A forecast for each SKU
Statistical models and machine learning models compete on a backtest of each SKU. The champion must beat a naive and a seasonal-naive benchmark before Planner trusts it.
South Africa
Miridia Planner forecasts each SKU from its own history, plans the stock for each location, and turns the plan into purchase orders. It is part of Miridia, the supply chain and automation platform, and it is priced in rand.
Last updated
Who it is for
Most South African companies plan demand in a spreadsheet: last year, plus a percentage. That method fails in three places. It misses the December peak when the peak moves, it treats a month of stockouts as a month of low demand, and it cannot tell which of the changes that people make helped.
The usual alternatives are an enterprise planning platform, which is priced and implemented for a large company, or a planning add-on that needs an ERP below it. Miridia Planner is built for the company in between. It plans the demand, and the rest of Miridia raises the purchase orders, receives the stock, and runs production from the plan.
Miridia Planner
Statistical models and machine learning models compete on a backtest of each SKU. The champion must beat a naive and a seasonal-naive benchmark before Planner trusts it.
Planner proposes corrections for outliers, stockouts, and gaps, and a planner approves each one. The raw history always stays, so a correction can be undone.
Add growth, promotions, and manual changes for each SKU. Sales, finance, and operations then agree one forecast, and each cycle is saved as a snapshot.
Test a price change or a promotion with elasticity before you commit to it. A scenario that you like becomes part of the plan.
Lead-time demand, safety stock, reorder points, days of cover, and order quantities rounded to the minimum order of each supplier, on order sheets grouped by supplier.
Planner measures forecast value added at each layer, and it archives each forecast run. You see which changes helped and which made the forecast worse.
Integrations
Planner reads your sales history from the channels that Miridia connects to, or from a CSV file that you upload. After the plan, the purchase orders and transfers run in Miridia.
| Sales history | Shopify, WooCommerce, Amazon, and the other online stores. Yoco, Shopify POS, and Square. |
|---|---|
| Upload | A validated CSV upload, for history from a system that Miridia does not connect to. |
| Accounting | Sage Cloud, Xero, QuickBooks, and Zoho Books. Sage Evolution, Odoo, and others connect through consulting services. |
| Automation | A reorder signal, a stockout risk, or a published consensus can start a workflow in n8n, Zapier, or Power Automate. |
If you use a planning add-on on top of an ERP today, compare the two approaches before you decide.
New to the terms? The glossary explains safety stock, forecast value added, and MAPE and WAPE with worked examples.
Local conditions
A South African retailer sells most in December, when a northern-hemisphere business is in its winter quiet months. A forecast tool that ships with seasonal settings for Europe or the United States starts from the wrong pattern. Planner does not start from a pattern. It learns the seasons of each SKU from your own order history, and each model must beat a seasonal-naive benchmark on your data.
Lead times are the second local problem. When imported stock waits at the port for extra weeks, the variation in the lead time can matter more than the variation in demand. Planner calculates safety stock from the demand during the lead time, and it recalculates the reorder points when the forecast changes. For a store network, two-echelon replenishment shows what stock each branch needs, and what to transfer before you buy.
Trust
Pricing
Miridia Planner is an advanced module. You add it to a Miridia plan, and it is priced separately. Ask us for a price on the discovery call.
The Starter plan is R3,000 a month, and yearly billing saves 10%. It includes orders, inventory, purchasing, bills of material, work orders, integrations, workflows, the full API, and the Ask Your Data chat, for up to 2,500 orders a month, 3 active integrations, and 10 team members.
We prefer that you learn this now, and not in the third month.
Demo
The demo runs on Cold Craft, a fictional ice cream maker on the KZN South Coast. It has 11 products priced in rand, two locations, weekly orders from its customers, and purchase orders to its suppliers. Book a demo, and we will backtest Planner on a sample of your own history.
Demand planning software forecasts how much of each product your customers will buy, and turns the forecast into a stock plan. Miridia Planner does both, and the rest of Miridia raises the purchase orders and transfers from the plan.
It depends on your data, so Planner measures it for you. Each SKU is backtested on its own history, the champion must beat a naive benchmark, and forecast value added shows the effect of each change. Ask us to run a backtest on your data before you decide.
Yes. Planner reads the sales history from your connected channels, and it accepts a validated CSV upload for history from other systems. Sage Cloud connects directly, and Sage Evolution connects through consulting services.
Planner learns the seasons of each SKU from your own history, so a December peak is part of the model when it is part of your data. It does not start from a northern-hemisphere pattern.
Planner is an advanced module, priced separately on the Starter and the Enterprise plan. The Starter plan is R3,000 a month. Ask us for the Planner price on the discovery call.
Miridia handles personal data in line with POPI, and with GDPR for European and UK customers. Each organisation has its own database schema, and you can export all your data through the API. The Enterprise plan adds a self-hosted option.
MAKE. MOVE. MEASURE.
Book a demo. We will backtest a sample of your SKUs and show you the accuracy before you decide.