The end of sampling: testing 100% of transactions
Audit sampling was a sensible answer to paper records. With every transaction in a database, testing all of them is often faster than choosing which ones to test. Here is what changes when you do.
Sampling made sense when every voucher was a piece of paper. Checking all of them was impossible, so auditors checked a few and used statistics to reason about the rest.
Records are no longer on paper. When every transaction sits in a database, testing all of them takes a query, not a warehouse of clerks.
What sampling misses
A sample is designed to estimate how common a problem is. It is not designed to find the rare, expensive exception: the one duplicate payment, the one journal posted at midnight, the one vendor sharing a bank account with an employee. Those are exactly the items an audit committee asks about.
What full-population testing gives you
- Every exception, not an estimate. You see each item that breaks a rule, with its value.
- Less time per audit. The tests run in minutes; people spend their time on the exceptions.
- Comparable results. The same test runs every quarter, so you see trends.
- Stronger evidence. A finding backed by every transaction is hard to argue with.
Tests that work on the whole population
| Area | Example tests |
|---|---|
| Purchase to pay | Duplicates, split purchase orders, invoices before orders |
| Order to cash | Price overrides, credit notes after period end, unusual discounts |
| Journals | Weekend and late-night postings, round amounts, unusual users |
| Payroll | Ghost employees, duplicate bank accounts, pay after leaving |
| Access | Segregation of duties conflicts, dormant users with activity |
What doesn't change
Judgement. A test flags an exception; a person decides whether it matters. Full-population testing removes the arithmetic from auditing, not the auditor. It also needs clean data: knowing which fields to trust is part of the work.
Where to start
Pick one process with high volume and clear rules, such as purchase to pay. Run the tests above over a full year. Compare the result with the last sampled audit of the same process. The difference is usually the best argument for doing it everywhere.
The Audit Suite runs these tests out of the box, and our audit effort work moves your audit plan from samples to populations, one process at a time. For the idea behind it, see Sampling is a habit, not a method.