E-commerce Operations · Retail & E-commerce · Accounting & Bookkeeping · Payments

High-Volume E-Commerce Reconciliation and Accounting Automation

Anonymised engagement

Client: Accounting firm · Bulgaria · month-end for an online retailer with 1,000-1,500 orders a month
Industry: E-commerce Operations
Department: Finance & Accounting
Region: Bulgaria

At 1,500 orders a month, month-end stops being an accounting task and becomes a throughput problem. Card batches, cash collected by couriers, returns crossing the fiscal year: every extra order meant more manual work, until the month closed late. Now the whole month reconciles and books itself in under two minutes.

A working model of the approach on sample data. It plays on its own and pauses when scrolled away; nothing here touches a live system.

Month-end close · 1,500 orders

2026-09

12 import files, ~2,300 lines, under two minutes.Sample data · in production for 8 months · 98% auto-matched, 100% accurate on what is booked

The challenge

Our client is a Bulgarian accounting firm. The business behind this automation is one of its clients: a fast-growing online retailer selling through its own web store, with 1,000 to 1,500 orders a month.

Money arrives in every way an online shop gets paid. Card and wallet processors settle a day of orders in a single batch. Transfers come through Wise and the bank. National courier networks collect cash on delivery and remit it days later against waybill reports. Returns generate credit notes and offsets that have to find their original order, sometimes in the previous fiscal year. The books live in AJUR, a legacy Bulgarian accounting system that accepts data only through its own import files.

What it was costing them

Work that scaled with orders, not with the fee. Every sale added matching and booking time for a small accounting firm.
A close that broke in peak months. Campaigns and seasons doubled the volume exactly when the close was due.
Cash on delivery, where money goes missing. Courier remittances that do not tie to waybills are easy to lose in the volume.
Returns across the year-end. A credit note that cannot find its order distorts two fiscal years.

Every way an online shop gets paid, one ledger

The approach

We built an unattended month-end automation on a self-hosted n8n platform. The accountant enters the period in a short form; everything else runs by itself.

Every source, one pass

The sales register, the payment processors, the bank statements, both courier settlement reports, credit notes and offsets are collected, normalised and matched in a single run. n8n orchestrates systems that were never designed to talk to each other so they behave as one pipeline.

Matching that mirrors how the money moves

Money movement	How it is closed
Card and wallet processor batches	Orders matched against the settlement that paid for them
Courier cash on delivery	Remittances reconciled against the waybill payment reports
Credit notes	Matched back to the order they reverse, across the fiscal year where needed
Offsets	Resolved against the balances they clear

Accounting output, not a spreadsheet

A full set of AJUR import files, one per source and entry type: turnover, credit notes, offsets and a combined file. Written in AJUR's exact block structure, field separators, operation codes and encoding.
A process report giving every transaction its status and its match.
Account mapping and posting logic in Excel files the accountant owns. No code change to adjust a rule.

Nobody has to remember what was already done

Run ledger. Every delivered file is recorded.
Duplicate-run guard. Before delivering, the automation checks whether any source file has changed since the last batch. A re-run with nothing new finishes in seconds, delivers nothing and emails to say so.
Closing the loop. Each run sends a run report and a completion email with the period, the scope and the counts.

What each run reports

A migration proven, not promised

The automation originally ran on an earlier workflow platform. We moved it to self-hosted n8n as a verbatim port and verified the output byte for byte against reference runs from the old platform before switching over. The port surfaced and fixed two latent defects in the original logic along the way.

Technology

Layer	What we used
Orchestration	n8n, self-hosted
Sources	Stripe, PayPal, Wise and bank exports; courier waybill and cash-on-delivery reports; sales register
Workspace and configuration	Google Drive with n8n service connectors, Excel
Accounting system	AJUR, through its native import format
Control	Run ledger, duplicate-run guard, email reporting

The outcome

	Before	After
Time to close a month	Days of skilled work	Under 2 minutes
Sale records matched	By hand	98% automatically
Accuracy of what is booked	Checked line by line	100%
Re-running a period	Another few days	Seconds, and free

A full month of up to 1,500 orders reconciled and booked in under two minutes: 12 import files and about 2,300 accounting lines.
98% of sale records matched automatically, with 100% accuracy on what is booked. The remaining 2% are custom bank transactions with no description to match on, and they arrive as a short review list.
Eight months in production.
Around 192 hours a year of manual reconciliation removed on the firm's own conservative estimate, with room for twice that.
Re-running changed how the team works. Because a re-run is safe and free, they run the month early, look at the exceptions, fix the source data and run it again.
The framework now serves more than one of the firm's clients, each new build reusing a substantial part of the first.

Closing a month of up to 1,500 orders

The risk most teams miss

Order volume grows, new payment or courier channels are added, and a manual close falls a little further behind every month. Nobody notices until year-end. Automation that scales with volume removes the ceiling before it is hit.

Built with

The tools this ran on, by the layer each one carried.

Next step

Start the same conversation

Tell us the process that keeps breaking. We will come back with what it would take to fix it, what it would cost, and what we would measure.

Book a working session See ways to start

Or write to us directly: •••••@•••••••.•••

A reply from a partner within one business day
Fixed scope and fixed fee, agreed before we start
The maps and the backlog are yours either way