The money layer

Everything the other apps do eventually turns into an accounting entry. Get this layer wrong and every report in the company is wrong, usually quietly, and usually discovered at the audit.

You do not need to be an accountant to configure it. You do need to understand what the accountant is going to check.

Documents become entries

This is the mental model to carry into every project. When a client says a report is wrong, the question is never “how do we fix the report”. It is: which document produced this entry, and which piece of configuration sent it to that account?

Before — on paper 8 steps

1. The accountant prints the bank statement.
2. Sales invoices are in one Excel file, purchases in another, petty cash in a notebook.
3. Each payment is ticked against an invoice with a pen.
4. The warehouse is asked for a stock value and sends a figure from a count taken three weeks ago.
5. Journal entries are typed by hand into the accounting program.
6. The trial balance does not agree. Two days go on finding a transposed digit.
7. The tax return is compiled by reading every invoice again.
8. The figures reach management on the twentieth of the following month.

Three weeks of reconstruction, and figures that arrive too late to act on.

After — in Odoo 4 steps

1. The bank statement is imported and most lines match themselves.
2. The stock value is already in the ledger, posted by each movement as it happened.
3. The tax report is produced from invoices that were already recorded.
4. Exceptions are reviewed, the period is locked, and the figures go out in the first week.

The close becomes a review of what the system already recorded, instead of a rebuilding of it.

What to learn, in order

The chart of accounts. Install the localization for the country first — it brings a chart, taxes and usually statutory reports. Adjust it; do not build one from nothing.

Journals. Sales, purchase, bank, cash and miscellaneous. They control document numbering and the default accounts. Setting up journals badly produces numbering the auditor will not accept, and renumbering later is painful.

Taxes. How the amount is computed, whether it is included in the price, and which accounts it posts to. Then fiscal positions, which swap taxes and accounts automatically by customer type — domestic, export, free zone, exempt. Fiscal positions are the tool that stops clients asking for a custom module.

Payments and reconciliation. Registering payments, matching them to invoices, importing bank statements and reconciling. Your client will do this every day, so it must be smooth.

Stock valuation. The connection back to stage 04. Costing method — standard, FIFO or average — and whether valuation is manual or automated. With automated valuation, stock movements post to the ledger by themselves. This is where “stock does not match the accounts” is prevented, and the setting lives on the product category.

Analytic accounting. A second dimension for reporting — by project, branch, cost centre, vehicle, contract. When a client asks to see profitability per something, this is usually the answer, not new accounts and not a custom report.

Localization. Egypt ETA e-invoicing, UAE VAT, and whatever your market requires. Read the country page before agreeing to build anything.

The questions to ask every finance client

Ask these in discovery, and write the answers down:

• How many legal entities, and do they trade with each other?
• Which currencies, and how are exchange differences handled?
• What is your costing method today, and does anyone actually rely on it?
• How long does the close take now, and what is the slowest part?
• Which statutory reports must come out of the system?
• Who is allowed to post, and who is allowed to change a posted document?
• What did the auditor complain about last year?

The last one is the highest-yield question in the whole discovery process, and almost nobody asks it.

Where beginners go wrong

• Building a chart of accounts by hand when a localization exists.
• Taxes without fiscal positions, then handling exports manually forever.
• Leaving valuation manual by default and discovering the gap at the first close.
• Creating dozens of accounts to get a report that analytic accounting would have produced.
• Ignoring lock dates, so users edit last year’s postings.
• Not involving the client’s accountant early. They will be asked to sign off. Bring them in during design or they will refuse in month five, and they will be right to.

Practise this

1. Install a country localization on a fresh database and look at what arrived — accounts, taxes, reports.
2. Set up a tax included in the price and one added on top, and compare the invoice.
3. Build a fiscal position that zero-rates an export customer, and test it.
4. Switch a product category to automated valuation with FIFO, receive goods, deliver them, and find every entry produced.
5. Import a bank statement and reconcile it against invoices.
6. Set up two analytic accounts for two branches, post to both, and produce a profit figure per branch.
7. Close a period, set a lock date, then try to post into it.

Exercise 4 is the one worth repeating until you can predict the entries before you look at them. It is the exact point where operations and finance meet, and it is where the difference between a good consultant and an average one shows.