Demand Forecasting and Automated Replenishment

A distributor with ten thousand active item codes and three warehouses does not really have a forecasting problem, it has an attention problem. A buyer can only look at so many item codes before the day runs out, so purchasing collapses into two habits: reorder what a supplier rep reminded you about, and reorder what a customer just shouted about. Standard SAP Business One helps once you have already decided what to buy. The MRP wizard and the Order Recommendation Report both respect Lead Time, Order Multiple and Minimum Order Quantity from the Item Master Data planning tab, and they will net your on-hand against open documents. Neither of them builds a demand forecast. They consume the demand you hand them. Nothing in the standard product learns that a product line sells three times as fast in June, that one item swings so widely it needs six weeks of cover while another needs six days, or that the number typed into the Minimum Inventory Level field on an item-warehouse row eighteen months ago is now wrong in both directions across your catalogue.

What ECOSIRE builds

This is a build-to-order add-on. There is no instant download and nothing pre-packaged for your business. After a scoping call and a fixed written quote we build the forecasting and replenishment layer against your company database, your item master, your warehouse structure and your actual purchasing rules, install it into your test company first, then into production. Typical lead time is one working week from a signed quote.

The build has three parts. A forecast engine reads your posted sales history. A policy layer turns each forecast into a safety stock and a reorder point per item per warehouse. A suggestion workspace lets a buyer review, edit and release purchase documents. All three are implemented as SAP Business One extensions: Service Layer (REST) for reads and document creation, the DI API where a screen or document behaviour needs the client, and UDFs, UDTs and UDOs for everything we store. We do not write to SAP tables directly and we do not replace standard documents.

The forecast engine

Demand history is taken from posted A/R Invoices (OINV / INV1) and Deliveries (ODLN / DLN1), with the lookback window, the document types counted and the treatment of credit memos all configurable. Intercompany transfers, samples and warranty replacements can be excluded by document series, item group or UDF marker so they do not pollute the baseline.

On top of that history we fit several models and let you choose per item or per item group: moving average, weighted moving average, single exponential smoothing, Holt smoothing where a trend is real, and a multiplicative seasonal model where a seasonal index exists. Seasonal indices are calculated per item group in monthly or weekly buckets and stored in a User-Defined Table so a planner can inspect and adjust them rather than trusting a black box. Every item carries a stored forecast accuracy record, tracking mean absolute percentage error and bias over the last periods, so you can see which items the statistics handle well and which ones need a human number.

Human numbers are first class. A forecast override is a User-Defined Object with a quantity, a period, a reason code and a change history, layered on top of the statistical baseline rather than overwriting it. A tender win, a promotion, a new branch opening or a discontinued line all enter through overrides and remain visible as such in every report.

Policy per item and per warehouse

A forecast on its own does not tell a buyer when to act. The policy layer calculates safety stock from demand variability and lead time variability at a service level you choose, and derives a reorder point and a reorder-up-to level from it. Policies are set per item-warehouse row (OITW), not globally, because a slow line in your secondary warehouse should not carry the cover you need at your main branch.

An ABC and XYZ classification job segments items by revenue contribution and by demand variability, and each cell gets a default service level, review period and rounding rule. That is what makes the system usable at ten thousand item codes: the buyer configures nine policy cells, not ten thousand items, and only overrides the exceptions. Calculated minimum levels can either be written back to the standard Minimum Inventory Level on OITW so the native MRP wizard consumes them, or held in UDFs for review first, depending on how much you want the add-on to drive standard planning.

The suggestion workspace

The replenishment run nets projected demand against on hand, committed and ordered quantities from OITW plus open Purchase Order and Sales Order lines, then produces a suggestion list grouped by vendor. Suggested quantities honour Order Multiple, Minimum Order Quantity, Lead Time and Procurement Method from the Item Master Data planning tab, and vendor choice can follow the preferred vendor, Business Partner special prices, container fill targets or a split across two suppliers.

The buyer works the list rather than the whole catalogue: adjust quantity, change required date, switch vendor, defer a line, or add a note. Release creates Purchase Quotations or Purchase Orders through the Service Layer, carrying the suggestion reference into a UDF on the document so any purchase order can be traced back to the forecast, policy and person that produced it. Documents above a configurable value or below a configurable margin route through SAP Business One approval procedures before they become approved documents.

Reporting and audit

Delivered reports cover forecast versus actual by item group, coverage in days by warehouse, suggestion audit trail, excess and obsolete exposure, and a slow-mover list with last movement dates. Crystal Reports layouts are supplied for the ones your team prints or emails, and the underlying queries are documented so your own analysts can extend them.

Who this is for

Distributors, wholesalers and importers running SAP Business One 10.0 with more item codes than a buyer can personally remember, seasonal or promotion-driven demand, and lead times long enough that a wrong call costs a season. It suits companies that already trust their inventory data and want the buying decision improved, rather than companies still fixing counting discipline.

How delivery works

We start with a scoping call, usually ninety minutes, on your system: which documents represent real demand, how many warehouses and item codes, current lead times, who buys, what needs approval, and which reports you actually use. We then send a fixed written quote covering scope, screens, reports, install and training, with anything excluded stated plainly.

On acceptance we build against a copy of your database or an agreed test company. You get a walkthrough of the forecast engine on your own history, including a backtest that shows which model was chosen per item group and why, before any policy is allowed to write anything back. Installation goes into your test company first for user acceptance, then into production in an agreed window, with the UDF, UDT and UDO deployment scripted so it is repeatable across company databases.

The add-on runs on SAP Business One 10.0 on Microsoft SQL Server or SAP HANA, on premise or on a hosted server, and is delivered per tenant.

After go-live

You receive a two-week go-live support window for defect fixes and configuration adjustment, which is when most service level and rounding tuning actually happens, plus recorded training for buyers and administrators. Longer support and enhancement arrangements are quoted separately if you want them, and the source code and full data dictionary are handed over so you are never locked to us for a change.