Replacing the legacy finance core

Customer billing

A large customer base, its own regulatory obligations and its own release cycle. It stays, and the receivable becomes an interface rather than a sub-ledger. This is the largest single scope decision on the programme and it should be made once, in writing.

Works and asset management

The maintenance system keeps the work orders and the physical asset register. Finance takes the costs and owns the financial asset register. The reconciliation between the two registers is in scope even though neither system is.

Payroll and HR

Out of scope for this release, in scope as an interface. The payroll journal and the net pay file both land in the finance core, so 'not changing payroll' still means testing payroll.

The operational technology estate

No financial transaction originates there. Named here only so that it is named, because it is the assumption most often left unstated.

Inventory

The model flags it as a prerequisite this scope leans on. If stores are staying where they are, that is a decision; if the ledger will take stock movements from a system nobody is changing, that interface belongs on the list below.

Revenue postings, receivable balances and cash applied←Customer billing system

Today's interface posts a summary the ledger cannot decompose. If it is rebuilt at the same grain, regulatory revenue reporting stays a spreadsheet exercise. This is the interface to design first, not to port.

Statements in, ABA and direct entry files out↔Banking platform

Format, testing window and the bank's own change lead time are on the critical path and are outside the programme's control. Book it early.

Period journal in, net pay file out↔Payroll system

Not changing payroll does not mean not testing payroll. A failed payroll journal is a month that does not close, and a failed pay file is a pay day that does not happen.

Labour, plant and materials costs; capitalised project costs; asset register alignment←Works and asset management system

The highest-value and highest-risk interface on the programme. It carries most of the capital spend and it is the point where the financial and physical asset registers are supposed to agree.

Statutory accounts, periodic returns, regulatory cost allocation→Departmental and regulator reporting

Currently assembled from extracts by a small number of people. The replacement either fixes this or inherits it, and inheriting it is a choice that should be made deliberately.

Supplier invoices in, remittance advice out↔Suppliers, e-invoicing and portals

The chance to move non-PO spend onto a purchase order is here, and it closes once the new process is live and the workarounds have set.

Trial balance, cash position and dimensional extracts→Data warehouse and reporting

Every downstream report is written against the old chart of accounts. Chart changes are reporting changes, and the effort belongs in this programme rather than in the reporting team's backlog.

Payroll journal←Payroll

General Ledger · per pay run · without it: Labour cost is missing from the period and the month cannot close.

Trial balance extract→Business Intelligence

General Ledger · daily · without it: Reporting runs on stale numbers.

Cash position and forecast→Business Intelligence

Cash Management · daily · without it: Treasury decisions are made on yesterday's balance.

Payroll journal→General Ledger

Payroll · per pay run · without it: Labour cost is missing from the period.

Net pay file→ABA

Payroll · per pay run · without it: Staff are not paid.

Capitalised project costs←Project Costing

Asset Management · monthly · without it: Capital work in progress never becomes an asset.

Asset register alignment←Asset Lifecycle Management

Asset Management · continuous · without it: The financial register and the physical register describe different assets.

Customer orders←Sales Force Automation

Order Processing · continuous · without it: Orders are rekeyed and differ from what was agreed.

Stock movements→General Ledger

Inventory Management · continuous · without it: Inventory value in the ledger stops matching the warehouse.

Reporting extracts←General Ledger

Business Intelligence · daily · without it: Management reporting diverges from the ledger.

Regulated and non-regulated activity

Where an economic regulator sets prices, cost has to be allocated by regulated service and the allocation has to be defensible line by line rather than reconstructed in a spreadsheet afterwards. That is a chart of accounts and cost allocation design decision, and it is the one that most constrains every other posting decision. Making it after the chart is built is the expensive order.

Two asset registers that are both correct

The statutory register carries fair value under the applicable financial reporting standards; a regulatory asset base carries a different valuation on a different roll-forward for price setting. They are not a reconciliation error — they are two answers to two questions. The system has to produce both without either being maintained by hand.

Capital is the main event

Most of what an asset-intensive organisation spends is capital. Capitalisation policy, treatment of overheads, work in progress ageing and the point at which a project becomes an asset are finance design decisions with an operational input, and they determine whether the capital programme can be reported at all.

Assets that arrive without an invoice

Contributed infrastructure and customer contributions arrive as assets and revenue with no supplier invoice and no cash receipt. They need a designed path in, or they become a year-end journal nobody can evidence.

Public sector reporting obligations

Reporting under the jurisdiction's financial management legislation, the model financial report, periodic returns to Treasury and audit by the auditor-general are fixed inputs. They shape the close calendar, the evidence the system must retain, and the go-live window — a cutover that lands across statutory reporting is a cutover that will be moved.

