improvement | Reduce recurrence and mature controls |
Maturity perspective
Data quality maturity generally progresses from reactive correction to proactive prevention and optimization.
| Maturity level | Characteristics |
|---|---|
| Ad hoc | Issues fixed manually when noticed |
| Reactive | Cleansing projects respond to recurring problems |
| Managed | Rules, owners, and issue processes exist |
| Measured | Metrics, scorecards, and thresholds guide action |
| Optimized | Prevention, automation, governance, and continuous improvement are embedded |
Exam logic
If a scenario asks how to improve maturity, prefer actions that institutionalize ownership, standards, measurement, and root-cause prevention over isolated clean-up.
Common scenario patterns
Scenario: many missing values
Best response depends on why values are missing.
| Likely cause | Better response |
|---|---|
| Field not required at entry | Add validation or workflow requirement |
| Users do not know value | Provide training or change process |
| Value not applicable | Adjust rule to account for applicability |
| Source system does not capture value | Change source process or source mapping |
| Optional data being treated as mandatory | Revisit business requirement |
Scenario: duplicate customers
Good answer usually includes:
- Define matching criteria.
- Profile duplicates.
- Establish survivorship rules.
- Assign stewardship review for ambiguous matches.
- Merge or link records carefully.
- Improve capture and matching controls at onboarding.
- Monitor duplicate rate.
Scenario: inconsistent report numbers
Good answer usually includes:
- Compare business definitions.
- Review lineage and transformations.
- Check timing and refresh cycles.
- Reconcile source-to-target counts and totals.
- Identify authoritative source.
- Establish governed KPI definition.
Scenario: invalid reference codes
Good answer usually includes:
- Align code sets and definitions.
- Validate against approved reference data.
- Fix source mappings.
- Add interface controls.
- Establish owner for reference data changes.
- Monitor exceptions.
Scenario: recurring defects after cleansing
Good answer usually includes:
- Perform root cause analysis.
- Fix upstream process or system controls.
- Clarify ownership and rules.
- Add monitoring.
- Avoid another isolated cleansing-only project.
Data quality tools
Tools support data quality work, but they do not replace governance or business rules.
| Tool capability | Supports |
|---|---|
| Profiling | Discovery of patterns and defects |
| Rule engine | Automated validation |
| Data cleansing | Standardization and correction |
| Matching and deduplication | Entity resolution |
| Monitoring | Ongoing scorecards and alerts |
| Metadata management | Definitions, lineage, rules |
| Workflow | Issue tracking and stewardship tasks |
| Reference data management | Controlled code lists |
| Master data management | Golden records and entity consistency |
| Data observability | Pipeline and anomaly monitoring |
Tooling trap
Selecting a tool before defining business requirements, ownership, and rules is usually not the strongest answer.
Data quality dimensions by example
Use this table for rapid classification practice.
| Example defect | Most likely dimension |
|---|---|
| Required customer email is blank | Completeness |
Email is abc.example.com | Validity or conformity |
| Email belongs to someone else | Accuracy |
| Customer has three active profiles | Uniqueness |
| Customer ID on order does not exist | Integrity |
| System A shows Gold tier, System B shows Silver tier | Consistency |
| Daily feed arrives after reporting deadline | Timeliness |
| Address has not been updated after verified move | Currency |
| Date of birth is 01/01/1800 | Reasonableness or validity, depending on rule |
| Amount is rounded to whole dollars when cents are required | Precision |
| Report users cannot access approved data in time | Accessibility |
Formulas worth knowing conceptually
You do not need to over-memorize formulas, but you should understand the structure of basic quality rates.
\[ \text{Validity Rate} = \frac{\text{Records Passing Validation Rule}}{\text{Records Tested}} \]\[ \text{Defect Rate} = \frac{\text{Records Failing Quality Rule}}{\text{Records Tested}} \]\[ \text{Completeness Rate} = \frac{\text{Applicable Records With Required Value Present}}{\text{Applicable Records}} \]Key exam point: define the applicable population carefully. A completeness denominator should exclude records where the field is genuinely not applicable.
Common traps on the CDMP Quality exam
Trap 1: choosing cleansing when governance is needed
If the issue is recurring, cross-system, or definition-based, cleansing alone is insufficient.
Trap 2: confusing technical validity with business correctness
A value can pass a system edit and still be wrong for the business.
Trap 3: ignoring root cause
Sustainable data quality requires preventing recurrence, not only fixing defective records.
Trap 4: treating all data equally
Focus on critical data elements and business impact.
Trap 5: assuming IT owns data quality
IT implements many controls, but business ownership and stewardship are central.
Trap 6: measuring without acting
Metrics are useful only when connected to thresholds, accountability, and remediation.
Trap 7: over-validating at the wrong point
Too many entry controls can slow business processes or create workarounds. Place controls thoughtfully.
Trap 8: using dimensions mechanically
Some issues can involve multiple dimensions. Choose the dimension most directly tested by the scenario.
Trap 9: trusting external data blindly
Third-party or external data requires assessment, contracts or expectations, monitoring, and lineage.
Trap 10: overlooking metadata
Definitions, lineage, and rules are often the missing link in data quality scenarios.
Fast decision rules
Use these when answering scenario questions quickly.
| If the question emphasizes… | Think first about… |
|---|---|
| Incorrect real-world value | Accuracy |
| Missing required value | Completeness |
| Wrong format or invalid code | Validity or conformity |
| Same entity appears multiple times | Uniqueness and matching |
| Broken parent-child relationship | Integrity |
| Different values across systems | Consistency |
| Late arrival | Timeliness |
| Outdated value | Currency |
| Conflicting report meanings | Metadata and governance |
| Repeated issue after fixes | Root cause and preventive controls |
| Unclear responsibility | Data ownership and stewardship |
| Too many defects in source data | Upstream controls |
| Data migration failure | Profiling, mapping, reconciliation, cleansing |
| Poor dashboard trust | Definitions, lineage, controls, and metrics |
| Question asks “best long-term solution” | Governance, root cause, prevention, monitoring |
Mini review: from issue to action
| Issue type | Immediate action | Sustainable action |
|---|---|---|
| Missing mandatory fields | Identify affected records | Add capture control and ownership |
| Invalid codes | Reject or correct exceptions | Govern reference data and mappings |
| Duplicate entities | Match and merge carefully | Improve onboarding and matching controls |
| Late feeds | Alert and reload if needed | Define service expectations and monitoring |
| Inconsistent definitions | Reconcile current outputs | Establish glossary and governed KPI |
| Bad source extract | Quarantine and investigate | Add source validation and contract expectations |
| Transformation error | Correct mapping | Add testing and lineage documentation |
| Untrusted report | Validate calculations | Govern metrics and certify data source |
What to practice next
After reviewing these concepts, move into IT Mastery practice for the DAMA International DAMA CDMP Data Quality Specialist (CDMP Quality) exam. Start with topic drills on data quality dimensions, profiling, rules, issue management, governance roles, and root-cause scenarios. Then use original practice questions and mock exams with detailed explanations to test whether you can choose the best action in realistic data quality situations.
Continue in IT Mastery
Use this Quick Review as a final concept map, then move into IT Mastery for focused topic drills, mixed practice sets, timed mock exams, and detailed explanations. The practice questions are original IT Mastery practice items; they are not official DAMA questions, copied live-exam content, or exam dumps.