Mopane Group:
four ERPs, one company
Mopane Group is invented. So are its customers, suppliers and every number. But its problem is real for most groups we work with: four companies in four countries, each on its own ERP, and no single place where the group exists as one set of figures.
Sources, makes and sells to retailers
Building and industrial supplies. South Africa imports from Asia and buys locally, makes three own-brand lines in a Durban plant, and supplies its sister companies across the border. Most sales go to retail chains and independent hardware stores.
| Company | Currency | ERP | How data arrives | Sales, 24 months |
|---|---|---|---|---|
| Mopane Distribution South Africa · head office, plant | ZAR | SAP Business One 10.0 (HANA) | Service Layer, hourly | R136.3m |
| Mopane Namibia | NAD | SAP Business One 9.3 (SQL Server) | SQL view export, every 4 h | R40.7m |
| Mopane Zambia | ZMW | SAP Business One 10.0 (partner cloud) | Service Layer, twice a day | R33.7m |
| Mopane Botswana | BWP | Sage 300 2024 | Nightly file over SFTP | R30.5m |
Figures on this page are read from the demo warehouse, PostgreSQL on our server, to September 2026. Three different releases of the same ERP, one different ERP, three currencies besides the rand, and data arriving anywhere from 1 to 26 hours late. Each system keeps its own item, customer and supplier codes.
Five tabs, five questions
Each tab answers one question a different person in the group asks. Charts filter each other: click a bar and the rest of the tab follows. Tables sort when you click a column heading.
"How is the group doing?"
For the CFO and the board. Four ledgers converted to rand at monthly rates, with sales between group companies removed, as consolidation requires.
- Sales by company by month
- Gross margin by product category
- Stock and dead stock (nothing sold from that branch in 180 days) by company
- Debtors ageing: 8.6% of the book is over 90 days
"This is the group as one company. None of the four ERPs can show you this on its own."
TryClick one company's bar in the stock or debtors chart. The other charts on the tab filter to that company. Click it again to clear.
"Are we buying well, and does it arrive?"
For buyers and the supply chain team. Purchases by supplier type, on-time delivery by supplier, an average 69 days from purchase order to receipt for imports, and stock moving between sister companies.
"A third of imports land late at Durban. And the stock sent to Zambia is in South Africa's books as sold before Zambia has booked it as bought: about R113k sat in that gap at the end of September."
TryIn the supplier-lane table at the bottom, sort by on-time %, then compare lead times for the worst three.
"What is the plant really costing us?"
For the plant manager and finance. SAP Business One production orders from the Durban plant: 153,142 units made, 2.5% scrapped (PVC pipe the worst at 3.9%), 88% of orders closed on time, and actual cost per unit against the standard cost the price list is built on.
"Resin prices are creeping up, and the price list still assumes the old standard cost. That's margin leaking on every length of pipe."
TryIn the two trend charts, click the other products in the legend to hide them and leave PVC pipe on its own.
"Which customers actually make us money?"
For sales and credit control. Retailers are 68% of sales. Every retail customer, across all four companies, with sales growth, margin and how much of what they owe is over 90 days.
"Your biggest customers are your thinnest margins, and here is who pays late."
TryIn the customer table, sort by over-90 debt, then by margin. See who is on both lists.
"Can we trust these numbers?"
For group IT and finance. Whether each system's feed arrived and how late, the rows that couldn't be mapped (with their value), and a data-quality scorecard per company: unmapped items, duplicate customers, stock below zero, a feed outside the 24-hour target.
"Zambia and Botswana sell solar kit under their own item codes. Until someone maps them, those sales don't exist at group level. This tab tells you exactly what to fix, and what it's worth."
TryFind Botswana in the feed and scorecard tables. The Sage feed is the slowest, and it has the most to map.
What's actually running
PostgreSQL warehouse
Staging tables shaped like each ERP's own (SAP B1 invoice and production-order tables, Sage 300's numeric dates), a star schema in rand, and the views the dashboards read. Rejected rows are kept with their value.
Apache Superset
29 charts on one tabbed dashboard, with click-to-filter between charts. Superset connects to the warehouse with a read-only database user.
Locked down
The demo sits behind Cloudflare Access. The guest login is read-only and sees only the Mopane data: no SQL editor, no editing, no other databases.
Three steps to the live dashboard
The demo isn't public. We open it to people we're talking to, so we can walk you through it and answer questions as you go.
Ask us
Email [email protected] from your work address. Tell us which systems you run. We add your email to the demo's guest list.
Prove it's you
Go to bi.2nth.app and enter that email. Cloudflare sends you a one-time code. No password to remember, and the session lasts 8 hours.
Sign in to Superset
Use the read-only guest login we send you. You land on the Mopane Group dashboard. Start with Overview.
Before you ask
Is any of this real data?
No. Mopane Group, its subsidiaries, customers and suppliers are invented, and every number is generated. The structure is real: the tables, the system differences and the problems are the ones we find in groups like yours.
Do we have to replace our ERPs?
No. This works alongside the systems you run today and only reads from them. Many groups do it before an ERP migration, because the mapping work is the same and the dashboards are useful straight away.
Which systems can you connect?
Anything we can read from: SAP Business One, Sage 300 and Sage 50, ERPNext, Xero, Pastel, a SQL database, or a nightly file export. If your system can produce a report, it can usually feed the warehouse.
Where would our data live?
In a PostgreSQL database on your own server, in your cloud account, or hosted by us. It's an ordinary PostgreSQL database, so your team can query it directly and you can take it with you.
What does it cost?
There's no licence fee for Superset or PostgreSQL, and no charge per user. You pay for the build and for running it, which depends on how many systems and how much history. Tell us what you run and we'll put a proposal together.
Can we brand it, or put dashboards in our own portal?
Yes. Superset can be themed, and dashboards can be embedded in your own website or intranet, with each viewer seeing only their own rows.
Want the tour?
We'll open the demo for you, walk through the five tabs on a call, and talk about what your version would look like.