Information Systems and Data Management
What this domain tests
This domain concerns how organisations turn data into reliable information. Candidates should understand the data lifecycle: capture, validation, storage, processing, reporting and disposal. Expect questions on database models (relational and non-relational), primary and foreign keys, normalisation, data warehousing, and the distinction between transactional (OLTP) and analytical (OLAP) systems. Familiarity with master data, metadata and data quality dimensions (accuracy, completeness, timeliness, consistency, validity, uniqueness) is essential.
Governance and quality
Data governance assigns ownership and accountability. A data owner is accountable for a data domain; a data steward handles day-to-day quality. Understand policies, standards and the role of a data dictionary or catalogue. Recognise the tension between accessibility and control: more sharing raises value but also risk. Privacy concepts such as purpose limitation, minimisation and retention limits belong here too.
Traps that appear in questions
- Assuming more data is always better; relevance and quality matter.
- Confusing a data warehouse (structured, curated, historical) with a data lake (raw, varied, schema-on-read).
- Treating normalisation as universally good; denormalisation can aid reporting performance.
- Forgetting that poor input validation propagates errors downstream (garbage in, garbage out).
- Mixing up backup copies with the system of record.
Worked example / rule of thumb
A sales report shows two different totals for the same month. Before blaming the tool, check the definition: are returns netted, is the currency converted at transaction or period rate? Reconciling definitions in the data dictionary resolves most discrepancies. Rule of thumb: if two numbers disagree, first compare definitions, then compare source systems, then compare timing.
Exam technique
Read the scenario for the business purpose: operational processing or decision support. That single distinction usually eliminates half the options. Where a question asks for the 'best' control, prefer preventive validation at the point of entry over detective reconciliation later.