One finance leader's journey from spreadsheets to a single finance ledger

Sector

FMCG distribution and wholesale

Footprint

Multiple warehouses across northern and southern Vietnam

Finance team

A small central team plus accountants attached to each site

Systems in place

An existing accounting system, GDT e-invoicing, several bank portals and a lot of Excel

The finance director did not arrive with a technology problem. They arrived with a calendar problem: the close consumed the first ten working days of every month, and by the time the numbers were trusted they described a period that had already gone.

• Supplier invoices arrived by email, by post and by hand, in three formats and two languages.
• Goods receipts lived with the warehouses, so matching meant a phone call before it meant a posting.
• Customer credit limits sat in a spreadsheet that only one person maintained.
• Aging was rebuilt each month from bank statements, so the collections list was always describing the past.
• Every question from the board started with several days of preparation.

The work was sequenced deliberately. Nothing was switched off before its replacement had run in parallel for a full close.

1. 1

   Chapter 1. Agree what a clean record looks like

   Before any automation, the team settled the definition of a supplier, a customer and a tax code. Duplicate master records were the root cause of most matching failures, and no capture tool would have solved that on its own.

2. 2

   Chapter 2. Capture, then match

   Supplier invoices were captured into BancoOS and matched against the purchase order and the goods receipt. Partial deliveries and multi warehouse receipts were handled as line level matches rather than being escalated to a person by default.

3. 3

   Chapter 3. Move approvals off email

   Approval thresholds were written into the workflow rather than into a policy document. The audit trail became a by-product of the work instead of something assembled afterwards.

4. 4

   Chapter 4. Bring receivables onto the same ledger

   Customer records carry payment terms and a credit limit, invoices move from draft to issued through GDT, receipts are recorded and allocated against specific invoices, and the allocation ledger is the single answer to what a customer owes.

5. 5

   Chapter 5. Let the numbers be derived

   Aging, overdue days and the collections worklist are calculated from the ledger. The monthly spreadsheet that used to produce them was retired once two closes had matched.

The payables changes were about removing handoffs, not about replacing the accounting system.

• Invoices are captured once, in whatever format they arrive.
• Three way matching runs against the purchase order and the warehouse goods receipt.
• Exceptions are routed by reason, so a price variance and a missing receipt do not sit in the same queue.
• Approvals carry the evidence with them, so the reviewer is not searching for the attachment.
• Posting to the existing accounting system stays intact, with the GL treatment unchanged.

Distribution runs on credit, so the customer ledger mattered as much as the supplier one. This is the half most projects leave for later.

• Customers are verified by tax code, so the same buyer is not held under three records.
• Payment terms and credit limits live on the customer record and are enforced at invoicing.
• Invoices move from draft to issued through GDT e-invoicing rather than a separate portal.
• Receipts are allocated to specific invoices, so a part payment is visible at invoice level.
• Aging and the collections worklist are derived, so the team chases from live numbers.

The finance director describes the change as a shift in where their attention goes, not as a headcount decision. The team is the same size.

• The close is a review of exceptions rather than a reconstruction of the month.
• Collections is a daily worklist owned by named people.
• Board questions are answered from the ledger in the meeting.
• New warehouses join the same process instead of inventing a local one.

• Fix master data first. Every downstream automation inherits its quality.
• Run the old process in parallel for one full close. Trust is built by matching, not by promises.
• Do not treat receivables as phase two. The two ledgers share customers, tax codes and bank data.
• Give exceptions an owner. A queue without a name attached becomes a backlog.
• Judge the project on where the team's hours go, not on a single headline number.

Most of what is described here came out of sitting with finance teams in warehouses and back offices rather than in a product meeting. If that is the kind of work you want to do, we hire for it.

See open roles at BancoOS