Procurement probity

Government purchasing policy, social procurement obligations and the associated disclosure apply to the buying process regardless of what the finance system does. The system has to evidence them, not just permit them.

Australian payroll and banking obligations

Single Touch Payroll, superannuation guarantee and ABA payment files apply even where payroll is out of scope, because the files and the journals still land here. Allowances under an enterprise agreement are the part most likely to be assumed rather than tested.

Does the receivable stay in the billing system, or move?

It decides the size of the programme. Moving it brings the whole customer base and its regulatory obligations with it; leaving it makes the revenue interface the most important design artefact on the programme.

By: Before scope is baselined — it is not a design decision, it is the scope. · Accounts Receivable

What is the chart of accounts, and does it carry the regulatory allocation natively?

Every posting decision, every report and every interface depends on it. A chart designed for statutory reporting alone leaves the regulatory submission where it is now.

By: Before configuration starts. Chart changes after the first sub-ledger posts are the most expensive change a finance programme makes. · General Ledger

Which system is master for the asset register, and for which attributes?

The financial register and the physical register will not agree unless somebody decides which one is right about what. Left undecided, the reconciliation becomes a permanent manual activity.

By: Before the works interface is specified. · Asset Management

How far does procurement change?

Payables inherits whatever procurement hands it. Three-way matching cannot be delivered without purchase orders and receipts, and a matching engine with no receipting discipline produces blocked invoices and manual workarounds.

By: With scope, because it determines whether supply chain is a stakeholder or a workstream. · Procurement

Does stores stay where it is?

The model flags inventory as a prerequisite this scope depends on without including. Either it is staying and the interface is named, or it is moving and the scope is larger than stated.

By: Before the first release is planned. · Inventory Management

Where does the cutover sit against the statutory calendar?

A cutover landing near a statutory reporting date, a regulatory submission or year-end will be moved, late and expensively. The calendar is fixed; the cutover is not.

By: At mobilisation, because it constrains everything else. · Cutover and Go Live

Who backfills the finance team?

The subject matter experts this programme needs are the people who close the month. Without funded backfill, the programme runs on evenings and the close slips, and both fail slowly.

By: Before mobilisation, with the budget. · Digital Transformation Readiness Checklist

Revenue posts from billing at a grain the regulatory view cannot use.

So: The regulatory submission stays a manual assembly, and the largest single benefit of the programme is not delivered.

Instead: Design the revenue interface against the regulatory reporting requirement first, and prove it in the foundation release rather than at user acceptance.

The works and capitalisation interface is treated as a technical build.

So: Capital work in progress is wrong from go-live, and the capital programme cannot be reported in its first year.

Instead: Own it as a joint finance and asset design, with both register owners named and a reconciliation designed before the interface is specified.

Data migration starts when the programme needs it rather than when it needs to start.

So: Opening balances that nobody can reconcile to the legacy ledger, discovered at cutover rehearsal.

Instead: Cleansing begins before mobilisation, with quality measured and an owner per data domain; reconciliation totals agreed in advance of the first trial load.

Segregation of duties is designed around the team that exists.

So: Conflicts are built into the role design and become an audit finding the system now enforces.

Instead: Build the conflict matrix from business risk, then design roles against it; where separation is genuinely impossible, document the compensating control rather than leaving it silent.

Customisation is rebuilt because it exists.

So: The estate's unupgradeability is carried into the new system on day one.

Instead: Every customisation carried forward states its business reason and is assessed against standard behaviour; the ones whose reason no longer exists are dropped deliberately.

Change is funded as training.

So: A technically successful go-live where the finance team runs the old process through the new system and the benefits do not appear.

Instead: Change resourced from mobilisation with an accountable business owner, and adoption measured after go-live rather than assumed.

An eighteen-month programme needs a sponsor who will still be accountable at the end of it, and a succession position if they will not.

The most common reason a finance replacement slips. Named people, named percentages, funded backfill — or the programme is resourced on goodwill.

Supplier and asset master duplication and completeness, measured rather than estimated, with an owner per domain and cleansing already under way.

Including the scheduled extracts and spreadsheets nobody owns. The catalogue is the honest measure of what replacing this system actually costs.

What was changed, and why. Without the why, every customisation is either rebuilt or dropped on a guess.

Measure decision latency before mobilisation. A programme needs decisions faster than business as usual, not slower.

Close dates, reporting deadlines, audit and the regulatory submission — plotted before the cutover date is chosen rather than after.