Audit data analytics: what it is and where it helps
What counts as audit data analytics, the procedures it suits best, when sampling is still the better tool, and how a firm can start on one client.
By Achekam · · 4 min read · For Accountants & audit firms in South Africa, Nigeria, Kenya
"We should be doing more data analytics" is said in many audit firms in Johannesburg, Lagos and Nairobi, usually at a planning meeting and usually without anyone defining what it means. For some it means a fancy dashboard; for others, a spreadsheet filter on the general ledger. Both can be useful. Neither is automatically audit evidence.
Before a firm invests time in analytics, it helps to know what the term covers, which procedures it genuinely improves, and where traditional sampling remains the better choice. This article answers those questions with examples a small or mid-sized firm can use.
A working definition
As reported by ACCA, the IAASB defines data analytics for audit as the science and art of discovering and analysing patterns, deviations and inconsistencies, and extracting other useful information in the data underlying or related to the subject matter of an audit, through analysis, modelling and visualisation, for the purpose of planning and performing the audit.
Three parts of that definition are worth holding on to. It is about the data underlying the audit, not the summarised figures. It looks for patterns and exceptions, not just totals. And it serves planning and performing the audit, which means it should lead somewhere: a risk identified, an item selected, a conclusion supported.
Procedures that suit analytics
ACCA gives examples of audit procedures performed with data analytics:
| Procedure | What the analysis does |
|---|---|
| Net realisable value of inventory | Compares the last purchase and last sale of each item, and the price, to find items selling below cost |
| Revenue trends | Analyses revenue by product and region to show unusual movements |
| Three-way matching | Matches purchase orders to invoices and payments |
| Segregation of duties | Uses transaction metadata to show who created and approved entries |
What these have in common is a large population of similar transactions and a clear rule for what counts as an exception. A distributor with thousands of inventory lines is a far better candidate for the first procedure than a consultancy with no inventory.
Looking at every item
ACCA explains that data analytics often makes it possible to test every item in a population, improving the coverage of audit procedures and reducing or eliminating sampling risk.
That is the real attraction. A sample of invoices tells you about the sample; an analysis of every invoice tells you about all of them, provided the data is complete and the test is well designed. The exceptions it produces still need following up, and a test that flags too many items to investigate has simply moved the work rather than reduced it.
Fraud risk is not only a data problem
COSO states that its Fraud Risk Management Guide, Second Edition, expands information on data analytics while continuing to emphasise the importance of interviewing and whistleblower systems.
That balance is a useful corrective. Analytics can show that journals were posted at odd hours; it cannot tell you why. The conversation with the financial controller, and the tip from an employee, remain part of how fraud is found.
Whole-journal extracts and risk-scored populations in Achekam
Two parts of Achekam put this into practice.
Journal entry testing. Staff load a whole-population journal extract as a CSV file and map its columns, and Achekam flags entries such as manual, weekend, after-hours and post-close postings for staff to select and test. The CSV is not stored; its hash and row count are. A badge on that screen reading "Halo-Style" is only a label. The approach to journal testing itself is covered in journal entry testing.
Audit Intelligence. Achekam risk-scores a loaded population and proposes a sample. A reviewer who did not prepare the plan approves it before it is drawn, and every draw can be reproduced from its stored seed. The methods available are random, systematic, monetary unit, specific items and all items, and no sample size is proposed without approved materiality and a confidence level.
The "all items" option is where analytics and sampling meet: when the population and the test allow it, the firm can choose to examine everything.
When sampling is still the better tool
Analytics is not always the answer. Sampling remains sensible when:
- the evidence needed is physical or external, such as inspecting an asset or obtaining a confirmation;
- the population is small enough that a well-chosen sample covers it efficiently;
- the client's data cannot be extracted completely or reliably;
- the test needs judgement on each item that no rule can capture.
Choosing between methods is covered in more depth in choosing a sampling method.
Starting small on one client
A practical way in is one client and one procedure, rather than a firm-wide programme. A sensible first project:
- Pick a client with a clean, complete general ledger and a cooperative finance team.
- Choose one procedure, such as journal entry testing or three-way matching.
- Agree with the client how the data will be extracted and how completeness will be checked.
- Run the test, follow up every exception, and document what was done.
- Review afterwards what it cost, what it found and what a sample would have found.
The lessons from that first client decide whether, and where, analytics earns a place in the firm's methodology.
Read ACCA's article
ACCA's technical article on data analytics and the auditor considers how auditors use analytics, including testing whole populations to reduce sampling risk, and the matters to weigh when doing so. Circulate it before the next planning meeting where someone proposes "more analytics", so the discussion starts from a shared definition.
General guidance only, not advice on your own circumstances. Tax and payroll rules change: check the current position with the authority or your adviser.
