In this topic 6 sections
Start with the decision and its information domains
A useful information search begins with a decision: what must change, who can act and what evidence could alter the choice? The accounting system supplies part of the answer. Operating records, staff experience, project documents and external conditions may explain why the financial result occurred and whether it will continue.
Internal information comes from within the organisation, including budgets, maintenance logs, customer records and managers' observations. External information includes official statistics, supplier information, industry research and competitor disclosures. Neither domain is automatically superior. Internal records can contain errors or incentives to misreport; an external source may be authoritative but poorly matched to a particular business.
Search across functions before commissioning new collection. Purchasing may already hold delivery data that operations needs, while customer support may reveal problems absent from the sales ledger. Record the source, purpose, coverage, collection date and responsible owner so another person can assess suitability.
Concept sources: ABS, Quality Management of Statistical Outputs Produced From Administrative Data, March 2011, introduction — administrative sources can support validation, comparison and richer linked information.
Choose an aggregation method that answers the question
Aggregation turns separate observations into a view that a manager can use. Five useful approaches need different checks.
| Analytical move | Original application | Check before relying on it |
|---|---|---|
| Assemble a picture from partial evidence | Estimate service demand from enquiries, bookings and local population data. | Identify assumptions and gaps; proxies are not direct observations. |
| Triangulate independent evidence | Compare reported delays with timestamps and customer feedback. | Several reports copied from one source are not independent confirmation. |
| Create a new measure | Divide energy use by completed processing units. | Match the numerator and denominator to the same period and scope. |
| Summarise observations | Calculate totals, averages, indices or ratios across locations. | Check weighting, exceptional items and differences hidden by averages. |
| Place different measures together | Present cost, completion time and safety incidents in one review. | Retain each measure's meaning instead of forcing unlike units into a total. |
Integration is the work that makes these combinations meaningful. Reconcile identifiers, definitions, currencies, units and reporting periods. Investigate conflicting totals and duplicated records before presenting a neat consolidated result. A shared customer code helps link systems, but it does not establish that every system defines an active customer identically.
Concept sources: ABS, administrative-data quality guidance, March 2011, validation and data-linking uses.
The balanced scorecard links financial results with customer, internal-process and learning-and-growth measures. Cascading relevant measures through organisational levels and comparing actual results with targets connects local activity to strategy.
Concept sources: Kaplan, Conceptual Foundations of the Balanced Scorecard, 2010, four perspectives and organisational alignment.
Preserve detail behind the summary
Granularity is the level represented by each record: one job, one site-day or one region-month, for example. Choose the level needed for the decision. Executives may need trends across regions; a supervisor investigating delays needs individual work orders.
Keep a traceable route from summary to underlying records where authorised. Drill-down helps explain an exception and reproduce a calculation. More detail does not automatically mean greater accuracy: thousands of incorrectly coded transactions remain incorrect. Excess detail can also obscure a decision or expose personal information.
- Define the question
- Identify sources and common definitions
- Check and link records
- Aggregate at the required level
- Explain exceptions
- Retain an authorised path to supporting detail
Distinguish the dimensions of a source
Classify information along separate dimensions rather than treating the labels as substitutes. Internal or external describes organisational origin. Financial or non-financial describes content. A margin percentage is financial information even though it is not expressed in dollars; cost per completed inspection combines financial and operational inputs.
Primary and secondary describe the relationship to collection and use. Primary data is newly collected for the study's purpose; secondary data already exists for another purpose. An internally held payroll dataset reused for workforce research can therefore be secondary data. Ask who collected the information, why, how and with which limitations. Ownership alone does not establish reliability.
Concept sources: ABS, Data sources, released 2 February 2023, primary and secondary collection. This later conceptual clarification qualifies the guide's simplified association of primary with internal and secondary with external.
Protect information when combining and sharing it
Limit access to the records required for a person's work. Apply approval, secure storage and audit trails to sensitive source data and exported reports. Combining harmless-looking datasets can reveal sensitive facts that neither dataset exposes alone.
Aggregation can reduce disclosure risk, but small groups may still identify individuals. Removing names is not enough when unusual characteristics or other available information enable re-identification. Consider the audience and access environment as well as the released fields.
Concept sources: NIST SP 800-53 Revision 4, April 2013, updated January 2015, AC-6 and the AU control family; OAIC, De-identification and the Privacy Act, re-identification and access-environment guidance. These support control concepts, not a complete statement of current legal obligations.