The structure every financial report is built on — and the one design decision that is genuinely hard to reverse.
A chart of accounts is the structured list of accounts used to classify financial transactions. Modern ERPs usually implement it as segments — company, cost centre, natural account, project, and so on — combined into a code. The natural account says what happened; the other segments say where and why.
Nearly every reporting limitation traces back to the chart of accounts. If cost cannot be analysed by product line, it is usually because no segment captures product line. Adding one later does not populate history, so comparatives break. This is why it is designed before configuration and why rushing it is expensive in a way that is not obvious at the time.
Company, cost centre, natural account at minimum. Product, project, region or channel where the business manages by them.
Detailed enough to report, coarse enough to post consistently. If posters cannot choose reliably, the detail is false precision.
Not every combination is valid. Cross-validation rules stop meaningless codes being created.
Roll-ups for reporting, defined so statutory and management views both derive from one structure.
Who may create an account, and on what basis. Ungoverned charts grow accounts nobody can explain.
A business creates separate expense accounts for each department: Salaries — Sales, Salaries — Production, Salaries — Admin, and so on. With eight departments and forty expense types, the chart runs to several hundred accounts.
It works, and it scales badly in a specific way. Adding a ninth department means creating forty new accounts. Reporting salaries across the business means summing eight accounts. Any analysis that crosses the two dimensions requires either a hierarchy that mirrors every combination or a spreadsheet.
The structural answer is to make department a separate segment rather than part of the natural account. Forty expense accounts and a department segment express the same information, and adding a ninth department becomes one new segment value. This decision is made at implementation and is painful to reverse afterwards, because changing it means remapping history. It is worth more design time than it usually gets — the question to ask of every proposed account is whether it describes what was spent or who spent it.
| Term | What it means |
|---|---|
| Natural account | What was spent — salaries, rent, materials. The core segment. |
| Segment | A dimension of the account code, such as cost centre, department or location. |
| Account hierarchy | The rollup structure used for reporting, which can differ from the posting level. |
| Control account | An account whose balance is maintained by a sub-ledger rather than by direct posting. |
| Statutory mapping | The link from your accounts to the format required by regulators. |
Hundreds of near-identical accounts get posted to inconsistently, which destroys the comparability the detail was meant to give.
Reorganisations then require renumbering. Use a separate segment.
Invalid combinations accumulate and have to be cleaned up at year end.
The users of the output should shape the structure, not only the people posting to it.
| Metric | What it tells you | What distorts it |
|---|---|---|
| Accounts actually posted to | Unused accounts are clutter that makes coding errors more likely. | |
| Postings to suspense or miscellaneous | Coding difficulty. High volumes mean the chart does not fit the business. | |
| Reclassification journals | Corrections after posting, which usually indicate ambiguous account definitions. | |
| Depth of reporting hierarchy | Very deep hierarchies tend to indicate dimensions that should be segments. |
We build and support the systems this page describes. Start with a free working proof of concept — no cost, no commitment.